Download: XMPP: The Definitive Guide

XMPP: The Definitive Guide XMPP: The Definitive Guide Building Real-Time Applications with Jabber Technologies Peter Saint-Andre, Kevin Smith, and Remko Tronçon Beijing • Cambridge • Farnham • Köln • Sebastopol • Taipei • Tokyo XMPP: The Definitive Guide by Peter Saint-Andre, Kevin Smith, and Remko Tronçon Copyright © 2009 Peter Saint-Andre, Kevin Smith, Remko Tronçon. All rights reserved. Printed in the United States of America. Published by O’Reilly Media, Inc., 1005 Gravenstein Highway North, Sebastopol, CA 95472. O’Reilly books may be purchased for educational, business, or sales promotion...
Author: Walton Shared: 7/30/19
Downloads: 640 Views: 3629


XMPP: The Definitive Guide,

XMPP: The Definitive Guide Building Real-Time Applications with Jabber Technologies

Peter Saint-Andre, Kevin Smith, and Remko Tronçon Beijing • Cambridge • Farnham • Köln • Sebastopol • Taipei • Tokyo,

XMPP: The Definitive Guide

by Peter Saint-Andre, Kevin Smith, and Remko Tronçon Copyright © 2009 Peter Saint-Andre, Kevin Smith, Remko Tronçon. All rights reserved. Printed in the United States of America. Published by O’Reilly Media, Inc., 1005 Gravenstein Highway North, Sebastopol, CA 95472. O’Reilly books may be purchased for educational, business, or sales promotional use. Online editions are also available for most titles ( For more information, contact our corporate/ institutional sales department: (800) 998-9938 or email is hidden.

Editor: Mary E. Treseler Indexer: Joe Wizda Production Editor: Loranah Dimant Cover Designer: Karen Montgomery Copyeditor: Genevieve d’Entremont Interior Designer: David Futato Proofreader: Loranah Dimant Illustrator: Robert Romano Printing History:

April 2009: First Edition. Nutshell Handbook, the Nutshell Handbook logo, and the O’Reilly logo are registered trademarks of O’Reilly Media, Inc. XMPP: The Definitive Guide, the image of a kanchil mouse deer on the cover, and related trade dress are trademarks of O’Reilly Media, Inc. JABBER® is a registered trademark licensed through the XMPP Standards Foundation. Many of the designations used by manufacturers and sellers to distinguish their products are claimed as trademarks. Where those designations appear in this book, and O’Reilly Media, Inc. was aware of a trademark claim, the designations have been printed in caps or initial caps. While every precaution has been taken in the preparation of this book, the publisher and authors assume no responsibility for errors or omissions, or for damages resulting from the use of the information con- tained herein. ISBN: 978-0-596-52126-4 [M],

Table of Contents

Preface .xi

Part I. An Overview of XMPP

1. Introduction .3 What Can You Do with XMPP? 3 Services 3 Applications 5 Brief History 7 Open Source and Open Standards 8 Extensibility 9 Summary 9 2. Basics of XMPP .11 Architecture 11 Addresses 14 Domains 15 Users 15 Resources 15 Internationalization 16 XMPP URIs 16 Streaming XML 16 Communication Primitives 18 Message 18 Presence 19 IQ 20 Extensibility 23 Asynchronicity 24 Error Handling 24 Hello Hello World World: Building a Basic XMPP Application 25 Summary 27 v,

Part II. The XMPP Toolkit

3. Presence .31 Is Anybody Home? 31 Authorization Required: The Subscription Handshake 31 How Presence Is Propagated 33 Availability Status 35 Presence Priorities 36 Directed Presence 37 Going Offline 37 Rich Presence 38 Presence and Rosters 39 Using Presence 42 Presence-Based Routing 42 Access Control 43 Presence As a Transport 43 Summary 44 4. Instant Messaging .45 I Think, Therefore IM 45 Chat Sessions 47 Are You There? Chat State Notifications 48 Looks Matter: Formatted Messages 52 Who Are You? vCards 53 Talk to the Hand: Blocking and Filtering Communication 55 Blocking: The Simple Approach 55 Advanced Blocking and Filtering 57 More Messaging Extensions 58 Summary 58 5. Discovering the World .59 Items and Info 59 Using Service Discovery with Servers and Services 61 Using Service Discovery with Clients 64 Explicit Service Discovery 64 Entity Capabilities: Service Discovery Shorthand 66 Summary 68 6. Data Forms .69 Basic Structure 69 Using Data Forms 71 Defining Your Terms: Form Types 73 vi | Table of Contents, Including Media in Data Forms 74 Summary 75 7. Multi-Party Interactions .77 Starting the Party 77 Groupchat Basics 78 Crowd Control 81 What’s in a Nick? 85 Configure This! 87 Privacy, Security, and All That Jazz 91 MUC As a Data Transport 92 Summary 93 8. Publish/Subscribe .95 Why It Matters 95 Quickstart 97 Subscriptions 98 Publishing and Receiving Notifications 100 Payloads: To Send or Not to Send? 102 Items: To Store or Not to Store? 103 Discovering Nodes 104 Node Management 107 Creating and Deleting Nodes 107 Node Configuration 108 Managing Node Access 112 Item Aggregation via Collection Nodes 114 Personal Eventing: PubSub Simplified 117 Summary 122 9. Jingle: Jabber Does Multimedia .123 To Instant Messaging and Beyond 123 The Jingle Model 124 Making a Call 127 A Swarm of NATs 131 Jingle on ICE 132 Additional Jingle Actions 135 Summary 136 10. Sending Binary Data .137 Starting Small: Bits of Binary 137 Moving On Up: Transferring Midsize Files In-Band 139 Thinking Big: Sending Large Files Out-of-Band 142 Sending Data Directly 142 Table of Contents | vii, Sending Data Through a Proxy 143 Negotiating File Transfer 145 File Transfer Using Stream Initiation 146 Session Negotiation Using Jingle 149 Summary 152 11. Remote Commands .153 Controlling Clients 153 A Simple Command 154 Commands and Data Forms 156 Providing Custom Commands 160 Advanced Workflows: SOAP, RPC, IO Data 163 Summary 163 12. Connection Methods and Security .165 Negotiating an XMPP Stream 165 Authentication Options 171 Encrypting the Connection 172 Server Federation 174 Server Components 179 BOSH: XMPP over HTTP 180 Serverless Messaging 189 XMPP Security 192 Encryption 193 Authentication and Identity 194 Spam and Abuse 195 Summary 196

Part III. Putting It All Together

13. Design Decisions .199 Is XMPP the Right Choice? 199 How the XMPP Community Works 201 Writing XMPP Software 202 Mixing, Matching, and Extending Existing XMPP Software 202 Client Extension, Bot, Component, or Server Module? 203 Rolling Your Own Client or Server 205 Extending XMPP 207 How to Design Custom Extensions 207 Standardizing New Extensions 209 Summary 210 viii | Table of Contents, 14. Building an XMPP Application .211 The CheshiR Microblogging Platform 211 First Sprint: The CheshiR XMPP IM Bot 211 Analysis 211 Design 212 Coding 214 Second Sprint: Configuring the CheshiR XMPP IM Bot 216 Analysis 216 Design 216 Coding 217 Third Sprint: Scaling the CheshiR XMPP Service Using a Server Component 218 Analysis 218 Design 219 Coding 220 Fourth Sprint: Registering with the CheshiR Server Component 221 Analysis 221 Design 222 Coding 222 Fifth Sprint: Extending the Server Component with Rosters 224 Analysis 224 Design 224 Coding 224 Future Sprints 227 A CheshiR Server Module or Dedicated Server? 227 Summary 228

Part IV. Appendixes

A. A Guide to XMPP Specifications .231 B. Popular Servers, Clients, and Libraries .253 C. Further Practical Considerations .263 Glossary .269 Bibliography .273 Index .277 Table of Contents | ix,

Preface Why XMPP?

In 1800, it took one or two years to send a message from London to Calcutta and receive a reply. You needed to find a ship’s captain you trusted, who piloted his sailing ship around the Cape of Good Hope and probably stopped in various ports along the way. Then your contact in Calcutta needed to write a reply and send it back to London in a similar fashion. Not exactly instant messaging! With the invention of the steamship and the opening of the Suez Canal, the time was reduced to a month or two. Air mail reduced the time further to a week or two, and eventually to a few days (“when it absolutely, positively has to be there overnight”). The deployment of commercial email systems introduced us to wait times of only a few minutes (depending on how often you polled your server). And instant messaging (IM) systems such as ICQ® took communication to its logical conclusion: nearly immediate interaction. As a result of these developments, the useful half-life of information has shrunk sig- nificantly, in many cases to mere seconds. For many people, IM trumps email. Blogging trumps newspapers and magazines. Microblogging trumps blogging. Groupchat trumps email discussion lists. Shared editing and whiteboarding trump carefully crafted presentations. Immediate notifications trump once-a-day updates. And the list goes on. What all these technologies have in common is that the interactions happen in close to real time. To make this possible, we need technologies for real-time communication. Ideally such technologies would be open standards providing the real-time equivalent of HTTP, HTML, and the other building blocks of today’s Internet, because over the long term open standards provide stronger security, greater extensibility, and the pos- sibility for more innovation at the edges than do closed technologies. The Extensible Messaging and Presence Protocol (XMPP) is just such an open tech- nology for real-time interaction. Consider some of its advantages: • XMPP is proven. Over 10 years of development has resulted in a stable, widely deployed, seriously tested, Internet-scale technology, with dozens of interoperable codebases, tens of thousands of deployed services, and millions of end users. xi, • XMPP is secure. It provides built-in support for channel encryption and strong authentication, inherent resistance to many forms of malware, a diverse ecosystem of implementations, a decentralized network without a single point of failure, and significant deployment at some of the most security-conscious financial organiza- tions and government agencies worldwide. Work on more advanced features (such as user-friendly end-to-end encryption) continues so that XMPP will be even more secure. • XMPP is decentralized. Unlike standalone communication silos, XMPP technolo- gies are deployed in a decentralized client-server architecture with an unlimited number of servers. Any person or organization can run their own XMPP server and connect it to the rest of the network using standard Internet infrastructure such as the Domain Name System (DNS), and certificates are freely available through the XMPP Standards Foundation (XSF) to enable secure federation of XMPP traffic. • XMPP is extensible. Because XMPP is at its core a technology for rapidly delivering XML from one place to another, it has been used for a wide range of applications beyond instant messaging, including gaming, social networking, Voice over IP (VoIP), real-time collaboration, alerts and notifications, data syndication, geolo- cation, intelligent workflows, machine-to-machine communication, and custom applications. • XMPP is scalable. The “push” model of information transfer used in XMPP solves serious scaling problems associated with traditional HTTP-based polling ap- proaches; as a result, it enables you to build applications that were literally im- possible until now. • XMPP is a standard. The core aspects of XMPP have undergone rigorous public review within the Internet Engineering Task Force (IETF), and extensions to XMPP are published in an open, developer-oriented standards process run by the XSF. This approach has resulted in strong technologies that can be freely implemented under any licensing terms, from open source to shareware to proprietary code. • XMPP is a community. Open standards, a large number of software products, and a communications network are all good, but the “secret sauce” of XMPP may be its vibrant and friendly community of technologists, developers, open source projects, commercial software companies, service providers, and end users. This community is committed to working together to solve problems and build great new applications. For these reasons, more and more software developers and service providers are using XMPP to build real-time applications or add real-time interfaces to existing applica- tions. And you can, too, because XMPP provides a simple but powerful set of tools that can help you solve real-world problems. This book will show you how. xii | Preface, Jabber and XMPP Throughout this book, we use the terms “Jabber” and “XMPP” inter- changeably. These technologies were originally developed by Jeremie Miller and the Jabber open source community in 1998–1999. When the community submitted its core protocols to the Internet Engineering Task Force (IETF) in 2002, it chose the name “Extensible Messaging and Presence Protocol” to distinguish the protocol from the broader technology and developer community. You can think of the relationship as “XMPP is to Jabber as HTTP is to the Web.” The term Jabber was proactively trademarked by Jabber, Inc. (now part of Cisco Systems, Inc.) in 2000 to protect the open source community, but the XSF sub- licenses the term for use in open source projects and other community activities.

Is This Book for You?

This book may be for you if: • You are a software developer who needs a helpful guide to building a real-time application or extending an existing system, as well as relevant reference materials to use during your project. • You are a product manager or software architect who is looking for suggestive ideas and case studies regarding real-time systems. • You are a software architect or developer who needs a brief but thorough overview of XMPP. • You are a researcher, teacher, or student who is designing a research project. • You are interested in new technologies and the emergence of the real-time Internet. Above all, this book provides a practical guide to XMPP. Because XMPP is a well- documented protocol, we regularly refer you to the XMPP specifications for relevant details (these specifications come in two flavors: the core protocols are defined in the Requests for Comments or “RFC” series published by the IETF, and dozens of exten- sions are defined in the XMPP Extension Protocol or “XEP” series published by the XSF). Because XMPP is widely supported by a large number of servers, clients, and code libraries, both open source and commercial, we refer you to those projects for assistance with real-world implementation. Instead of covering all protocol details and possible implementations, we show how XMPP technologies can be used to solve cer- tain classes of problems by helping you to “think in XMPP” and covering the “gotchas” that can trip up those who are new to XMPP technologies. Throughout this book, we assume that you are familiar with the very basics of computer networking, common Internet applications (such as email and the World Wide Web), and structured data formats (such as HTML). However, we often treat these technol- ogies as the starting points for our discussion or as “contrast objects” for XMPP, which Preface | xiii, differs from applications such as the Web in important ways that we’ll describe as we go. Finally, we include some examples using the Python programming language, so some familiarity with Python can also help you understand the concepts we describe.

