Publish / Subscribe Systems

Professor Sasu Tarkoma, University of Helsinki, FinlandSasu Tarkoma is a full professor in the Department of Computer Science at the University of Helsinki. He leads the Networking and Services specialization line at the Department. His research interests include publish/subscribe, mobile computing, and Internet technology. He received a PhD in computer science from the University of Helsinki. He has authored two textbooks and over 100 scientific articles.

• About the Author xiiiNotes on Contributors xvPreface xvii1 Introduction 1 1.1 Overview 11.2 Components of a Pub/Sub System 41.2.1 Basic System 41.2.2 Distribution and Overlay Networks 51.2.3 Agreements 61.2.4 The Event Loop 71.2.5 Basic Properties 71.3 A Pub/Sub Service Model 91.4 Distributed Pub/Sub 101.5 Interfaces and Operations 111.6 Pub/Sub Semantics for Targeted Delivery 131.7 Communication Techniques 151.8 Environments 171.9 History 181.9.1 Research Systems 191.9.2 Standards 221.9.3 Internet Technology 231.9.4 A Taxonomy 241.10 Application Areas 261.11 Structure of the Book 27References 292 Networking and Messaging 31 2.1 Networking 312.1.1 Overview 312.1.2 Sockets, Middleware, and Applications 332.1.3 Naming and Addressing 342.1.4 Organization 352.1.5 Firewalls and NATs 352.2 Multicast 362.2.1 IP (Network Layer) IP-Multicast 362.2.2 Application-Layer Multicast 382.3 Reverse Path Forwarding and Routing 382.4 Causality and Clocks 392.4.1 Causal Ordering and Lamport Clocks 392.4.2 Vector Clocks 402.4.3 Total Ordering 402.4.4 Discussion 412.5 Message Passing and RPC/RMI 422.5.1 Store and Forward 442.5.2 Concurrent Message Processing 442.5.3 Semantics and QoS 462.6 Web Services 462.6.1 Overview 472.6.2 Asynchronous Processing 482.6.3 The Connector Model 492.6.4 Web Service Platform 502.6.5 Enterprise Service Bus (ESB) 522.6.6 Service Composition 522.7 Session Initiation Protocol (SIP) 532.7.1 SIP Framework 532.7.2 Method Types 542.7.3 Establishing a Session 552.7.4 Extensions 552.8 Summary 56References 563 Overlay Networks and Distributed Hash Tables 59 3.1 Overview 593.2 Usage 613.3 Consistent Hashing 623.4 Geometries 633.5 DHTs 643.5.1 DHT APIs 653.5.2 Chord 653.5.3 Pastry 673.5.4 Discussion 723.6 Gossip Systems 733.6.1 Overview 733.6.2 View Shuffling 753.6.3 Gossip for Pub/Sub 763.7 Summary 77References 774 Principles and Patterns 79 4.1 Introduction 794.2 General Pub/Sub Model 804.2.1 Principles and Characteristics 804.2.2 Message Service 824.2.3 General Patterns 824.2.4 Event Notification Patterns 824.3 Architectural Patterns 834.4 Design Patterns 854.4.1 Structural Patterns 854.4.2 Behavioural Patterns 864.4.3 Concurrency Patterns 864.5 Design Patterns for Pub/Sub 864.5.1 Broker 864.5.2 Observer 874.5.3 Model-View-Control (MVC) 894.5.4 Rendezvous Point 914.5.5 Handoff with Rendezvous 914.5.6 Client-Initiated Connection 924.5.7 Other Patterns 934.6 Event Notifier Pattern 944.6.1 Overview 944.6.2 Structure 954.6.3 Distributed Event Notifier 974.6.4 Design Considerations 984.7 Enterprise Integration Patterns 1014.8 Summary 103References 1035 Standards and Products 105 5.1 CORBA Event Service 1055.2 CORBA Notification Service and Channel Management 1065.3 OMG Data Distribution Service (DDS) 1095.3.1 Overview 1105.3.2 QoS Policies 1115.3.3 Real-Time Communications 1115.3.4 Applications 1125.4 SIP Event Framework 1135.5 Java Delegation Event Model 1145.6 Java Distributed Event Model 1145.7 Java Message Service (JMS) 1155.7.1 Two Communication Models 1165.7.2 Message Types and Selection 1175.7.3 JMS Process 1185.7.4 Message Delivery 1205.7.5 Transactions 1215.7.6 Advanced Issues 1215.7.7 JMS in Java EE and Implementations 1215.8 TibCo Rendezvous 1225.9 COM+ and .NET 1235.10 Websphere MQ 1255.10.1 Overview 1255.10.2 Pub/Sub in WebSphere MQ 1265.11 Advanced Message Queuing Protocol (AMQP) 1275.12 MQ Telemetry Transport (MQTT) 1295.13 Summary 130References 1326 Web Technology 133 6.1 REST 1336.2 AJAX 1346.3 RSS and Atom 1356.4 SOAP 1376.5 XMPP 1396.6 Constrained Application Protocol (CoAP) 1406.7 W3C DOM Events 1416.8 WS-Eventing and WS-Notification 1426.9 Summary 143References 1437 Distributed Publish/Subscribe 145 7.1 Overview 1457.2 Filtering Content 1487.3 Routing Function 1507.4 Topic-Based Routing 1537.4.1 Mechanisms 1547.4.2 Channelization Problem 1547.4.3 Distributed Overlay with Many Topics 1557.4.4 Dynamic Clustering in Topic-Based Pub/Sub 1557.4.5 Summary 1557.5 