Effective multicast programming in large scale distributed systems

Many distributed applications have a strong requirement for efficient dissemination of large amounts of information to widely spread consumers in large networks. These include applications in e‐commerce and telecommunication. Publish/subscribe is considered one of the most important interaction styles with which to model communication on a large scale. Producers publish information on a topic and consumers subscribe to the topics they wish to be informed of. The decoupling of producers and consumers in time, space, and flow makes the publish/subscribe paradigm very attractive for large scale distribution, especially in environments like the Internet. This paper describes the architecture and implementation of DACE (Distributed Asynchronous Computing Environment), a framework for publish/subscribe communication based on an object‐oriented programming abstraction in the form of Distributed Asynchronous Collection (DAC). DACs capture the variants of publish/subscribe, without blurring their respective advantages. The architecture we present is tolerant of network partitions and crash failures. The underlying model is based on the notion of Topic Membership: a weak membership for the parties involved in a topic. We present how Topic Membership enables the realization of a robust and efficient reliable multicast on a large scale. The protocol ensures that, inside a topic, even a subscriber who is temporarily partitioned away eventually receives a published message. Copyright © 2001 John Wiley & Sons, Ltd.     

Interest spaces: A unified interest-based dissemination framework for opportunistic networks

Embedded ubiquitous computing systems (EUCS) promise to grow remarkably in the near future. For such systems, new communication technologies are constructed to cope with the increasing functional and temporal demans. Because of the ubiquitousness of mobile devices nowadays, for example, traditional publish/subscribe is no longer an adequate model for data dissemination in mobile networks. Since any node can publish content at any time, the network can get congested easily, so a dissemination paradigm where mobile nodes contribute with a fraction of their resources is needed, through the use of opportunistic networks. Furthermore, a suitable organization for data dissemination in mobile networks should be centered around interests. Thus, we propose a unified interest-based dissemination framework for opportunistic networks entitled Interest Spaces, which simplifies dissemination by just allowing applications to mark data items with certain tags, letting the framework handle the caching, routing, forwarding, and disseminating. Similarly, applications that need to subscribe to channels simply have to specify the tags they are interested in, and the framework does the rest. However, the Interest Spaces framework also allows applications to have more control over the dissemination process if they wish, by specifying various criteria and dissemination rules. In this article, we present the architecture and components of Interest Spaces. We focus on data aggregation at the context layer, showing through simulations the benefits it brings in opportunistic dissemination. We also present a real-life use case for Interest Spaces through Chatty, an opportunistic chat application.     

一种信息共享的主动服务框架

信息共享服务是Internet核心功能之一。提出了一种基于Internet与下一代广播网相互融合的信息共享主动服务框架,探讨了网络层次模型、业务流生成及内容解析等核心技术。该系统立足于Internet信息资源,通过对信息合理的收集、组织与分发,结合统一内容定位和信息过滤技术,实现个性化、主动化的信息服务。最后讨论了基于Web服务的信息订阅/发布主动服务系统的实现模型与实现框架。     

Covering-Based Routing Algorithms for Cyclic Content-Based P/S Overlays

Content-based routing (CBR) publish/subscribe (P/S) system is an important class of distributed systems. This system differs from classical paradigms as messages are routed based on their content rather than their destination address, so as to provide a fine-granularity event dissemination, and support more flexibility decoupling applications. Covering-based routing is a typical optimization method of CBR and has been widely used as a building block in many distributed P/S systems, for it maintains a compact routing table and reduces the costs of communications and matching computations. So far as we know, this optimization method can only be implemented on acyclic overlay network, but cannot be directly utilized on cyclic networks. As the CBR in cyclic systems becomes a new focus of research, developing covering-based protocols and algorithms for cyclic P/S system is becoming significantly important. This paper contributes the cyclic covering-based routing protocol with corresponding algorithms to support covering-based protocol in cyclic P/S system, and implements it in PADRES, a distributed event management infrastructure based on the publish/subscribe model.     

一个新的分布式安全电子拍卖协议

基于Shamir的（t＋1,n）门限方案和公钥加及数字签名提出一个适于分布式松耦合发布／预约系统使用的安全电子拍卖协议。协议除满足了安全分布式拍卖的基本需求外,还保证了代理拍卖服务器的不可否认性及投标者的匿名性。与先前工作相比,新协议具有较多的安全特性,而且更适合于分布式大规模拍卖。松耦合发布预约结构广泛使用于可扩展 自适应的分布式系统中,它的灵活性使之为许多应用所采用。但是,在不安全的通信环境中,这种结构缺乏容错性和安全支持,新协议可以在分布式松耦合发布／预约系统中实现安全性和容错性,同时具有适当的运行效率。最后,我们对协议设计的正确性和效率进行了详尽分析。     

内容发布订阅系统路由算法和自配置策略研究

路由算法和动态自配置特性是实现大规模基于内容的发布订阅系统的两个关键问题.尽管已经有多种路由算法被提了出来,但是它们没有充分利用组播技术提高系统性能和节省网络带宽;此外,已有系统的网络都是静态的,不能够进行网络的自动配置.首先,提出了具有组播集群的层次性系统模型,设计了混合式路由算法,充分利用物理网络组播的特性,节省网络带宽.然后,提出了组播集群复制协议和基于内容的组播树协议CMTP,分别处理节点或者链路失效导致的网络分割以及路由的重建.实验结果表明,这些算法和协议的引入节省了网络带宽,显著提高了系统的性能,保证了系统的自配置特性.     