Getting the Most Out of This Book

To get the most out of this book, we do not recommend that you read it cover to cover in one sitting (although you are welcome to do so!). Instead, first explore the sections that interest you or that you need to complete a particular task, perhaps after reading the introductory materials in Part I. You might also consider skimming over the details of each XML example on your first reading so that you get the general idea of each use case and protocol extension. The book is organized as follows: • Part I provides an overview of XMPP. The first chapter talks about XMPP at a high level and introduces you to some ways XMPP is being used to build real-time applications. The second chapter describes the basics of XMPP technologies, in- cluding architectural issues, addressing, and communication primitives. Read this section first if you’d like a relatively quick orientation to XMPP technologies. • Part II consists of a series of “developer stories” that illustrate how the tools in the XMPP toolkit can help you solve particular classes of problems. Each chapter in Part II introduces the XMPP concepts and services that you need in a given problem domain, describes how to use those tools, and provides examples showing how specific protocols come into play. Read the chapters here that interest you most. The order doesn’t matter, because we recap concepts where needed, and provide cross-references to more detailed treatments in other chapters. • Part III shows you how to put it all together by walking you through the thought processes and design decisions involved in building an XMPP-based application. Read this part after you have a feel for XMPP from the first two parts, and as you begin to dig into a large project that uses XMPP to construct a business application or real-time service. • Part IV consists of the appendixes, which help you understand the terminology of XMPP; introduce you to the wealth of XMPP servers, clients, and code libraries; and guide you through the large “stack” of XMPP protocol specifications so you can quickly find what you need. Use these appendixes as reference material on an ongoing basis, or as a quick index to the myriad of XMPP resources available on the Internet. xiv | Preface,

Conventions Used in This Book

The following typographical conventions are used in this book: Italic Indicates new terms, URLs, email addresses, filenames, and file extensions. Constant width Used for protocol examples and sample code, as well as within paragraphs to refer to protocol aspects such as XML elements, attributes, and namespaces, code fea- tures such as variable and function names, databases, data types, environment variables, statements, keywords, etc. Constant width bold Indicates user input in examples showing an interaction. Also indicates empha- sized code elements to which you should pay particular attention. This icon signifies a tip, suggestion, or general note. This icon indicates a warning or caution.

About the Examples

In Parts I and II, we include a large number of protocol examples (but not nearly as many as you will find in the XMPP specifications, which are extremely thorough). Each example contains a snippet of XML that would be sent over the wire to communicate a message, share presence information, retrieve data, initiate a command sequence, return an error, and the like. These chunks of XML are essentially copied directly from the XMPP specifications with additional notes to highlight their most important and relevant aspects. However, sometimes our examples are incomplete or shortened for readability, so be sure to check the official XMPP specifications for the most accurate examples and protocol descriptions! Most of the examples in this book use Lewis Carroll’s Alice’s Adventures in Wonder- land as the backdrop (Alice and her friends are much more interesting than “User A” and “User B”!). The domain names in these examples are things like wonderland.lit, which clearly don’t work on today’s Internet, because the .lit top-level domain has not yet been assigned. This is intentional (we don’t want to bother anyone who owns a real domain name like Preface | xv, In Part III, we intersperse protocol examples with software code showing one possible implementation of several protocol interactions. This software code is written in the Python programming language, a popular, easy-to-read language for scripting and ap- plication development.

Using Code Examples

This book is here to help you get your job done. In general, you may use the code in this book in your programs and documentation. You do not need to contact us for permission unless you’re reproducing a significant portion of the code. For example, writing a program that uses several chunks of code from this book does not require permission. Selling or distributing a CD-ROM of examples from O’Reilly books does require permission. Answering a question by citing this book and quoting example code does not require permission. Incorporating a significant amount of example code from this book into your product’s documentation does require permission. We appreciate, but do not require, attribution. An attribution usually includes the title, author, publisher, and ISBN. For example: “XMPP: The Definitive Guide, by Peter Saint-Andre, Kevin Smith, and Remko Tronçon. Copyright 2009 Peter Saint-Andre, Kevin Smith, and Remko Tronçon, 978-0-596-52126-4.” If you feel your use of code examples falls outside fair use or the permission given above, feel free to contact us at email is hidden.

Safari® Books Online

When you see a Safari® Books Online icon on the cover of your favorite technology book, that means the book is available online through the O’Reilly Network Safari Bookshelf. Safari offers a solution that’s better than e-books. It’s a virtual library that lets you easily search thousands of top tech books, cut and paste code samples, download chapters, and find quick answers when you need the most accurate, current information. Try it for free at

How to Contact Us

Please address comments and questions concerning this book to the publisher: O’Reilly Media, Inc. 1005 Gravenstein Highway North Sebastopol, CA 95472 800-998-9938 (in the United States or Canada) 707-829-0515 (international or local) 707-829-0104 (fax) xvi | Preface, We have a web page for this book, where we list errata, examples, and any additional information. You can access this page at: To comment or ask technical questions about this book, send email to: email is hidden For more information about our books, conferences, Resource Centers, and the O’Reilly Network, see our website at: Finally, the authors of this book can usually be found on the XMPP network in the email is hidden chat room.


We would like to thank Mary Treseler for her editorial guidance throughout this project, and her patience with an enthusiastic but not entirely disciplined group of authors. We’d also like to thank our technical reviewers for their thorough comments on, and insight into, the contents of this book; it was greatly improved by their input. Thank you Dave Cridland, Brian Dainton, Kellan Elliott-McCrea, Michelle Fisher, Nathan Fritz, and Jack Moffitt.

Peter Saint-Andre

Thanks are due to the many developers who helped me understand these technologies as they were being designed in the early days of the Jabber open source community. I would like to especially recognize the help of my friend Peter Millard, who patiently answered my never-ending questions about Jabber technologies from 1999 until his death in 2006. I dedicate my work on this book to his memory. I would not have been able to contribute to XMPP all these years without the generous support of my employer, Jabber, Inc. (now part of Cisco Systems, Inc.). Most fundamentally, my wife, Elisa, has always cheerfully tolerated my obsession with XMPP despite countless hours working on specs, posting to discussion lists, writing blog entries, traveling to conferences, and all the rest.

Kevin Smith

Many of the members of the XMPP community have been supportive over the last seven or so years since my first involvement, and I’d like to acknowledge particularly those people I’ve worked with in the Psi and Sleek projects, and those I’ve worked with on the XMPP Council in expanding my knowledge of XMPP and software development. Thanks to Peter and Remko especially, for all the fun we’ve had with this book. Preface | xvii, My wife, Cath, has my unending gratitude for her support in my numerous XMPP- related and other free-time-swallowing commitments.

Remko Tronçon

My first words of gratitude go to my coauthors, Peter and Kevin. Not only did they make the writing of this book an incredibly fun experience, but they are also the reason why I got into XMPP in the first place. Thanks to Kevin, the other Psi developers, and the whole Psi userbase, I got the chance to take my first steps into the XMPP world. Thanks to the support of “patron saint” Peter and the rest of the XMPP community, I was able to take this involvement one step further, and joined the conversation to define the XMPP standards. The XMPP community is without a doubt one of the most pleas- ant and accessible groups of people out there on the interwebs. Thanks to everyone out there who ever talked to me! My most important source of inspiration, however, comes from outside the digital world. Kim has always unconditionally supported me in all my time-consuming activ- ities, and has continuously pushed me to work harder, even in times when she hardly received any of the attention she deserved. xviii | Preface, PART I An Overview of XMPP,

CHAPTER 1 Introduction What Can You Do with XMPP?

The Extensible Messaging and Presence Protocol (XMPP) is an open technology for real- time communication, using the Extensible Markup Language (XML) as the base format for exchanging information. In essence, XMPP provides a way to send small pieces of XML from one entity to another in close to real time. XMPP is used in a wide range of applications, and it may be right for your application, too. To envision the possibilities, it’s helpful to break the XMPP universe down at a high level into services and applications. The services are defined in two primary spec- ifications published by the Internet Engineering Task Force (IETF) at (the “RFC” series), and in dozens of extension specifications published by the XMPP Standards Foundation at (the “XEP” series); the applications are soft- ware programs and deployment scenarios that are of common interest to individuals and organizations, although the core services enable you to build many other applica- tion types as well. RFC Revisions As of this writing, [RFC 3920] and [RFC 3921] are under active revision to incorporate errata, clarify ambiguities, improve their readability, de- fine additional error codes, etc. These documents, called [rfc3920bis] and [rfc3921bis] in the terminology of the IETF, provide the most ac- curate definition of XMPP and might have been published as replace- ment RFCs (with new numbers) once you read this book. For the latest versions of the revised specifications, visit


In this context, a service is a feature or function that can be used by any given applica- tion. XMPP implementations typically provide the following core services:, Channel encryption This service, defined in [RFC 3920] and explained in Chapter 12 of this book, provides encryption of the connection between a client and a server, or between two servers. Although channel encryption is not necessarily exciting, it is an im- portant building block for constructing secure applications. Authentication This service, also defined in [RFC 3920] and explained in Chapter 12 of this book, is another part of the foundation for secure application development. In this case, the authentication service ensures that entities attempting to communicate over the network are first authenticated by a server, which acts as a kind of gatekeeper for network access. Presence This service, defined in [RFC 3921] and explained in Chapter 3 of this book, en- ables you to find out about the network availability of other entities. At the most basic level, a presence service answers the question, “Is the entity online and avail- able for communication, or offline and not available?” Presence data can also include more detailed information (such as whether a person is in a meeting). Typically, the sharing of presence information is based on an explicit presence subscription between two entities in order to protect the privacy of user information. Contact lists This service, also defined in [RFC 3921] and explained in Chapter 3 of this book, enables you to store a contact list, or roster, on an XMPP server. The most common use for this service is an instant messaging “friend list,” but any entity that has an account on a server can use the service to maintain a list of known or trusted entities (e.g., it can be used by bots). One-to-one messaging This service, defined in [RFC 3920] and explained in Chapter 4 of this book, en- ables you to send messages to another entity. The classic use of one-to-one mes- saging is personal IM, but messages can be arbitrary XML, and any two entities on a network can exchange messages—they could be bots, servers, components, de- vices, XMPP-enabled web services, or any other XMPP entity. Multi-party messaging This service, defined in [XEP-0045] and explained in Chapter 7 of this book, ena- bles you to join a virtual chat room for the exchange of messages between multiple participants, similar to Internet Relay Chat (IRC). The messages can be plain text, or can contain XML extensions for more advanced functionality, such as room configuration, in-room voting, and various session control messages. Notifications This service, defined in [XEP-0060] and explained in Chapter 8 of this book, ena- bles you to generate a notification and have it delivered to multiple subscribers. 4 | Chapter 1: Introduction, This service is similar to multi-party messaging, but it is optimized for one-to-many delivery with explicit subscriptions to specific channels or topics (called “nodes”). Service discovery This service, defined in [XEP-0030] and explained in Chapter 5 of this book, ena- bles you to find out which features are supported by another entity, as well as any additional entities that are associated with it (e.g., rooms hosted at a chat room service). Capabilities advertisement This service, defined in [XEP-0115] and explained in Chapter 5 of this book, is an extension to the presence service that provides a shorthand notation for service discovery data so that you can easily cache the features that are supported by other entities on the network. Structured data forms This service, defined in [XEP-0004] and explained in Chapter 6 of this book, ena- bles you to exchange structured but flexible forms with other entities, similar to HTML forms. It is often used for configuration and other tasks where you need to gather ad-hoc information from other entities. Workflow management This service, defined in [XEP-0050] and explained in Chapter 11 of this book, enables you to engage in a structured workflow interaction with another entity, with support for typical workflow actions, such as moving to the next stage of a business process or executing a command. It is often used in conjunction with data forms. Peer-to-peer media sessions This service, defined in [XEP-0166] and explained in Chapter 9 of this book, ena- bles you to negotiate and manage a media session with another entity. Such a session can be used for the purpose of voice chat, video chat, file transfer, and other real-time interactions. These are some of the core services available to you (or your application) as a participant in an XMPP network. The XMPP developer community has defined additional features in various XMPP extensions, but here we focus on the services that we think you will find most useful in building real-time applications.


