第 1 讲 分布式系统概述 §1.1 分布式系统的定义 §1.2 分布式系统分类 §1.3 分布式系统体系结构

§1.1 分布式系统的定义  A distributed system is: A collection of independent computers that Appears to its users as a single coherent system.

分布式系统定义 Figure 1-1. A distributed system organized as middleware. The middleware layer extends over multiple machines, and offers each application the same interface.

分布式系统的基本目标  Making resources accessible  Distribution transparency  Openness  Scalability

Making Resources Accessible  Making it easy for users and applications to access remote resources  Share remote resources in a controlled and efficient manner  Benefits Better economics by sharing expensive resources Easier to collaborate and exchange information Create virtual organizations where geographically dispersed people can work together using groupware Enables electronic commerce  Problems Eavesdropping or intrusion on communication Tracking of communication to build a profile

Transparency in a Distributed System Figure 1-2. Different forms of transparency in a distributed system (ISO, 1995). Transparency is NOT always good! Cost in performance!

Openness  In form of services  Services are described via interfaces Interface Definition Language (IDL)  Interoperability, Portability, Extensibility

Scalability  Scalability can be measured against three dimensions Size: be able to easily add more users and resources to a system Geography: be able to handle users and resources that are far apart Administrative: be able to manage even if it spans independent administrative organizations

Scaling Techniques  Hiding communication latency Code at client: e.g. Javascript to check a form  Distributing components Divide a component into smaller parts E.g. dividing the DNS name space into zones  Replicating components Caching is a special form of replication

Centralized versus distributed  Centralized approach  Distributed approach No global clock No complete/global information  Decisions based only on local information Failure locality

§1.2 分布式系统分类  Distributed Computing Systems  Cluster Computing Systems  Grid Computing Systems  Cloud Computing Systems  Distributed Information Systems  Transaction Processing Systems  Enterprise Application Integration  Distributed Pervasive Systems  Home Systems  Electronic Health Care Systems  Sensor Networks

Cluster Computing Systems  Figure 1-6. An example of a cluster computing system.

Grid Computing Systems  Figure 1-7. A layered architecture for grid computing systems.

Cloud Computing Systems servers storage Provisioning Manager Virtual Machines Monitoring Manager Provisioning Interface

Transaction Processing Systems  Figure 1-9. A nested transaction.

Transaction Processing Systems  Figure The role of a TP monitor in distributed systems.

Enterprise Application Integration  Figure Middleware as a communication facilitator in enterprise application integration. RPC, RMI and MOM are examples.

Distributed Pervasive Systems  Embrace contextual changes.  Encourage ad hoc composition.  Recognize sharing as the default.

Electronic Health Care Systems  Figure Monitoring a person in a pervasive electronic health care system, using (a) a local hub or (b) a continuous wireless connection.

Wireless Sensor Networks  Centralized storage and process.  Distributed storage and process.

§1.3 分布式系统体系结构  Software architecture how various software components are organized and how they interact  System architecture An instance of a software architecture after deciding on the software components, their interaction and their placement.

Architectural Styles Layered architectures Object-based architectures Event-based architectures Data-shared architectures

System Architecture  Centralized architectures Application layering Multitiered architectures  Decentralized architectures Structured peer-to-peer Unstructured peer-to-peer  Hybrid architectures Edge-server systems Collaborative distributed systems

Centralized Architectures

Application Layering Figure 2-4. The simplified organization of an Internet search engine

Application Layering Figure 2-5. Alternative client-server organizations (a)–(e).

Multitiered Architectures Figure 2-6. An example of a server acting as client.

Decentralized Architecture  Structured peer-to-peer Distributed Hash Table Content-Addressable Network  Unstructured peer-to-peer

Structured Peer-to-Peer Architectures Data mapping in Chord and CAN.

Unstructured Peer-to-Peer Architectures  The overlay network resembles a random graph. Each node maintains a list of c neighbors, where, ideally, each of theses neighbors represents a randomly chosen live node from the current set of nodes. The list of neighbors is referred to as a partial view.  How to maintain a partial view? Nodes are in push or pull mode. Using only one mode leads to isolated sub-networks so most nodes will do both (exchange mode) To add to the group, simply contact any node. To leave, simply leave.

Topology Management of Overlay Networks A two-layered approach for constructing and maintaining specific overlay topologies

Network with Superpeers

Hybrid Architectures: edge-server systems Figure Viewing the Internet as consisting of a collection of edge servers.

Hybrid Architectures: collaborative sys. Figure The principal working of BitTorrent

Summary  分布式系统定义  分布式系统的目标 Making resources accessible Distribution transparency Openness Scalability  分布式系统的类型 Distributed computing system Distributed information system Distributed pervasive system  分布式系统体系结构 Centralized Decentralized Hybrid

Homework Questions