主动发布订阅分组拓扑和自配置策略

为了提高主动发布/订阅系统中网络节点间的数据传输性能,加强系统的稳定性,针对主动发布/订阅系统的特性,提出了基于事件代理分组的拓扑模型。该方法的主要思路是采用多级分组的方法,先根据发布事件主题的匹配度,将属于同一匹配阀值区间的节点分为同一组;组内节点再根据事件属性的相似性利用聚类算法划分为不同子组。该自配置策略实现了动态分组的创建,代理节点加入和退出的管理。仿真实验表明,分组拓扑模型能有效减少节点间的平均转发次数及系统的拓扑维护开销,提高系统健壮性。     

内容发布订阅系统的组播路由

基于内容的发布订阅系统多数采用泛洪和匹配优先的路由算法。本文提出了基于事件空间划分的发布订阅系统的实现方法,采用组播的方式实现了目标代理到多个订阅者之间的通知路由算法。与采用贪婪路由思想的单播路由算法相比较,该算法提高了在通知路由阶段的发布订阅系统的路由效率。算法对解决基于内容的发布订阅系统中相关的路由问题具有重要的作用和意义。最后给出了算法的仿真实验。     

A scalable and self-adapting notification framework for healthcare information systems

There has been a great interest in publish/subscribe systems in recent years. This interest, coupled with the pervasiveness of light-weight electronic devices, such as cellular phones and personal digital assistants, has opened a new arena in publish/subscribe networks. Currently, many broker overlay networks are static and rarely change in structure. Often, a network overlay structure is predefined or manually modified. This paper presents a dynamic broker network for disseminating critical lab and patient information in a Healthcare information system. The reported work builds upon previous network optimization research on ad hoc publish/subscribe networks. The underlying framework utilizes user-defined cost functions to satisfy quality of service constraints. In essence, the broker network optimization problem is reduced to an incremental search problem to generate low cost network configurations. Certain reliability issues are also addressed by providing a scheduling algorithm to selectively retransmit information and handle broker connectivity failures.     

基于MapReduce的语义发布/订阅系统

为了提高已有的语义发布/订阅系统中事件与订阅匹配的时间效率,提出了基于MapReduce[1]的语义发布/订阅系统[2]。对语义发布/订阅系统的处理流程进行了认知与分解,指出订阅与事件的匹配时间效率是系统必须要解决的问题,在语义发布/订阅模型的基础上设计了新的匹配模块。对过去已完成的匹配方案进行了分解,采用基于MapReduce的并行处理技术[3]对事件与订阅的匹配进行并行处理,从而提高系统的时间效率。该系统通过加入多台具有相同处理能力的处理机,搭建了一个能够并行处理的运行环境,验证了该系统的准确性和有效性。     

Rule-based composite event queries: the language XChangeEQ and its semantics

Web systems, Web services, and Web-based publish/subscribe systems communicate events as XML messages and in many cases, require composite event detection: it is not sufficient to react to single event messages, but events have to be considered in relation to other events that are received over time. This entails a need for expressive, high-level languages for querying composite events. Emphasizing language design and formal semantics, we describe the rule-based composite event query language XChange^EQ. XChange^EQ is designed to completely cover and integrate the four complementary querying dimensions: event data, event composition, temporal relationships, and event accumulation. Semantics are provided as a model theory with accompanying fixpoint theory, an approach that is established for rule languages but has not been applied to event queries so far. Because they are highly declarative, thus easy to understand and well suited for query optimization, such semantics are desirable for event queries.     

面向大规模分布式计算发布订阅系统核心技术

发布/订阅系统技术具有异步、松散耦合和多对多通信的特点,适应了目前动态多变的大规模分布式计算环境的需求,有着广阔的应用前景.分析了国内外发布/订阅系统的研究现状,并从拓扑结构、事件模型和订阅模型等不同角度进行了系统的分类,然后分别就其关键问题从匹配算法、基于内容的路由算法、形式化建模和服务质量等方面进行了阐述,并对已有的典型系统进行了分析比较,指出了当前该领域研究存在的问题和不足.同时,分析了在支持语义和近似匹配来增强系统智能性所面临的挑战,展望了发布/订阅系统在支持移动计算、P2P等新型计算环境下的研究趋势.     

支持移动推送的一种分层式体系结构

移动环境中依赖内容发送的信息服务越来越受到欢迎，本文探讨了有代表性的应用情况，分析了移动推送服务的特征并提出了一种体系结构，该体系结构基于发布／订阅方式，支持松散耦合实体之间的多对多交互。     

可配置可扩展发布订阅系统的研究及其实现

分析了当前发布订阅系统的主要类型和存在的问题.基于发布订阅系统的主要设计模型,采用分层与插件相结合的动态体系结构设计方法,设计和实现了可配置和可扩展的发布订阅系统ACT-PS(ACT-Publish/Subscribe System),支持基于应用需求的配置和扩展.在具体的电子政务数据交换系统当中的配置运行表明,ACT-PS系统为通过可配置和可扩展的方式建立发布订阅系统提供了新的途径.     

一种基于谓词覆盖技术的启发式事件匹配算法

基于内容路由的发布/订购(Pub/Sub)技术具有异步、松散耦合和多对多通信等特点,使得它能更好地应用于大规模分布式交互系统.而高效率的发布事件匹配算法是实现基于内容路由的大规模Pub/Sub系统所要解决的关键问题.针对基于内容的发布/订购模型主要性能的事件匹配问题做了重点研究,提出了谓词关系(二叉树)的概念.并在此基础上,通过将谓词覆盖技术同启发式匹配策略相结合,设计并实现了启发式发布事件匹配算法--PRBT-Match算法.实验结果表明,PRBT-Match算法相对于其他常用匹配算法有更好的匹配性能.