Given that you have a dozen core services at your disposal, what can you build? Here are a few possibilities: Instant messaging The classic instantmessaging systems that most people are familiar with combine three of the core services: presence, contact lists, and one-to-one messaging. Such What Can You Do with XMPP? | 5, systems can and often do include more services and features, but if you have these three services, you can build a bare-bones IM application. Groupchat The multi-party messaging service enables you to build groupchat systems similar to IRC. Often, groupchat systems are used for more specific applications, such as real-time trading systems in the financial industry, situation rooms for first res- ponders and military personnel, and virtual classrooms. Gaming Combined with custom extensions, both one-to-one messaging and multi-party messaging enable you to build simple gaming systems. For example, the Chesspark service ( is built entirely using XMPP. Other game de- velopers are using XMPP to add presence and contact list features to existing multi-party games. Systems control The combination of one-to-one messaging and data forms makes it possible to deploy lightweight systems for control of and interaction with remote systems. Deployed applications in this domain include network management, scientific telemetry, and robotic control. Geolocation The XMPP notification service is payload-agnostic. One defined payload format is geolocation, which enables you to build fascinating location-based applications, such as vehicle tracking. Middleware and cloud computing A number of companies and research groups are actively working on XMPP-based systems for computation services, lightweight middleware, and management of cloud computing infrastructures. While the use of XMPP may be surprising here because such applications have traditionally relied on heavyweight messaging technologies, we have seen XMPP begin to nibble away at the lower end of this market. It appears that companies that already have an XMPP infrastructure in place figure they might as well use it for non-IM use cases. These systems often use the workflow extensions we explore in Chapters 6 and 11 for structured message exchange. Specific applications include bioinformatics. Data syndication Popular social networking applications are increasingly using the XMPP notifica- tion service to solve a particular problem they have: constant polling for updated information. Existing HTTP-based deployments have been found not to scale, be- cause quite often a particular feed has not changed since the last time it was polled. By contrast, the XMPP notification service sends out an update only when a feed has changed, saving a significant amount of bandwidth and server resources that otherwise would be wasted on polling. 6 | Chapter 1: Introduction, Voice over IP (VoIP) The Google Talk application that launched in August 2005 first popularized the use of XMPP for voice chat. Since then, the XMPP extensions for media session services (called Jingle) have been formalized through the XSF, and have been implemented and deployed by the likes of Nokia and the One Laptop Per Child project. The same extensions can also be used to negotiate a wide range of media session types, including video, file transfer, whiteboarding, and collaborative editing. Identity services Given the existence of stable identifiers (JabberIDs) and a robust authentication service, it is possible to use XMPP in building identity and authorization services such as OpenID and OAuth. Other application examples include data transfer, live chat integrated into websites, mobile device communications, and presence-enabled directories. We will mention relevant applications throughout this book to illustrate the most popular and interest- ing uses of XMPP. Although we highlight many applications of XMPP, unfortunately we can’t cover all of them. Not only do we lack the space and time, but the list keeps growing every day. Moreover, the most cutting-edge uses of XMPP are not standardized yet, which makes them too much of a moving target to describe in a book. Examples of ongoing work at the time of this writing include collaborative document editing, whiteboarding, calen- dar integration, file sharing, and personal media networks. If you want to learn more about these topics, we suggest that you get involved with the XMPP community (see Chapter 13) as we define new ways of using XMPP. What does the future hold for XMPP technologies? Although we don’t know for sure, the trends seem clear: deployment of XMPP systems at more organizations and service providers, XMPP interfaces to more web applications, use of XMPP features to solve more business problems, and continued growth in the XMPP developer community. It’s an exciting time to be working on XMPP technologies, and we invite you to join the conversation!

Brief History

Jabber/XMPP technologies were invented by Jeremie Miller in 1998. Jeremie was tired of running four different clients for the closed IM services of the day, so in true open source fashion, he decided to scratch an itch, releasing an open source server called jabberd on January 4, 1999. Before long, a community of developers jumped in to help, writing open source clients for Linux, Macintosh, and Windows; add-on components that worked with the server; and code libraries for languages such as Perl and Java. During 1999 and early 2000, the community collaboratively worked out the details of Brief History | 7, the wire protocols we now call XMPP, culminating in the release of jabberd 1.0 in May 2000. As the community grew larger and various companies became interested in building their own Jabber-compatible (but not necessarily open source) software, the loose col- laboration evident in 1999 and 2000 became unsustainable. As a result, the community (spearheaded by a company called Jabber, Inc., acquired by Cisco in late 2008) formed the Jabber Software Foundation in August 2001. Ever since, this nonprofit membership organization, renamed the XMPP Standards Foundation in early 2007, has openly documented the protocols used in the developer community, and has defined a large number of extensions to the core protocols. After several years of implementation and deployment experience, members of the developer community decided to seek a wider review of the core protocols by formal- izing them within the IETF, which has standardized most of the core technologies for the Internet (including TCP/IP, HTTP, SMTP, POP, IMAP, and SSL/TLS). Given that most good protocols seem to be three- or four-letter acronyms ending with the letter “P,” the relevant IETF working group labeled its topic the Extensible Messaging and Presence Protocol (XMPP). After less than two years of intensive work (mostly focused on tightening communications security), the IETF published the core XMPP specifi- cations in its Request for Comments (RFC) series as [RFC 3920] and [RFC 3921] in October 2004. Publication of these RFCs has resulted in widespread adoption of XMPP technologies. In August 2005, the Google Talk IM and Voice over Internet Protocol (VoIP) service was launched on a basis of XMPP. Thousands more services have followed. Prominent and emerging software companies use XMPP in their products, including the likes of Apple, Cisco, IBM, Nokia, and Sun. Countless businesses, universities, and govern- ment agencies have deployed XMPP-based instant messaging systems for their users. Many game developers and social networking applications are building XMPP into their services, and a number of organizations have used XMPP as the “secret sauce” behind some of their most innovative features.

Open Source and Open Standards

Although XMPP was originally developed in the Jabber open source community, the protocol itself is not an open source project like Apache, but rather an open standard like HTTP. As a result, XMPP is an open technology that is not tied to any single software project or company. The XMPP specifications define open protocols that are used for communication among network entities. Much as HTTP and HTML define the protocols and data formats that power the World Wide Web, XMPP defines the protocols and data formats that power real-time interactions over the Internet. The protocols are as free as the air, which means they can be implemented in code that is licensed as free software, open source software, shareware, freeware, commercial prod- ucts, or in any other way. This open standards model is different from the open source 8 | Chapter 1: Introduction, or free-software model for software code, wherein the code is often licensed so that modifications must be contributed back to the developers. That said, XMPP emerged from an open source developer community, specifically the community that formed around the open source jabberd server that Jeremie Miller released. Thus there are many open source implementations of XMPP, which can be downloaded for free by end users, system administrators, and developers alike. Much of this software is focused on instant messaging, as befits a technology that started as an open alternative to closed IM silos that did not interoperate. There are open source clients for just about every operating system and device; as a result, millions of end users communicate using XMPP-based services. There are open source servers that can be deployed at companies, schools, and service providers; as a result, tens of thousands of XMPP services inter-connect every day. There are open source libraries for all the major programming languages, which can be used to write bots, components, and other real-time applications; as a result, there are thousands of active developers in the XMPP community. Much of this software is linked to from, and we provide an overview of some of the most popular codebases in Appendix C.


The original Jabber developers were focused on building an instant messaging system. However, the extensible nature of XML has made XMPP attractive to application de- velopers who need a reliable infrastructure for rapidly exchanging structured data, not just IM features. As a result, XMPP has been used to build a wide range of applications, including content syndication, alerts and notifications, lightweight middleware and web services, whiteboarding, multimedia session negotiation, intelligent workflows, geolocation, online gaming, social networking, and more. Over the years, the developer community defined a large number of extensions to the core protocols. These extensions are developed through an open, collaborative stand- ards process and published in the XSF’s XMPP Extension Protocol (XEP) series at http: // As you’ll discover, the core protocols and various extensions provide a long “runway” for just about any feature you might need in developing real-time ap- plications. But if you find that a feature is missing from the XMPP protocol stack, it is easy enough to extend the protocol for your own purpose, and (optionally) work with the community in standardizing these new features, as we discuss in Chapter 13.


In this chapter, we looked at the core services XMPP provides and sampled the kinds of applications you can build with those services. Next, you’ll get acquainted with the basic workings of XMPP, after which we’ll dive into each of the core XMPP services in detail. Summary | 9,

CHAPTER 2 Basics of XMPP

This chapter outlines the fundamental features used by all XMPP-based applications. We first describe the generic architecture of XMPP systems and then the addressing scheme for XMPP communications, the three communication “primitives,” the model for sharing information about availability on the network (called presence), and the processes for session establishment.


All good Internet technologies have an “architecture”—a way that various entities fit together, link up, and communicate. For example, the World Wide Web consists of millions of web servers running software like Apache, and many more millions of web clients (browsers) running software like Firefox, all using standard protocols and data formats like HTTP and HTML. As another example, the email infrastructure consists of millions of email servers running software like Postfix, and many more millions of email clients running software like Thunderbird, all using standard protocols like SMTP, POP, and IMAP. Similarly, the Internet’s infrastructure for instant messaging, presence, and other forms of real-time communication increasingly consists of hundreds of thousands of Jabber servers running software like ejabberd and Openfire, and millions of Jabber clients running software like Adium, Gajim, Pidgin, and Psi, all using the standard protocol we call XMPP. XMPP technologies use a decentralized client-server architecture similar to the archi- tectures used for the World Wide Web and the email network. The diagram in Fig- ure 2-1 is a simplified representation showing three servers, each with three clients. The beauty of using a decentralized client-server architecture is that it enables an in- telligent separation of concerns (client developers can focus on user experience, and server developers can focus on reliability and scalability), it is much easier for organi- zations to manage than a purely peer-to-peer technology, it is quite robust because the full system does not have a single point of failure (anyone can run their own XMPP, Figure 2-1. XMPP uses a client-server architecture similar to email and the World Wide Web server and thereby join the network), and the servers can enforce important security policies such as user authentication, channel encryption, and prevention of address spoofing. Finally, the XMPP community has always worked to keep clients simple and to push as much complexity as possible onto the servers, further enabling widespread adoption of the technology. (We discuss the core XMPP design principles more fully in Chapter 13.) However, there are some important architectural differences between the Web, email, and Jabber. When you visit a website, your browser connects to a web server, but web servers typically do not connect to each other in order to complete a transaction (see Fig- ure 2-2). Instead, the HTML of the web page may refer to other web servers (e.g., to load images or scripts), and your browser opens sessions with those web servers to load the full page. Thus, the Web typically does not involve inter-domain connections (often called federation, and shown in Figure 2-1 by the double line). When you send an email to one of your contacts at a different domain, your email client connects to your “home” email server, which then seeks to route the message to your contact. Thus, unlike the Web, the email system consists of a federated network of servers. However, your message might be routed through multiple intermediate email 12 | Chapter 2: Basics of XMPP, Figure 2-2. The World Wide Web has many servers and clients, but very few server-to-server connections servers before it reaches its final destination (see Figure 2-3). Thus, the email network uses multiple hops between servers to deliver messages. Figure 2-3. The email network has many servers and clients, plus the servers are interconnected in a multi-hop network Like email, but unlike the Web, XMPP systems involve a great deal of inter-domain connections. However, when you send an XMPP message to one of your contacts at a different domain, your client connects to your “home” server, which then connects directly to your contact’s server without intermediate hops (see Figure 2-4). This direct federation model has important differences from the indirect federation model used in email (in particular, it helps to prevent address spoofing and certain forms of spam). Architecture | 13, Figure 2-4. The XMPP network has many servers and clients, plus the servers are interconnected in a single-hop network Table 2-1 summarizes these differences. Table 2-1. Client-server architectures Feature Web Email Jabber Interdomain Connections No Yes Yes Multiple Hops N/A Yes No Although clients and servers are the fundamental entities on an XMPP network, other entities play a part, too. Automated clients called bots provide a wide range of com- munication services, including assistance in chat rooms and human-friendly interfaces to non-XMPP services such as social networking applications. Furthermore, most XMPP servers are written in a modular way that enables administrators to add speci- alized services or server components, such as multi-user chat rooms, publish- subscribe topics, gaming arbiters, and the like. We discuss bots and components later in this book, especially in Part III.


Because XMPP communications happen on a network, every XMPP entity needs an address, called a JabberID (JID). XMPP typically relies on the Domain Name System (DNS) to provide the underlying structure for addressing, instead of using raw Internet Protocol (IP) addresses. After all, it’s much easier to remember that there is an XMPP service running at than to remember Similarly, JabberIDs for users look like email addresses (e.g., email is hidden) because the format 14 | Chapter 2: Basics of XMPP, email is hidden is already familiar to people; furthermore, this format uses the com- plete DNS infrastructure as its address space, unlike older IM systems that used num- bers or names without any domain identifier.


Every JabberID contains a domain portion, which typically maps to a fully qualified domain name (FQDN). When you install your favorite XMPP server software, you choose a domain name for the deployment, such as or Using DNS service label records, your domain name maps to one or more particular machines, such as or Those machine names in turn map to particular IP addresses, such as or (We discuss de- ployment scenarios further in the appendixes.) However, for the purposes of addressing on the network, all we need to care about is the domain name itself (e.g., or, rather than the lower-level machine names and IP addresses. Finally, for ASCII characters, the domain portion of a JID is case-insensitive (so that JABBER.ORG is the same as; as we explain later, the rules for non-ASCII characters are a bit more complex.


When you create an account at an XMPP service such as, you choose a JabberID that functions as your virtual identity on the network. Alternatively, your JabberID might be assigned to you automatically. Your JabberID looks much like an email address (e.g., email is hidden). Depending on deployment policies, it might even be the same as your email address at a service or company (e.g., your Google Talk address on the XMPP network looks the same as your Gmail address on the email network). As for the domain portion of the JabberID, the username portion of a JID is case-insensitive for ASCII characters (so that email is hidden is the same as email is hidden). XMPP developers usually call an address of the form email is hidden a bare JID.


When you connect your client to an XMPP server, you choose (or the server assigns to you) a resource identifier for that particular connection. This resource is used for routing traffic to that connection instead of any other connections you might have open at the moment. The resource is added to the end of your account address, such as email is hidden/roundabout or email is hidden/home. This enables someone to query or exchange messages with a particular device that is associated with your ac- count; it also means that each device is a separate “point of presence,” with different availability states, capabilities, etc. The resource is often the name of your computer, your location, or the client software you are using, but can be any string (including Addresses | 15, spaces and other special characters). Contrary to the other parts of a JID, the resource portion is case-sensitive (e.g., email is hidden/home is different from remko@el- XMPP developers usually call an address of the form email is hidden/ resource a full JID.