Filter-Based Routing 1557.6 Content-Based Routing 1577.6.1 Addressing Model 1587.6.2 Propagating Routing Information 1597.6.3 Routing Behaviour: Subscriptions 1607.6.4 Routing Behaviour: Advertisements 1617.6.5 Routing Tables 1627.6.6 Forwarding 1637.6.7 Performance Issues 1647.6.8 A Generalized Broker with Advertisements 1647.7 Rendezvous-Based Routing 1667.8 Routing Invariants 1677.8.1 Configurations 1677.8.2 Pub/Sub Configurations 1687.8.3 False Positives and Negatives 1697.8.4 Weakly Valid Routing Configuration 1697.8.5 Mobility-Safety 1707.8.6 Stabilization and Eventual Correctness 1707.8.7 Soft State 1717.9 Summary 172References 1748 Matching Content Against Constraints 177 8.1 Overview 1778.2 Matching Techniques 1788.3 Filter Preliminaries 1808.4 The Counting Algorithm 1818.4.1 Overview 1828.4.2 Algorithms 1838.5 Matching with Posets 1868.5.1 Poset Preliminaries 1878.5.2 SIENA Poset 1888.5.3 Poset-Derived Forest 1918.5.4 Matching Events 1928.6 Tree Matcher 1938.7 XFilter and YFilter 1948.8 Bloom Filters 1968.8.1 Definition 1978.8.2 Summary Subscriptions 1988.8.3 Multicast Forwarding 1988.8.4 Content-Based Forwarding 1988.8.5 Multi-Level Bloom Filters 2008.9 Summary 200References 2029 Research Solutions 205 9.1 Gryphon 2059.2 The Cambridge Event Architecture (CEA) 2079.3 Scalable Internet Event Notification Architecture (SIENA) 2089.3.1 Event Namespace 2099.3.2 Routing 2099.3.3 Forwarding 2109.3.4 Mobility Support 2119.3.5 CBCB Routing Scheme 2119.4 Elvin 2139.4.1 Clustering 2139.4.2 Federation 2149.4.3 Quench 2149.4.4 Mobile Support 2149.4.5 Nondestructive Notification Receipt 2159.5 JEDI 2159.6 PADRES 2179.6.1 Modular Design 2179.6.2 Load Balancing 2189.6.3 Composite Events 2189.7 REDS 2199.8 GREEN 2209.9 Rebeca 2209.10 XSIENA and StreamMine 2219.11 Fuego Event Service 2229.11.1 Fuego Middleware 2229.11.2 Event Service 2239.11.3 Filtering 2249.11.4 Client-Side API 2249.11.5 Event Router 2249.11.6 Data Structures for Content-Based Routing 2259.12 STEAM 2279.13 ECho and JECho 2279.14 DHT-Based Systems 2289.14.1 Scribe 2289.14.2 Bayeux and Tapestry 2309.14.3 Hermes 2319.14.4 Other Systems 2339.15 Summary 234References 23510 IR-Style Document Dissemination in DHTs 239 10.1 Introduction 23910.2 Data Model and Problem Statement 24010.2.1 Data Model 24010.2.2 Problem Statement and Challenges 24110.3 STAIRS: Threshold-Based Document Filtering in DHTs 24210.3.1 Overview of DHT-Based P2P Networks 24210.3.2 Solution Framework 24210.3.3 Document Forwarding Algorithm 24410.4 Recent Progress and Discussion 24610.4.1 Recent Progress 24610.4.2 Discussion 24710.5 Summary 248References 24811 Advanced Topics 251 11.1 Security 25111.1.1 Overview 25111.1.2 Security Threats 25211.1.3 Security Issues in Pub/Sub Networks 25311.1.4 EventGuard 25411.1.5 QUIP 25511.1.6 Hermes 25511.1.7 Encrypting Attributes 25711.1.8 Privacy 25711.2 Composite Subscriptions 25811.3 Filter Merging 26011.4 Load Balancing 26311.5 Content-Based Channelization 26511.6 Reconfiguration 26611.6.1 Middleware Component Reconfiguration 26711.6.2 Topology Reconfiguration with Failures and Mobile Brokers 26711.6.3 Self-Organizing Pub/Sub with Clustering 26911.7 Mobility Support 27011.7.1 Generic Pub/Sub Mobility 27211.7.2 Graph Based Mobility with Optimizations 27411.8 Congestion Control 27711.8.1 Rate-Control Using Posets 27711.8.2 Explicit Signalling 27911.8.3 Rerouting to Avoid Congestion 27911.9 Evaluation of Pub/Sub Systems 28011.10 Summary 282References 28312 Applications 287 12.1 Cloud Computing 28712.1.1 Pub/Sub for Cloud 28812.1.2 The Windows Azure AppFabric Service Bus 28812.1.3 Amazon Simple Queue Service (SQS) 29112.1.4 PubNub 29112.2 SOA and XML Brokering 29212.3 Facebook Services 29412.3.1 Facebook Messages 29412.3.2 Facebook Chat and Messenger 29512.4 PubSubHubbub 29712.5 Complex Event Processing (CEP) 29912.6 Online Advertisement 30112.7 Online Multiplayer Games 30312.8 Apple Push Notification Service (APNS) 30312.9 Internet of Things 30412.10 Summary 305References 30613 Clean-Slate Datacentric Pub/Sub Networking 309 13.1 Datacentric Communication Model 30913.1.1 Naming of Data 31013.1.2 Content Security 31213.2 CCN 31413.2.1 CCN Node Operation 31413.2.2 CCN Transport Model 31513.2.3 Interest Routing 31613.3 PSIRP/PURSUIT 31713.4 Internet Interdomain Structure 31813.4.1 Policy Routing Problem 32013.4.2 PURSUIT Global Rendezvous 32113.5 Summary 323References 32514 Conclusions 327Index 333