A major difference between JabberIDs and email addresses is that XMPP is fully inter- nationalized. This means that XMPP domain names and user names are not limited to the boring old ASCII character range, but can include virtually any Unicode character. If you live in the Czech Republic, you could run a Jabber server at a domain such as č, and you could have an address such as jiři@č Or, if you enjoy math- ematics and happen to own the domain, your JabberID could be something fun like ∞ (try that with email!). For non-ASCII characters (i.e., most of the characters in the world), we don’t talk about case-sensitivity, but instead about case- folding. Although some rather complicated rules for character comparison and decom- position can come into play when using Unicode characters, these case-folding rules (defined by a technology called stringprep, as specified in [RFC 3454]) are typically enforced by a lower-level library, and so most developers don’t need to deal with this directly.


On the XMPP network itself, JabberIDs are provided as raw addresses without a Uni- form Resource Identifier (URI) scheme. This is similar to the convention of telling someone to visit (instead of, or sending an email to the email is hidden list (instead of mailto:email is hidden). However, there is an XMPP URI scheme that can be used to identify JabberIDs as URIs, such as or xmpp:email is hidden (note the lack of a “//”—think mailto, not http). This URI scheme is defined in [RFC 5122]. The XMPP community has also defined ways to include various “commands” in XMPP URIs, such as xmpp:email is hidden?message to send a message (see [XEP-0147] for details).

Streaming XML

XMPP is, in essence, a technology for streaming XML. When you want to start a session with an XMPP server, you open a long-lived TCP connection and then negotiate an XML stream to the server (the server also opens a stream in return, i.e., there is one stream in each direction). We discuss the details of XML streams in Chapter 12, but for now you can think of a stream as an XML document that is built up incrementally over time between your client and your server. Once you have negotiated an XML stream with your server, you and your server can exchange three special XML snippets over the stream: , , and 16 | Chapter 2: Basics of XMPP, . These snippets, called XML stanzas, are the basic units of meaning in XMPP, and once you have negotiated an XML stream you can send an unlimited number of stanzas over the stream. Example 2-1 illustrates a simplified XMPP session, including the interaction between streams and stanzas, as well as the outbound stanzas sent from the client (prefaced with “C:”) and the inbound stanzas delivered from the server (pref- aced with “S:”). Example 2-1. In an XMPP session, the stream element acts as a wrapper for an unlimited number of outbound and inbound XML stanzas; outbound stanzas sent from the client are prefaced with C:, and inbound stanzas delivered from the server are prefaced with S: C: C: C: S: C: Off with his head! S: You are all pardoned. C: C: The XMPP approach of opening a long-lived TCP connection and then asynchronously exchanging an unlimited number of XML snippets differs radically from the traditional approach used in web and email technologies, where you open a TCP connection, complete a transaction (say, retrieving a web page or downloading some email), then close the connection again. These transactional connections do not lend themselves to real-time communication, because the server does not have an “always-on” channel available to push information down to the client. As a result, constant polling for new information is the order of the day (this is true even in recent HTTP techniques such as Ajax and Comet, although the polling has gotten smarter over time). By contrast, in XMPP, the client can send out multiple requests without “blocking” while it waits for Streaming XML | 17, replies, and the server will return those replies dynamically as soon as they are answered. These design decisions have important implications for the XMPP user ex- perience and for the kinds of applications you can build with XMPP. But they also introduce new challenges for developers, who are not necessarily accustomed to think- ing in terms of asynchronous information flows and streaming XML snippets. As you explore XMPP, remember that you may need new tools and a new mindset to see the possibilities. The Layered Look Although XMPP is defined in a number of exhaustive (not to say boring!) specifications, in most cases, you won’t need to worry about lower lev- els, such as XML streams. Instead, existing code libraries typically ab- stract away from the raw XML layer so that you can focus on adding real-time features to your application.

Communication Primitives

These XML “stanzas” sound rather poetic, but what do they mean in practice? In XMPP, a stanza can be thought of as the basic unit of communication, similar to a packet or message in other network protocols (the term was suggested by Lisa Dusseault, who cochaired the IETF’s XMPP Working Group along with Pete Resnick). Several factors determine the meaning of a stanza: • The stanza element name, which is message, presence, or iq. Each kind of stanza is routed differently by servers and handled differently by clients. • The value of the type attribute, which varies depending on the kind of stanza in question. This value further differentiates how each kind of stanza is processed by the recipient. • The child element(s), which define the payload of the stanza. The payload might be presented to a user or processed in some automated fashion as determined by the specification that defines the namespace of the payload. The following sections provide a brief introduction to these factors, and we will explore them throughout this book as we unfold the meaning of various stanza kinds, type attribute values, and payload definitions.


The XMPP stanza is the basic “push” method for getting information from one place to another. Because messages are typically not acknowledged, they are a kind of “fire-and-forget” mechanism for quickly getting information from one place to an- other. Messages are used for IM, groupchat, alerts and notifications, and other such applications. 18 | Chapter 2: Basics of XMPP, Message stanzas come in five flavors, differentiated by the type attribute: normal Messages of type normal are most similar to email messages, since they are single messages to which a response may or may not be forthcoming. chat Messages of type chat are exchanged in a real-time “session” between two entities, such as an instant messaging chat between two friends. groupchat Messages of type groupchat are exchanged in a multi-user chat room, similar to Internet Relay Chat (we discuss groupchat messages in Chapter 7). headline Messages of type headline are used to send alerts and notifications, and a response is not expected at all (a client that receives a headline should not enable a user to reply). error If an error occurs in relation to a previously sent message, the entity that detects the problem will return a message of type error. In addition to the type attribute, message stanzas contain a to and from address, and can contain an id attribute for tracking purposes (we discuss IDs in more detail in relation to IQ stanzas, where they are used more widely). Naturally enough, the to address is the JabberID of the intended recipient, and the from address is the JabberID of the sender. The from address is not provided by the sending client, but instead is stamped by the sender’s server to avoid address spoofing. Messages also contain payload elements. The core XMPP specifications define some very basic payloads, such as and , which are used for person-to- person chat messages. For example, a simple message could look like this: Who are you? Query The order of attributes is insignificant (we usually show the attributes in alphabetical order, but they can appear in any order). Messages (and other kinds of stanzas) can also contain payloads that are not defined in the core XMPP specifications, as we explore throughout this book.


One of the distinctive features of real-time communication systems is presence, which we discuss in Chapter 3. Presence advertises the network availability of other entities, Communication Primitives | 19, and thus enables you to know whether other entities are online and available for com- munication. Many people liken presence to a “dial tone” for the real-time Internet. But this analogy implies that, by itself, presence is fairly boring: who picks up the phone to listen to the dial tone? The exciting thing about presence is that it is a catalyst for communication and collaboration over the Internet, because people are more likely to interact with you if they know you are online. But don’t worry: people can’t see that you’re online unless you authorize them. This authorization is called a presence subscription. In order for someone to see your pres- ence, that person needs to send you a subscription request, which you need to approve. Once you have approved the subscription, that user will automatically receive regular notifications about your network availability. This subscription model implies that the XMPP stanza is in essence a simple, specialized publish-subscribe method, wherein people who subscribe to your presence receive updated presence information when you come online, change your status to “in a meeting” or “at lunch,” and then go offline. At its most basic, presence is an on-off indication that an entity is either online or offline. However, core XMPP presence is often extended by some common states such as “away” and “do not disturb.” These states can be personalized using status messages such as “on a train” or “I’m writing, don’t bother me right now” (a state we have used quite a bit recently). For example: xa down the rabbit hole! In IM applications of XMPP, presence is typically displayed in your roster, which is a kind of presence-enabled contact list. Your roster contains a list of JabberIDs and the state of your presence subscriptions with those entities. When you come online, you announce your presence to your server and it handles the rest—both notifying your contacts that you are online and fetching their current presence for display in your client interface. We delve into these details in Chapter 3.


The Info/Query (or IQ) stanza provides a structure for request-response interactions and simple workflows, similar to the GET, POST, and PUT methods that you may be familiar with from HTTP. Unlike the stanza, an IQ stanza can include only one payload, which defines the request to be processed or action to be taken by the recipient. In addition, the entity that sends an IQ stanza must always receive a reply (usually generated by the intended recipient or the recipient’s server). Requests and responses are tracked using the id attribute, which is generated by the requesting entity and then included by the responding entity. Finally, the type attribute has special values for IQ stanzas: 20 | Chapter 2: Basics of XMPP, get The requesting entity asks for information, such as requirements for registering an account (similar to HTTP GET). set The requesting entity provides some information or makes a request (similar to HTTP POST or PUT). result The responding entity returns the result of a get operation (such as the information that an entity must provide to register an account), or acknowledges a set request (similar to an HTTP 200 status code). error The responding entity or an intermediate entity, such as an XMPP server, notifies the requesting entity that it was unable to process the get or set request (e.g., item because the request is malformed, the requesting entity does not have permission to perform the operation, etc.). The early use of HTTP-style numeric error codes has been superseded by XML elements for extensible error conditions. IQ or Message? XMPP message stanzas provide a “fire-and-forget” transport that is best used for human-readable text, alerts, notifications, and whenever you don’t need assurance that the content was truly delivered. IQ stanzas provide a more reliable transport that is optimized for a structured ex- change of data, typically non-human-readable data. (Although the Ad- vanced Message Processing extension defined in [XEP-0079] provides mechanisms that can make message stanzas a more reliable transport, it is not yet widely implemented or deployed.) Using the values of the IQ stanza’s type, we can generate a fairly structured IQ inter- action between two entities, as shown in Figure 2-5. To illustrate this kind of interaction in more detail, consider the process for getting your roster and then updating it: Communication Primitives | 21, Figure 2-5. The IQ stanza provides structured interactions between entities

Here Alice has asked her server (wonderland.lit) to give her the contact list she stores

on the server by sending an IQ-get containing an empty payload qualified by the jabber:iq:roster namespace. The server then replies with a non-empty payload quali- fied by that namespace, in this case, containing a single element for each contact in her roster. (The client knows that this roster result relates to its initial request because the server includes an id attribute with a value of rr82a1z7.)

Alice can also use an IQ-set to add a new contact to her roster:

Her server then simply acknowledges the roster update by returning an empty IQ-result:

22 | Chapter 2: Basics of XMPP, As you can see from these roster examples, the payload of an IQ-get or IQ-set is always defined by its own format qualified by a particular XML namespace, as specified in one of the many XMPP protocol documents. You can think of each payload format as a command to be processed by the recipient. An IQ-get requests a particular kind of information, such as a registration form, configuration data, service discovery infor- mation, or a contact list. An IQ-set creates, updates, or deletes a particular kind of information, such as a completed form, updated configuration data, or an addition to a contact list. Throughout the following chapters, we explore these IQ interactions in great detail as we describe how particular XMPP protocol extensions work.


An XML stanza can contain any number of other child elements, including XHTML- formatted message bodies, pointers to URLs, RSS or Atom notifications, forms to be filled out (or submitted forms), XML-RPC or SOAP data for web services, geographical locations, and a wide range of other payloads. (The “X” in XML and XMPP stands for “extensible,” so payload types are limited only by your imagination!) Because XMPP is a pure XML technology, it makes extensive use of XML namespaces as a way to “scope” stanza payloads. You can think of these namespaces as the XML equivalent of packages and namespaces in programming. So far, the XMPP developer community has defined dozens of extensions to the core XMPP stanza layer. Most often these extensions are published by the XMPP Standards Foundation at http://xmpp .org/, but you can also define your own private extensions for custom features. Extensions are matched on both the element name and the namespace. In the early days of the Jabber open source projects, developers used an element for extensions that would be placed in message or presence stanzas and a element for extensions that would be placed in IQ stanzas. You will see examples of these early extensions in this book (e.g., ), but it is important to realize that this usage was merely conventional; extensions developed later do not follow the same practice. What’s in a Name? You’ll notice many different types of XML namespaces. In general, the earliest namespaces were things like jabber:iq:roster, namespaces de- fined from 2001 to 2005 were like, and more recent namespaces are things like urn:xmpp:jingle. Don’t worry about the differences; they are all just namespace names. Communication Primitives | 23,


In XMPP, you exchange stanzas asynchronously with other entities on the network. This model is different from HTTP, where your client sends a request to a server and then waits for a reply before it makes another request. By contrast, in XMPP your client can “pipeline” requests to your server or to other entities and then receive replies as they come back. Certain events can also trigger information that is pushed to your client (e.g., when one of your devices adds an item to your roster, the item is pushed out to all of your devices so that they stay in sync). This rapid-fire, event-driven approach can be confusing at first to developers who are more accustomed to traditional web devel- opment, but it has a number of advantages, such as real-time notifications and the ability to work around the need to continually poll for updated information.

Error Handling

Unlike some communication technologies, XMPP does not acknowledge every packet or message that is sent over the wire. Typically, you assume that a message or presence stanza has been delivered if you don’t receive an error. IQ stanzas are more structured: you must always receive either an IQ-result or an IQ-error in response to an IQ-get or an IQ-set. Errors are reported by setting the stanza’s type attribute to a value of error, along with an child element that is qualified by the urn:ietf:params:xml:ns:xmpp-stanzas namespace. Here is an example: The type attribute of the element is one of auth, cancel, continue, modify, or wait (the values hint at how to handle the error). However, the primary meaning of the error is specified by the native child element, for example, or . XMPP error conditions generally follow the model of errors from HTTP and SMTP, except that they are structured not as numeric codes, such as 404, but as (you guessed it!) XML elements. A full list of stanza error conditions can be found in [RFC 3920]. Know the Code In the early days of the Jabber community, errors were specified with HTTP-style error codes, such as 404 and 501. When the Jabber proto- cols were standardized at the IETF, the error syntax was expanded to provide a more flexible data format. However, you may still see the HTTP-style error codes on the wire. For a full mapping between the older codes and newer conditions, refer to [XEP-0086]. 24 | Chapter 2: Basics of XMPP, The element can also include application-specific child elements that specify further details about the error condition. For example, the following error stanza in- dicates that a pubsub subscription request (which we discuss in Chapter 8) has failed because the pubsub node is closed to new subscriptions: In addition to the stanza error conditions described here, [RFC 3920] also defines stream error conditions as well as errors related to SASL authentication. The main difference is that stream errors are unrecoverable and result in closing the XML stream, whereas stanza errors are recoverable and therefore are used for reporting problems with particular stanzas (without termination of the underlying stream). We don’t show a lot of error flows in this book, because it would make the text twice as long. If you want all the details about errors that can result in a particular use case, refer to the relevant XMPP specification, which usually will show all the error cases in addition to the “happy path.”

Hello Hello World World: Building a Basic XMPP Application

Believe it or not, at this point you have enough technical baggage to go off and start implementing your own XMPP application. In fact, we’ll prove it to you: in this section, we will implement a simple XMPP service, using only the basic building blocks of XMPP introduced in this chapter. The task of our service is simple: reply to every incoming message with an identical message. You can see our service in action in Figure 2-6. Our service acts as a regular contact for its users, but automatically echoes back every mes- sage that it receives. Unmanned contacts like this service are typically called bots. Hello Hello World World: Building a Basic XMPP Application | 25, Figure 2-6. The “echo” service in action; “echo bot” acts as an ordinary IM contact, but automatically echoes back every message you send it

To implement the bot we just described, we chose Python as our implementation lan-

guage, and delegate the actual XMPP protocol details to SleekXMPP, one of the many available XMPP libraries (a few of which are listed in Appendix B). We now walk through the steps that lead to the implementation of this “echo” service, displayed in

Example 2-2.

Example 2-2. Implementation of a basic bot that echoes all incoming messages back to its sender def main() : bot = EchoBot("email is hidden/HelloWorld", "mypass") class EchoBot : def _init_(self, jid, password) : self.xmpp = sleekxmpp.ClientXMPP(jid, password) self.xmpp.add_event_handler("session_start", self.handleXMPPConnected) self.xmpp.add_event_handler("message", self.handleIncomingMessage) def run(self) : self.xmpp.connect() self.xmpp.process(threaded=False) def handleXMPPConnected(self, event): self.xmpp.sendPresence(pstatus = "Send me a message") def handleIncomingMessage(self, message) : self.xmpp.sendMessage(message["jid"], message["message"]) 26 | Chapter 2: Basics of XMPP, The first step in the implementation of our service is to make sure that the echo bot is available on the XMPP network. In order to do this, the service connects to an XMPP server under a given username, just like one would connect with an ordinary IM client. Our bot happens to be registered as echobot with the wonderland.lit server. Since there will be only one instance of our bot running, we pick HelloWorld as an arbitrary resource name to identify the instance of the bot. Putting all these pieces together, we get email is hidden/HelloWorld as the JID with which our bot connects to the server (and through which our service will be reachable). Connecting to the server is done by initializing a ClientXMPP object from the SleekXMPP library, and calling connect() to set up the connection to the server. The subsequent call to process() starts the event loop of the bot. The event loop is a seemingly infinite loop that waits for XMPP events to occur (incoming messages, notifications about connection errors, etc.); whenever an event occurs, the event loop calls the event han- dler method that is associated with the event. For reasons explained in the following paragraphs, our bot registers event handlers for two types of events: session_start and message. Once the bot is connected to the server, it needs to announce that it’s available for service. This is why the bot registers with the XMPP library to receive notification of the session_start event, which will fire when the bot is connected and the XMPP ses- sion has started. Upon the beginning of a session, the bot sends out basic availability presence by calling sendPresence(). As a result, every user that is subscribed to the bot’s presence will see the bot appear in his roster. The core functionality of our bot is triggered whenever a message is received. The handler of the message event extracts the sender and the body of the incoming message, and sends it back to the originator using sendMessage(). That’s all there is to it! With only a handful of lines of code, we created a fully functional XMPP service. We showed that XMPP applications are typically event-driven, imple- menting their functionality in terms of asynchronous events that occur. Of course, this book wouldn’t be over 250 pages long if all you could do with XMPP was read back simple messages to users. That’s why we develop a larger XMPP application in Chap- ter 14, showing more aspects of implementing XMPP applications.


In this chapter, we outlined the architecture, addressing, underlying data transport, and communication primitives of XMPP. As a result, you might already have some ideas about how to XMPP-enable an existing application (for example, you might send an XMPP stanza when someone checks a file into a source control system, or you might add presence indicators to a website directory). Summary | 27, However, we have only begun to scratch the surface of what XMPP can do. In the remainder of this book, our explorations will proceed in two directions: • We will move “up the stack” by describing many of the core XMPP extensions that enable more specific functionality, such as the extensions for service discovery (Chapter 5), multi-party messaging (Chapter 7), publish-subscribe (Chapter 8), and multimedia session management (Chapter 9). • We will move “down the stack” by describing some of the ins and outs of session establishment (including authentication, channel encryption, the HTTP binding, and serverless messaging over a local network). You may not need to know about these details to build your application, because typical XMPP libraries provide a “login” or “connect” function. However, these lower-level options give you addi- tional tools that can prove extremely useful in more advanced XMPP applications, so we discuss them in Chapter 12. Now let’s look at the key tools in the XMPP toolkit. 28 | Chapter 2: Basics of XMPP, PART II The XMPP Toolkit,

CHAPTER 3 Presence Is Anybody Home?

Imagine that you want to contact a friend or colleague. In the old days, you might have sent the person a letter (you know, one of those pieces of paper that is delivered to your home or office), and then waited for a reply. Or you might have phoned the person and hoped she was around (if not, you would have left a message with a person or an automated system, perhaps playing “phone tag” for a few days). Or, more recently, you might have sent an email, perhaps receiving a reply in 10 minutes or so, but perhaps days later. In XMPP, you can know when a contact of yours is online and available for commu- nication, using a technology called presence. So instead of waiting and wondering, or just getting lucky, your Jabber client will show you the network availability of your contacts, usually with an indicator such as a light bulb icon (on the theory that if some- one is home, the lights will be on). Figure 3-1 shows an example of such a presence- enabled contact list in an IM client. However, presence is not limited to pretty little icons; it enables you to get real work done. In this chapter, we delve more deeply into presence, and explore how you can use it to build smarter, more interactive applications.

Authorization Required: The Subscription Handshake

One of the reasons why presence provides a key tool for application development is that it’s voluntary. No one is forcing you to share information about your network availability with anyone else. But if you do choose to share that information, you have made a trust decision, which separates someone who can see your presence from everyone else on the network. As we shall see, this distinction provides important ben- efits to XMPP-based systems., Figure 3-1. Your roster provides a visual representation of the people you know and trust on the network, including information about their network availability or “presence” The trust or access decision behind presence happens naturally in IM systems, because people you approve are automatically added to your contact list (called a roster in XMPP), which is typically the “home base” for any instant messaging or real-time communications application. In addition, presence access is usually bidirectional: you allow a contact to see your presence, and your contact allows you to see his presence. This happens through a subscription “handshake,” as shown in Figure 3-2. If the handshake is completed suc- cessfully, the result is a bidirectional presence subscription between the two parties. (XMPP servers also add the contact to the user’s roster and add the user to the contact’s roster during this process, plus manage a state machine of subscription states, but we don’t need to worry about those details here; refer to [RFC 3921] for a full description.) Let’s see how the subscription handshake works in practice. To request someone’s presence, you send him a subscription request, which is a stanza of type subscribe: When the intended recipient receives your presence subscription request, he can either approve it (via a stanza of type subscribed) or deny it (via a stanza of type unsubscribed): As you might imagine, to create a bidirectional presence subscription, the person who approved the original subscription request needs to send a subscription request of his own: 32 | Chapter 3: Presence, Figure 3-2. A bidirectional subscription handshake: after the contact subscribed to the user’s presence, the user in turn subscribes to the contact’s presence Typically, your client will auto-reply at this point, rather than asking you to manually approve the reverse request: Once you are subscribed to another person’s presence, you will automatically be no- tified when the other party’s network or communications availability changes. This presence notification takes the form of a stanza with no type attribute (i.e., implicitly indicating availability): xa down the rabbit hole! The next section describes how that happens.

How Presence Is Propagated

Now that you and your contact are subscribed to each other, how does presence in- formation flow between the two of you? Here is a brief overview: 1. You negotiate an XML stream with your server (see Chapter 12). 2. You send an initial presence stanza to your server: How Presence Is Propagated | 33, Yes, this is the smallest XMPP stanza you will ever see! Initial presence can also include more detailed availability status information, as described next. 3. Your server checks your roster and sends a presence notification to each person who is subscribed to you, making sure to add the full JabberID of your connected resource as the from address: [etc.] 4. Now, everyone who is subscribed to your presence knows that you are online and available for communication. But how do you know if they are online? Here, your server once again comes to the rescue, because it sends a presence probe to everyone you’re subscribed to: [etc.] 5. Once your contacts’ servers receive the probes, they check permissions according to their records. If you are allowed to see your contacts’ presence, you will receive at least one presence notification from each of your contacts who is online, and often a notification if your contact is offline, including information about when her last presence notification was sent: [etc.] 34 | Chapter 3: Presence, The element is added by the contact’s server, and the UTC timestamp is the time when the presence stanza was sent by the contact (in this case, when the contact went offline). Note that you might receive more than one presence stanza, because any given contact might have multiple connected resources. (Look closely at the presence notifications received from the Mad Hatter in the previous example.) Do You Always Receive Unavailable Presence? Some server implementations do not return an unavailable presence notification in response to a presence probe; instead, they simply ignore the presence probe, on the theory that if the probing entity does not receive any presence notification, it will assume that the probed entity is not online.

Availability Status

So far, our examples of presence notification stanzas have been extremely simple (either available or unavailable). But a presence stanza can contain more information than basic on-off network availability. There are two primary presence elements that express more detailed information: the element and the element. The element is limited to four predefined values, which provide insights into a human user’s availability for and interest in communication (these are not shown directly to end users but are used to provide availability hints in a user interface): chat Announces that you are available for, and actively seeking, conversation (perhaps you’re feeling especially sociable). away Indicates that you are gone from your IM client, computer, or device for a short period of time; this state is often triggered without human intervention through a feature known as auto-away, commonly found in many IM clients. xa Indicates that you are gone for a longer period of time (xa is shorthand for “eX- tended Away”); your IM client can also automatically generate this state. dnd Announces that you are busy and don’t want to be interrupted right now (dnd is shorthand for “do not disturb”). Furthermore, the element enables a user to specify some free-form, human- readable text that describes the user’s availability in more detail. For example, a user might combine a value of away with a value of “Having tea with the White Rabbit,” or a value of dnd with a value of “On a deadline.” Availability Status | 35, The and elements are not limited to human users. They could also be used by automated processes; for example, a particular unit in a computing farm could be dnd if it cannot accept any new jobs at the moment. However, for more so- phisticated handling of presence from automated entities, it would probably be pref- erable to define a custom presence extension rather than overloading existing text values. Typically, these elements are used to send updated availability information during the life of a user’s presence session, as in the following example: away Having a spot of tea Just as with the initial presence notification, subsequent presence updates are also broadcast by the sender’s server to everyone who is subscribed to the user’s presence. If the presence subscription is bidirectional, the user’s server will often send the sub- sequent notifications only to contacts who are online. This optimization helps to reduce traffic, since presence uses a great deal of bandwidth in a real-time communications system.

Presence Priorities

The presence stanza can include one more optional element: . Unlike many other IM systems, XMPP allows you to connect multiple devices or clients to the same account at the same time. This introduces interesting possibilities for inter-device com- munication (e.g., you could control a set-top box at home from your computer at the office). However, it also introduces the need to differentiate between those devices. For addressing purposes, this is done through the resource portion of a JabberID, such as email is hidden/TV as opposed to email is hidden/office. (XMPP developers usually refer to a JID of the form email is hidden as a bare JID and a JID of the form email is hidden/resource as a full JID.) For presence purposes, each connected re- source can specify a priority, in the range from –127 to +128. A higher-priority resource is more likely to receive a message sent to the account’s bare JID. A resource with a negative priority will never receive such a message (although it will receive a message sent directly to that resource). The latter is useful for network-enabling a device that doesn’t intercept human-oriented chat messages. 7 -1 36 | Chapter 3: Presence,

Directed Presence

The presence notifications we’ve looked at so far have been broadcast—that is, they are sent to everyone in your roster (you indicate that you want a presence notification to be broadcast by leaving off the to address). But what if you want to send presence to someone who is not in your roster? Perhaps you want to chat with someone for a little while but don’t want to add that person to your roster and therefore share presence on a permanent basis. In this case, you can send directed presence to the other person, i.e., presence that has a to address. Consider what happens when Alice goes down the rabbit hole and meets the White Rabbit. Because the rabbit isn’t in her roster, she sends a message but also sends directed presence: If you please, sir- The White Rabbit is too frightened to reply, but his IM client at least sends directed presence back to Alice: This kind of temporary presence sharing without a long-term subscription is a best practice for brief interactions over the network. And as we’ll see in Chapter 4, directed presence is also used to join and leave multi-user chat rooms (another form of tempo- rary interaction).

Going Offline

When you’re tired of real-time interactions or just need to disconnect from the network, you can easily go offline by telling your server that you are now unavailable: There is no presence type of available, because presence implicitly describes net- work availability (i.e., if there is no type attribute, then the entity is assumed to be available). Going offline has several implications: • Your server broadcasts your unavailable notification to everyone in your roster. Going Offline | 37, • Your server also broadcasts your unavailable notification to all the entities to which you’ve sent directed presence (see the earlier example of the White Rabbit). • If you have no other online resources, when your contacts’ servers receive the un- available notification, they will probably stop sending presence notifications to you. • If you have no other online resources, your server will stop sending presence sub- scription requests to you, instead storing them up for delivery the next time you are online. • If you have no other online resources, your server will stop sending messages to you, instead storing them up for delivery the next time you are online (we describe these “offline messages” more fully in Chapter 4). Naturally, if you have other online resources, your server will continue to send messages and presence subscription requests to you, but the presence session for your newly offline resource is now over.

Rich Presence

The presence stanza provides a convenient and relatively efficient method for publish- ing information about a user to interested others. In the past, enterprising developers of XMPP clients have used this presence “transport” to push out information about much more than network availability. For example, why not use presence stanzas to advertise what music you’re listening to? Pink Floyd - Dogs This kind of information is typically called rich presence or extended presence, and can include a very wide range of transient data: your current mood or activity, the music you listen to, the videos you watch, the chat rooms or web pages you visit, the games you play, your physical location, etc. There are several problems with putting all these payloads inside presence stanzas: • It’s not very XML-friendly to put all of this information into the element as an unstructured text string. • Sending all of this information in presence will result in a lot more presence stanzas, and those stanzas will probably be bigger (perhaps much bigger) than existing presence stanzas. Given that presence already uses far more bandwidth in XMPP than messaging does, piling on more and bigger presence stanzas could seriously degrade network performance. • Not everyone in your roster will be interested in things like the music you listen to, so why send them that information? 38 | Chapter 3: Presence, • You might want to restrict who can know your physical location or other sensitive information to a special sub-group of those who can know your network availa- bility. Publish-Subscribe [XEP-0060] and the Personal Eventing Protocol [XEP-0163] provide this kind of access control, but basic presence is broadcast to everyone in your roster. Because of considerations like these, rich presence is typically not sent via the presence transport, but instead uses a specialized publish-subscribe method that we describe in Chapter 8.

Presence and Rosters

Figure 3-1 shows that presence information is usually displayed as one aspect of a user’s roster. By retrieving the user’s roster when the user logs in, an IM client is able to integrate the presence data it receives into a useful, familiar interface. Your roster is managed by your client, but it is stored on your “home” server. This enables you to connect from anywhere and still retrieve your contact list, which your client typically does when you start your session by sending an IQ-get to the server: The server ignores the from address on the roster request, because it always delivers the roster to the entity that requested it (for security reasons, you can’t request someone else’s roster). The to address on the roster request is the bare JID of the user. This means that the server handles the request on behalf of the user’s account. Equivalently, the sender could include no to address at all, since no to address is treated the same as a to address of the sending user. ([RFC 3921] recommends including no to address, and [rfc3921bis] recommends including it, but the result is the same.) The user’s server then retrieves the user’s roster from a server-side data storage mech- anism and returns it to the resource that made the request: Presence and Rosters | 39,

Each item in your roster has a JabberID associated with it, which acts as the “key” to

storing and identifying the item. Each item has a particular presence subscription state that reflects the presence authorizations we’ve already looked at, and can also specify a user-friendly name. Therefore, the data returned by the server will in fact be a little more complete:

The roster can be used not only to store a flat list of contacts, but also to group into

various categories. Because these roster groups are not exclusive (a bot could also be a friend—though that would be rather sad), it’s better to think of them as flexible tags instead of exclusive buckets. Roster groups become increasingly important in organ- izing your contact list as you befriend more and more people on the XMPP network (the average roster size for us is 800 and rising!). Therefore, the data returned by the server will in fact be even more complete: Wonderlanders Wonderlanders 40 | Chapter 3: Presence, Wonderlanders Nobility Family

Roster groups can be edited through the roster management protocol, via the

element: Bots

When one of your connected clients modifies the roster (e.g., by adding a new contact

or changing the group for a particular item), the server then pushes that change to all of your connected clients by sending an IQ-set containing only that item to each re- source. This IQ-set, called a roster push, enables all of the resources to remain synchronized: Bots Presence and Rosters | 41, Getting Pushy Because the server always pushes a roster change to a connected client, the client can simply wait for and process roster pushes. This is easier than closely monitoring incoming presence stanzas of type subscribed, unsubscribe, and unsubscribed. It also removes the need to perform a roster set before sending the presence subscription request (if the con- tact doesn’t exist or denies the request, then you would need to remove the roster item, whereas the server will automatically perform this cleanup). Thus the use of roster pushes helps to simplify the task of writing an XMPP client; this is consistent with the early Jabber philos- ophy of “simple clients, complex servers,” as defined in the Protocol Design Guidelines [XEP-0134] and described further in Chapter 12. Finally, in enterprise and school settings, it’s not uncommon for parts of your contact list to be centrally managed by appropriate IT staff, in which case, they might be pulled straight out of an LDAP database or other backend storage location (e.g., when Bob in the QA department is fired or when Alice drops the symbolic logic course, they will automatically disappear from the relevant contact list group). However, methods for doing so are not yet part of the standard roster functionality and tend to be specific to particular XMPP server implementations, so talk with the developers of your favorite open source project or commercial product.

Using Presence

In this chapter, we looked at several core features of XMPP: presence subscriptions (which are typically bidirectional); presence notifications, such as available, away, and unavailable; and rosters, including roster groups. These features provide a few building blocks for XMPP applications, so let’s look at how they are used higher up in the XMPP protocol stack, and how you might be able to use them, too.

Presence-Based Routing

Once you know that someone or something is online, and perhaps know the priori- ties of its resources, you can make some decisions about whether and how to deliver information to that entity. Some of these decisions are made by the server that handles traffic on behalf of an account. So, to use our example of a set-top box and an office computer, a message sent to email is hidden typically would be delivered to the office resource but not the TV resource (although some XMPP servers can be configured to deliver the message to all resources, or at least to all resources with non-negative priority). 42 | Chapter 3: Presence, However, the sender can also make messaging decisions based on presence data. Con- sider a workflow application, in which five different people have authority to approve a given purchase order. If two of them are offline, one of them has only a resource with negative priority, one of them has a resource with positive priority that is dnd, and the last one has a resource with positive priority and no value, it might make sense to send the approval form to both of the positive resources but not to the others. Such presence-based messaging determinations can expedite decision-making and applica- tion processing. (Often such addressing choices will also incorporate information about the capabilities of each resource, as we discuss in Chapter 5.)

Access Control

In our discussion of rich presence, we observed that basic presence is an all-or-nothing affair (either someone receives your presence information or they don’t). However, sometimes you want more granular control over who can chat with you or who can receive certain kinds of information (e.g., your geolocation). All of these matters es- sentially boil down to access control. Although basic presence publishing and roster data do not by themselves provide such access control methods, they can be used to build such methods. In particular, the communications-blocking techniques discussed in Chapter 4 makes use of information about both presence subscriptions and roster groups to make access control decisions. The same is true of the Personal Eventing Protocol, or PEP [XEP-0163], which is used to publish rich presence data, as described in Chapter 8.

Presence As a Transport

XMPP developers try to resist the temptation to push out arbitrary information in presence stanzas. As mentioned earlier, even though presence stanzas tend to be small, there are a lot of them, and so presence tends to be the most bandwidth-intensive aspect of XMPP (much more so than messaging). Therefore, it is best to keep presence stanzas small and send them only when the client generates information that is relevant to communication. Terms like “small” and “relevant” are, unfortunately, vague. Couldn’t the tune I’m listening to be relevant to communication (e.g., in an online music service)? Maybe. But XMPP developers tend to be conservative about the payloads they include in a presence stanza. One exception is capabilities information—data about what XMPP features and ex- tensions a device supports. Because there are a lot of XMPP features and extensions, and because users want to know when, for example, a friend has plugged in a video camera and can now engage in a video chat, it’s helpful to dynamically publish capa- bilities data in presence. We discuss this usage in depth in Chapter 5. Using Presence | 43,


Presence lies at the core of many uses of XMPP. Indeed, the fact that your XMPP server knows when you are online (and with which devices) is what makes real-time com- munication possible, whether it is instant messaging, multi-user chat, just-in-time no- tifications, or voice and video chat. As we’ve seen, your presence is broadcast only to other people or entities you have authorized through subscription requests that you approve. These subscription states are stored on the server in your roster, a presence- enabled contact list that is the central focus of instant messaging applications. This chapter also provided an overview of specific availability states, presence priorities, directed presence, and how presence is propagated on the network. In upcoming chap- ters, we will see how presence is used to interact with XMPP chat rooms; dynamically advertise device capabilities; and enable advanced personal eventing and “lifestream- ing” applications, such as microblogging, location sharing, and social music services. 44 | Chapter 3: Presence,

CHAPTER 4 Instant Messaging I Think, Therefore IM

The initial goal of the early Jabber project, well before the protocol was named XMPP, was to create an open Instant Messaging (IM) platform. Although IM is often thought of as person-to-person chat, at its core it really provides the ability to quickly route messages from one place to another over the network (no matter who or what the intended recipient is). For this reason, XMPP servers are optimized for handling large numbers of relatively small messages with very little latency. When you are exchanging instant messages, you don’t want to experience any delivery delays (which can be al- most as annoying in IM as they are on the phone). In XMPP, messages are delivered as fast as possible over the network. Let’s say that Alice sends a message from her new account on the wonderland.lit server to her sister on the realworld.lit server. Her client effectively “uploads” the message to wonderland.lit by pushing a message stanza over a client-to-server XML stream. The wonderland.lit server then stamps a from address on the stanza and checks the to ad- dress in order to see how the stanza needs to be handled (without performing any deep packet inspection or XML parsing, since that would eat into the delivery time). Seeing that the message stanza is bound for the realworld.lit server, the wonderland.lit server then immediately routes the message to realworld.lit over a server-to-server XML stream (with no intermediate hops). Upon receiving the message stanza, the realworld.lit server checks to see whether Alice’s sister is online; if so, the server immediately delivers the message to one or more of her online devices over a server-to- client XML stream (without storing it or otherwise performing much processing on it). As a result, the message is delivered very quickly from Alice to her sister. These design decisions have important implications. First and foremost, the clients and servers need to be event-driven and ready to take appropriate action whenever they receive an incoming stanza. XMPP servers don’t have the luxury of storing a message and waiting for a client to poll for it; instead, they deliver the message as soon as they receive it. Second, all entities (but especially the servers) need to be presence-aware, since it is the concept of being online that makes rapid delivery possible in the crucial, “last mile” between the recipient’s server and the recipient’s device(s). Third, fast and accurate handling of DNS lookups, domain name resolution, long-lived TCP connec- tions, connectivity outages, and network congestion is critical to the success of the overall system. Several types of XMPP messages exist, fundamentally differentiated by the value of the type attribute: normal This message type is delivered immediately or stored offline by the server, and handled by the client as a “standalone” message outside of any chat or groupchat session. This is the default message type. chat Messages of type chat are sent within a burst of messages called a “chat session,” usually over a relatively short period of time. Instant messaging clients show such messages in a one-to-one conversation interface for the two parties. groupchat XMPP servers usually route messages of type groupchat to a specialized component or module that hosts multi-user chat rooms, and this component then generates one outbound message for each of the room occupants. (We discuss groupchat messages in Chapter 7.) headline Headline messages usually are not stored offline, because they are temporal in nature. In addition, XMPP servers often send a message of type headline to all of the online devices associated with an account (at least those with non-negative values). error A message of type error is sent in response to a previously sent message, to indicate that a problem occurred in relation to the earlier message (the recipient does not exist, message delivery is not possible at the moment, etc.). Both chat and normal messages are usually handled by the recipient’s server in a par- ticular way: if the message is addressed to the bare JID (email is hidden) of the account, the server immediately delivers the message to the highest-priority resource currently associated with the account. If there is only one online resource, this decision is easy, but if there are multiple online resources, the recipient’s server delivers the message to the resource with the largest value for its presence priority. For example, a resource with a presence priority of 7 will receive messages addressed to the bare JID, but another resource with a presence priority of 3 will not. (Resources with negative priority will never receive a message sent to the bare JID, but all resources will receive a message addressed to the full JID of that resource.) Finally, although XMPP technologies put a premium on near real-time data delivery, almost all XMPP servers include support for “offline messages” if the intended recipient is not online when the server receives a normal or chat message addressed to that 46 | Chapter 4: Instant Messaging, JabberID. These messages are automatically pushed to the recipient’s client when the user next logs in. When the recipient’s server pushes out the offline message, it also adds a small extension noting when the message was originally received, using the protocol extension defined in Delayed Delivery [XEP-0203]. This enables the recipient’s client to properly order the messages it receives in a user interface.

Chat Sessions

When two people “IM” with each other, the conversation usually happens in a burst of messages over a short period of time. This pattern mimics real life, where you might chat with someone for 5 or 10 minutes when you meet them on the street or talk on the phone, but not chat with them again for a week or two. In XMPP, we call this kind of burst a chat session, and you can see an example of such a session in Figure 4-1. Figure 4-1. A chat session consists of a “burst” of messages sent over a short period of time XMPP chat sessions are not formally negotiated but proceed naturally. The entity that initiates the conversation sends a message to the bare JID of the responder, and this message is stamped by the initiator’s server with the full JID of the initiator. When the responder sends a reply, it too is stamped by the recipient’s server with the full JID of the responder. At this point, the initiator knows the responder’s full JID and the res- ponder knows the initiator’s full JID, so the parties have “locked in” to each other’s XMPP resource identifiers. Each party now addresses stanzas to the full JID of the other Chat Sessions | 47, party when sending subsequent messages, until and unless receiving a presence change from the other party (which might trigger resending a message to the bare JID). The features we discuss in the following sections all relate in one way or another to instant messaging, and to chat sessions in particular: chat state notifications tell you whether your conversation partner is actively engaged; XHTML lets you add a bit of dash and style to your messages; vCards enable you to learn something about the people you chat with; and blocking and filtering help you avoid unpleasant conversations with some of the unsavory characters you might meet online.

Are You There? Chat State Notifications

Consider the following IM conversation between you and your nine-year-old daughter: You: Hi honey! She: Hi You: How was school today? She: Great This is the moment where she starts typing about all the great and exciting things she learned about. Unfortunately, her typing skills aren’t at the 80 words per minute you’re hitting. While she’s composing her answer, you assume that she’s not in the mood to talk about school right now, so you continue the conversation: You: Did you visit grandma this afternoon? What did she tell you? Now you’re waiting for an answer. In the meantime, your daughter has been typing away about her day at school. After a while, she decides to pause composing her answer to your first question, and looks up from the keyboard she’d been concentrating on for the past few minutes. She now sees that you have already moved on from the previous question, so she’s left with the choice to delete everything she wrote so far, send half of the answer she wanted to send and move on, or just continue typing (thus slowing the conversation down even further). She bites the bullet, deletes everything she wrote so far, and moves on: She: Yes, I did Now, you’re waiting for the second part of the answer as to what grandma told her. After waiting for two minutes, you wonder whether she’s just typing slowly, or she just missed the fact that you asked a second question. So, just to be sure, you repeat the question: You: And? It turns out that she started writing the answer, but suddenly had to go downstairs to answer the phone. So, she comes back, and finishes the answer: She: Everything was fine. I have to go do my homework now. 48 | Chapter 4: Instant Messaging, Since the answer wasn’t coming immediately, you decided to do something else while waiting for it. When switching back, you notice that your daughter wants to finish the conversation, so you’ll have to say goodbye. Or, wait, maybe she finished it already, and started doing her homework, in which case you don’t really want to distract her. The problem with this (fairly common) scenario is that neither of you know anything about the other person’s activity level with regard to the conversation. The exact same conversation over the phone would have been a lot less awkward: it would have been easy to tell whether the other person was answering your question or not, and the sound of a dial tone would leave no doubt that the conversation was actually finished. In order to avoid the inconvenient situations like the preceding conversation, you need the no- tion of chat states in your IM system, as defined in Chat State Notifications [XEP-0085]. Chat states describe your involvement with a conversation, which can be one of the following: Starting Someone started a conversation, but you haven’t joined in yet. Active You are actively involved in the conversation. You’re currently not composing any message, but you are paying close attention. Composing You are actively composing a message. Paused You started composing a message, but stopped composing for some reason. Inactive You haven’t contributed to the conversation for some period of time. Gone Your involvement with the conversation has effectively ended (e.g., you have closed the chat window). During the conversation, your chat state will most likely change: after composing a message while in the composing state, you will become active while waiting for a reply to your message. However, it does not always make sense to go from one specific state to another one. For example, from composing a message, you can’t really become in- active for a long period without pausing for at least a short time. Figure 4-2 shows the possible transitions between chat states. Are You There? Chat State Notifications | 49, Figure 4-2. The transitions between chat states are well defined

Changing state in a conversation is done by embedding the corresponding chat state

element into a message stanza. For example, the mother-daughter conversation would start off like this: Hi honey!

By adding the element to your message, you indicate that you are actively

engaged with the conversation. Your daughter starts typing her response, so her client sends you a chat state update by adding a element to an empty message:

Shortly after the notification, the actual message comes in, making her an active par-

ticipant of the conversation again: Hi

The conversation goes on for a while, up to the point where you ask her about grandma:

Did you visit grandma this afternoon? What did she tell you?

This is where she suddenly stops typing to go answer the phone, and so after a few

seconds, her client notifies you of that fact by sending you a notification:

After a while, she resumes her answer:

Finally, skipping to the end of the conversation, she sends her goodbye and closes

her chat window: Everything was fine. I have to go do my homework now. s

The person you are communicating with may not always be interested in receiving

notifications about your chat state. For example, when she is using her mobile phone for IM, she would rather save on the usage of the limited network capacity, at the price of not being able to see when you are typing. In order to discover whether the other party is interested in your chat state, you start the conversation as usual, by adding an element to your message. If the reply comes back without any chat state information, you have to assume that the other person either does not know how to handle chat state updates, or does not want to receive them. From then on, you both continue the conversation, without adding any chat state information to your subse- quent messages. (Naturally, if you know that the other party does not support the chat Are You There? Chat State Notifications | 51, states protocol, you would leave off the notifications entirely. We talk about ways to discover support for various protocol extensions in Chapter 5.) Another reason why you may not want to send chat state notifications is privacy. You may not want other people to know when you are physically using your IM client (information that chat state notifications would reveal). However, it does not always have to be as drastic as disabling all types of notifications. You could configure your client to send only basic chat state information (i.e., whether you are active or com- posing), and not send any information about more fine-grained states, such as paused, inactive, or gone. This basic information would only reveal whether you started composing an answer or not, and leave out any hints to whether you physically went away from your IM client, or reconsidered talking and closed the conversation. So far, we have talked about chat state notifications only in the context of one-to-one conversations. To a certain degree, chat state notifications can be useful inside multi- party chats as well (we talk about groupchat in Chapter 7). However, note that if the number of participants starts growing, the total number of notifications sent will in- crease drastically as well.

Looks Matter: Formatted Messages

Some folks think plain-text messages are boring. For example, let’s say you are really excited about a new movie that you just watched, so you send a message to your friend: You: I love this movie I saw last night, it’s awesome! If you said that over the phone or in person, you’d probably emphasize some of the words: You: I love this movie I saw last night, it’s awesome! One way to represent that kind of emphasis is by using some special characters in the plain text: You: I /love/ this movie I saw last night, it’s *awesome*! That’s a bit of a kludge, though. Thankfully, XMPP enables you to customize the look or presentation of messages, using a subset of HTML as defined in XHTML-IM [XEP-0071]: I love this movie I saw last night, it's awesome!

I love, this new movie I saw last night, it's awesome!

52 | Chapter 4: Instant Messaging,
As you can see, your client sends the plain-text message body plus the marked-up ver- sion. That way, if your friend is using a client that doesn’t understand XHTML markup, the key content of the message still gets through. Although we formatted the italics and bold text using the XHTML and elements, you can also format text using Cascading Style Sheets (CSS). This en- ables you to include a number of popular stylistic formats, including colors, font families, text sizes, font weights (e.g., bold) and styles (e.g., italic), margins, text align- ment (e.g., center), and text decoration (e.g., underline). The XHTML-IM subset also provides support for some of the core HTML presentation features, including numbered and unordered lists, hypertext links, and images. Missing from that list are more advanced HTML features such as tables and media objects, as well as anything that normally goes in the tag of an HTML document, such as scripts. This is intentional, because some of these features could be used to include malicious code (yes, the designers of XMPP are always thinking hard about security!). Instead, XHTML-IM is focused on a simple subset of HTML features that can be used for lightweight presentation in the context of rapid-fire chat conversations. Even so, XMPP clients should exercise caution about receiving XHTML-formatted messages from unknown entities, since even the inclusion of image references could introduce security vulnerabilities. One such preventive measure is to accept XHTML- IM formatting only from people in your roster.

Who Are You? vCards

Sometimes you want to find out more information about the people you chat with. Perhaps someone has sent you a message out of the blue or asked to subscribe to your presence information. Before you continue the conversation or approve the subscrip- tion request, you wonder to yourself: just who is this person? Don’t worry, XMPP has you covered. The extension we’re interested in here is called vCard-temp [XEP-0054], and enables you to publish a kind of electronic business card called a vCard, and to retrieve vCards that other people have published. The vCard standard (originally published in vCard MIME Directory Profile [RFC 2426]) defines many of the basic data fields you might want to advertise, including your name, nickname, address, phone and fax number, company affiliation, email address, birthday, a pointer to your website, a photo of you, and even your PGP key. You don’t have to publish any of that information if you don’t want to, but doing so enables people to find out more about you, which can grease the wheels of communication. So let’s say that Alice in Wonderland sends an unsolicited message to a poor, hapless mouse: Who Are You? vCards | 53, O Mouse, do you know the way out of this pool?

Before replying, the mouse might check Alice’s vCard by sending an IQ-get to her JabberID:

Because the request was sent to Alice’s bare JID, Alice’s server replies on her behalf:

Alice http://wonderland.lit/~alice/

As a result, the mouse can at least visit Alice’s website and view a picture of her before

continuing the chat. Naturally, all of the data in a vCard can be faked, so it pays to take any given vCard result with a grain of salt. But in many situations, it’s better than nothing!

To update your vCard, send an IQ-set to your server. Here Alice adds an email address

and uploads the entire vCard to her server (no, it’s not possible to upload only a “diff,” as the vCard-temp specification does not provide for that feature): Alice http://wonderland.lit/~alice/ email is hidden 54 | Chapter 4: Instant Messaging, Is vCard Really “temp”? The vCard format used by the early Jabber developers was derived from an experimental XML representation of the official vCard format. Re- cently, the IETF has begun work on a more modern and stable approach to XML vCards, and it is possible that the XMPP community will adopt that standard instead of using vCard-temp (which has been “temp” since 1999!).

Talk to the Hand: Blocking and Filtering Communication

Lots of people use XMPP-based IM services (probably over 50 million of them, although we have no way of knowing, because XMPP is a distributed, decentralized technology). But you might not want to chat with them all. In fact, you might want to actively block a certain person from chatting with you—say, your old boss, a childhood enemy, or that weird guy you met in a chat room last week. Because the XMPP developers care about privacy, they have defined an extension for communications blocking (defined in Privacy Lists [XEP-0016]), as well as a stripped- down interface to privacy lists (defined in Simple Communications Blocking [XEP-0191]). First we’ll look at simple communications blocking because it’s, well, simple.

Blocking: The Simple Approach

Let’s say you want to block communications from your old boss at It’s easy enough to do if your server supports simple communications blocking—just send an appropriate IQ-set: Now, what does blocking email is hidden mean exactly? First of all, you want to appear offline to your old boss. When you add the block rule for that JabberID, your server sends out an unavailable presence packet, so that your old boss sees you go offline. From then on, whenever you update your presence (e.g., by coming online), the associated presence stanzas will not be sent to boss@bigcom (as far as he is concerned, it’s as if you never log in anymore). Second, your server needs to make sure that your old boss cannot find out that you are online in any other way. This means that your server will respond to every incoming IQ-get or IQ-set with a error, ignore any incoming Talk to the Hand: Blocking and Filtering Communication | 55, message (or, again, return a error), and drop any incoming stanza.

Finally, your server needs to prevent you from doing something daft, like sending a

message or IQ request to your old boss, so it will reply to any outbound stanza intended for email is hidden with a error.

You can also block entire domains. Let’s say that you have started to receive unsolicited

messages from a rogue server on the XMPP network (perhaps spammers.lit). You can block messages from any JabberID at that domain by setting another block rule:

Now when you retrieve your “block list,” you will see two items:

In simple communications blocking, it is also straightforward to unblock someone. Simply send an IQ-set with the JabberID contained in an element instead

of a element: 56 | Chapter 4: Instant Messaging,

Advanced Blocking and Filtering

Sometimes you want to have more control over blocking and filtering rules than simple communications blocking will give you. For example, when you are using your mobile phone to log into your IM server, you don’t want to receive status updates from your 200 coworkers, as this would clog up your very limited bandwidth. On the other hand, you do want to receive the occasional messages they send you. Moreover, you also don’t want to block all incoming presence packets, as you want to know which members of your family are online, so you can chat with them before leaving on an overseas trip. Thus you need a finer-grained protocol for controlling your traffic filtering rules. Here, again, XMPP comes to the rescue. Whereas simple communications blocking used a basic block list, the full-featured privacy protocol uses a more advanced privacy list. A privacy list is a list of rules that are matched against all traffic, both incoming and outgoing. If one of the rules matches an outgoing packet, the associated action of the rule is applied on the packet. For example, consider the following privacy list: Let’s see how to parse this into plain English: • An incoming message from email is hidden would match the first rule. There- fore, if your server receives an IQ or message stanza from your old boss, it will discard the stanza or return an error. • However, if your server receives a presence stanza from your old boss, that stanza is not matched by the first privacy rule, so your server proceeds to the next rule. Since you don’t work with your old boss anymore, he is not in the “Work” group of your roster. Therefore, your server proceeds to the next (and, in this case, final rule). Lo and behold, the inbound presence stanza matches the final rule, so your server allows the stanza through. Now you can see when your old boss is online, but he can’t communicate with you! The possible combinations of particular privacy rules provide a powerful tool for allowing and blocking communication, because your privacy list can include an un- limited number of privacy rules in any specified order (each identified by an element, as shown earlier). The action for any given rule is either allow or deny, and the rule type processes stanzas based on a specific or wildcard JabberID, on a roster group name, or on a presence subscription state. Finally, stanzas are matched based on Talk to the Hand: Blocking and Filtering Communication | 57, whether they are messages, inbound presence notifications (i.e., not including subscription-related presence stanzas), outbound presence notifications, IQs, or all stanzas (including subscription-related stanzas). In practice, these more advanced block and allow methods provide basic filtering instead of just simple blocking (al- though at the price of greater complexity).

More Messaging Extensions

This chapter provided an overview of various messaging-related extensions in XMPP. But not all of them! Here is a quick look at a few more. Refer to the specifications for all the details, and make sure you check for support in your favorite client, server, or library, because some of these are not yet widely implemented: • Extended Stanza Addressing [XEP-0033] lets you send a single message to multiple recipients at the same time, without using a dedicated chat room. • Advanced Message Processing [XEP-0079] provides a way to control the delivery of a message; examples include message expiration and preventing messages from being stored offline for later delivery. • Message Receipts [XEP-0184] do just what you would expect it to do based on the title: they provide an end-to-end mechanism for determining whether the intended recipient has indeed received a message (by contrast, Advanced Message Process- ing notifications are generated by servers, not clients). • Message Archiving [XEP-0136] defines a technology for storing messages on your server instead of archiving them to your local machine. There are many scenarios in which this is helpful: perhaps you are using a web client that does not have local storage, the device you are using (e.g., a PDA or mobile phone) has limited storage capacity, or you move between different devices quite a bit and you want all of your message history in one place.


Instant messaging is not only the most visible application of the ability to quickly route data from one point to another, but it is also the most popular (with over 50 million XMPP users worldwide). IM interactions usually take the form of chat sessions: short bursts of messages exchanged between two parties. The XMPP extension for chat state notifications provides support for chat sessions by communicating up-to-date infor- mation about the involvement of one’s conversation partner in the discussion. In XMPP, XHTML is used to provide user-friendly formatting, such as bold, italics, and colored text. Furthermore, vCards enable you to find out more about people you might want to chat with, and privacy lists can prevent unwanted communication from other entities. The XMPP developer community continues to work XMPP extensions that will optimize the IM experience. 58 | Chapter 4: Instant Messaging,

CHAPTER 5 Discovering the World

Throughout this book, we talk about many varieties of XMPP entities: servers, clients, bots, chat rooms, pubsub nodes, etc. On the public XMPP network, all of these entities come in multiple flavors. For instance, there are at least half a dozen popular XMPP server implementations, and many more XMPP clients for just about every device and operating system. Furthermore, there are hundreds of possible features that an XMPP entity can support, including standardized protocols (to which the XMPP Standards Foundation is always adding), user-configurable options, client plug-ins, server mod- ules, and more. And let’s not forget that many of these software projects are quite active, frequently releasing updated versions. This diversity is tremendously powerful, but it raises two important questions: 1. How can you learn what entities are out there on the network? 2. Once you find them, how can you determine which XMPP features they support? To answer these questions, you need service discovery (often called “disco” by XMPP developers). When might you want to use service discovery? You might be learning a foreign lan- guage, so you want to find a chat room where you can practice; you might be interested in using a specific publish-subscribe mechanism, so you want to discover a pubsub service where that mechanism is supported; you might like to figure out whether one of your friends or colleagues has video chat capabilities; you might even want to an- nounce to all your contacts that you’re interested in finding out what music they’re listening to. All of these tasks (and more) can be completed using the techniques dis- cussed in this chapter.

Items and Info

The XMPP service discovery protocol defined in [XEP-0030] provides two basic dis- covery methods. The first, known as disco#items, enables you to discover entities. The second, known as disco#info, enables you to discover which features a given entity supports. Let’s look at each of these in turn., It does you no good to discover features unless you have first discovered some entities. A client always knows about at least one entity: the server it connects to. And since XMPP servers typically host additional entities such as pubsub topics and multi-user chat rooms, clients often need to discover those additional entities. Such discovery happens using the disco#items half of the XMPP service discovery protocol by sending an IQ-get to the server. Here, the Mad Hatter queries the wonderland.lit server: That command means “please send me all the items that are associated with wonderland.lit; the server then replies with a list of associated entities, which the client tracks by the value of the id attribute: The only two associated entities in this case are conference.wonderland.lit and notify.wonderland.lit. But what are these entities? What features do they support? To find out, the Mad Hatter needs to query each one individually using the disco#info method. Here the Mad Hatter queries the conference.wonderland.lit service: The XML element for both the items request and the info request is , but the requests are differentiated by the XML namespace: the items request is qualified by the namespace, whereas the info re- quest is qualified by the namespace. Now, the conference.wonderland.lit service returns some information about itself. (Web developers can think of this as similar to the results returned in the HTTP Accept, Accept-Charset, Accept-Encoding, Accept-Language, and Accept-Ranges response head- ers, except that the disco#info response is more extensible.) By interpreting the XML in the foregoing example, the Mad Hatter learns that wonderland.lit hosts a service conference.wonderland.lit, which provides a text con- ferencing service that supports the Multi-User Chat protocol (the protocol/muc namespace defined in [XEP-0045]), in-band registration of usernames (the jabber:iq:register namespace defined in [XEP-0077]), and component vCards (the vcard-temp namespace defined in [XEP-0054]). What’s Your Identity? In most uses of service discovery, we’re mainly interested in the partic- ular features that another entity supports. The element provides a more general clue about what kind of entity this is. The first discovery protocol used in the XMPP community, called “Agent Infor- mation” (see the historical [XEP-0094]), did not disclose detailed fea- tures, but did advertise basic identities, which is why the modern service discovery protocol includes identities as well. (In fact, an entity can ad- vertise multiple identities at the same time, such as a groupchat serv- ice that is simultaneously a native XMPP Multi-User Chat service and a gateway to Internet Relay Chat.)

Using Service Discovery with Servers and Services

The disco#items and disco#info methods are typically used together to “walk the tree” of entities. Consider a typical sequence: 1. Send a disco#items query to the wonderland.lit server, discovering (among others) the conference.wonderland.lit service. 2. Send a disco#info query to the conference.wonderland.lit room, in order to dis- cover that the conference.wonderland.lit service is a multi-user chat service. 3. Send a disco#items query to the conference.wonderland.lit service, discovering (among others) the email is hidden room. 4. Send a disco#info query to the email is hidden room to find out more information about the room (e.g., its name, natural language, and other con- figuration options). 5. Send a disco#items query to the email is hidden service, discov- ering the email is hidden/alice user. Using Service Discovery with Servers and Services | 61,

Clearly, quite a bit of back-and-forth is needed to generate a complete picture of an

entity hierarchy. In common usage, a client would not walk the entire tree automati- cally. Instead, it would not go beyond the first few queries unless the user requests detailed information about, say, a particular chat room and its users.

What follows is the full sequence just outlined. First, query the wonderland.lit server for its associated items:

Second, query the conference.wonderland.lit service to see what kind of entity it is:

Third, query the conference.wonderland.lit service for its associated items:

62 | Chapter 5: Discovering the World,

Fourth, query a particular conference room for its information:

The room then returns a list of its configured features, such as the fact that it is public

(anyone can discover it), persistent (it won’t go away when the last person leaves), open (anyone can join it), semi-anonymous (only the room admins can find out your real

JabberID), unmoderated (new users have voice), and unsecured (no password is re-

quired to join). Here, the room also provides additional information about itself by including a data form of the kind we discuss in Chapter 6 (this extension mechanism for service discovery is defined in [XEP-0128]): 4

Furthermore, querying a conference room with the disco#items namespace may return

a list of JIDs in the room, as shown in the following examples: Using Service Discovery with Servers and Services | 63,

Using Service Discovery with Clients

When it comes to discovering the capabilities of other clients on the network, there are two tools at your disposal: explicit service discovery of the kind we’ve already discussed, and a kind of service discovery shorthand that is advertised in XMPP presence notifi- cations. We’ll look at each of these in turn.

Explicit Service Discovery

In the last section, we said that a client always knows about at least one entity: its server. Yet we’ve also seen that usually a client knows about some other entities: the items in its roster. What can those entities do? Presence plays an important part in helping us find out. As you may recall from Chap- ter 3, when a client goes online, its server sends presence probes to each of the user’s contacts. The server for each contact then returns information about the available re- sources for that contact. This information is not a service discovery list, but a series of presence stanzas, such as the following presence notifications that Alice’s sister receives when she logs in: 64 | Chapter 5: Discovering the World, Disco and Presence Presence helps here because in order to exchange IQ stanzas with an- other user, you need to know the person’s full JID (email is hidden/ resource). Because presence notifications come from the full JID, they are essentially a kind of push format for the data you would need to poll for via the disco#items namespace for each of your contacts.

Now let’s say that Alice’s sister wants to find out what each of Alice’s devices can do

(this kind of information can be used to populate a drop-down box of possible actions, such as “send a file” or “start voice chat”). To find out, she sends a disco#info request to each of Alice’s resources:

For each of Alice’s resources, she receives a reply telling her what features are supported:

Using Service Discovery with Clients | 65, That’s a lot of namespaces! Imagine if you sent a disco#info request to everyone in your roster, and you had 500 or 1,000 or 2,000 contacts. You’re probably thinking, “Isn’t all that traffic going to get expensive? There must be a better way!” And there is. We discuss it in the next section.

Entity Capabilities: Service Discovery Shorthand

The XMPP community has developed an optimized protocol for discovering supported features, at least when you share presence (which you typically do with people in your roster). This Entity Capabilities [XEP-0115] protocol uses presence as the transport for a kind of shorthand service discovery notation. (In fact, just as presence itself is a spe- cialized push version of disco#items data, the entity capabilities protocol uses the XMPP presence “transport” to push out a shorthand notation for disco#info data.) Entity Capabilities is based on the observation that many of the entities with which you interact will be running the same software, and that this software changes relatively infrequently (that is, if you see version 0.13 of the Psi client today, there’s a good chance you’ll see it again tomorrow). Under this scheme, entities advertise the features they support in their presence stanzas. However, it wouldn’t save any bandwidth to send along the full list of features. Therefore, the features are concatenated and hashed into a short verification code that uniquely identifies the current feature set. The first time your client sees this code, it sends a standard disco#info request to find out which features correspond to the code. The good thing is that whenever your client sees this code in the future, it automatically knows exactly which features the person supports, so no disco#info requests are needed, thus saving a lot of bandwidth (especially on login). How does this work in practice? When Alice logs in, her client concatenates and hashes all of the features it supports and then attaches a child to the presence packet containing the verification code in the ver attribute: The hash attribute identifies the hashing algorithm used (the default is SHA-1). The node attribute identifies the software application in use by means of a URI that provides further information about the codebase. Both Alice’s sister and the White Rabbit receive Alice’s presence notification. Alice’s sister has received this verification code before and her client has remembered the ca- pabilities associated with it, so she doesn’t need to send a disco#info request to Alice. However, this is the first time that the White Rabbit’s client has seen this verification code, and so it sends a disco#info request to Alice’s client for the full capabilities: This query can be recognized as a standard service discovery query, with an extra node attribute specified, constructed from the node and ver attributes of the ele- ment. The reason for the node specification, rather than using a disco query without the node attribute, is to prevent the possible race condition where the White Rabbit’s client “discos” Alice, but Alice’s client has changed its capabilities before receiving the disco request, replying with a set of capabilities that do not match the verification code. Alice replies to the disco#info request in the usual way: Upon receiving the response from Alice’s client, the White Rabbit’s client can cache the result, remembering that future clients seen with the same element support the features that Alice has advertised. The result: no need to send a flood of service discovery requests when you log in. (The XMPP network thanks you!) In Chapter 8, we’ll see how entity capabilities are used to optimize the processes of discovering and subscribing to real-time information sources over the XMPP network. Using Service Discovery with Clients | 67,


Service discovery is a key building block of XMPP-based services because it is extremely helpful to determine what entities are out there on the network and exactly which XMPP features those entities implement. Using only the ability to find other entities (disco#items) and to query their capabilities (disco#info), you can discover a wealth of services on the network and discover how those services and the entities in your roster can interact with you in real time. 68 | Chapter 5: Discovering the World,

CHAPTER 6 Data Forms The World Wide Web started out as a way to publish physics papers, and only later

gained transactional capabilities, such as electronic commerce. One of the key building blocks that enabled this transition was the HTML
tag, which made it possible for a website to offer an empty, but structured, form to the user and for the user to submit a completed form to the website.

XMPP includes a very similar technology for lightweight workflows called Data Forms, defined in [XEP-0004]. Because Data Forms are used throughout the technol-

ogies we’ll discuss in the following chapters, we introduce them now. In Chapter 11, we build on these basic concepts to explore more advanced workflows.

Basic Structure A data form is made up of several pieces, as shown in the following “stub” example:

My Special Form Please fill in the following form

Let’s walk through these pieces one by one.

, First, the form is structured as an element qualified by the jabber:x:data name- space. (Because of the namespace name, XMPP developers often refer to this technology as “x:data.”) The element includes a type attribute that specifies where this element lives in the workflow. The overall flow of a data forms exchange is illustrated in Figure 6-1. Figure 6-1. Data form interactions follow a simple workflow The form can include both a human-readable title and human-readable instructions. The bulk of the form consists of fields for which the user will provide values. As in HTML forms, there are several kinds of fields. In x:data, these are differentiated by the type attribute of the element, which can have any of the following values: boolean A true/false option, usually presented as a checkbox or a set of two mutually ex- clusive radio buttons. This is similar to an tag of type checkbox or radio in HTML forms. The lexical representation for a boolean field can be 0 or false for the logical value FALSE and 1 or true for the logical value TRUE. fixed A field that’s presented to the user, but isn’t editable; these are used for labels and instructions. This is similar to the

© DownloadDoc 2023

Terms and conditions | Privacy policy | Sitemap