具体描述
内容简介
TCP/IP网络互连技术系列的第Ⅲ卷讨论客户/服
务器编程和应用，讲述了构筑所有分布式计算系统的客
户/服务器计算模型的基本慨念，内容涉及各种不同的
服务器设计方法，以及用来构造客户/服务器的各种工具
和技术，包括远程过程调用RPC。书中包含了用来说明
每种设计和工具的运行程序示例的源代码。卷Ⅲ有三个
版本：BSD套接字版，AT&TTLI版，WindowsSockets
版。本书系WindowsSockets版，讨论在Internet上
应用软件通过TCP/IP进行通信的问题，适用于Windows
95，Wind0wSNT，Win32环境下编程和使用Windows
SocketAPI应用编程接口。
《现代云计算架构与实践》图书简介 聚焦前沿，洞悉未来：构建高可用、高性能的云原生应用 在数字化转型的浪潮中，云计算已成为驱动技术创新的核心引擎。本书并非聚焦于传统的网络协议栈的底层细节，而是将视角提升至宏观的应用架构层面，深入剖析如何利用云计算的最新特性，设计、部署和运维下一代企业级应用。本书旨在为架构师、高级开发人员和DevOps工程师提供一套全面、实用的云原生技术栈指南。 第一部分：云原生时代的架构范式转变 本部分将系统阐述从传统三层架构到现代微服务、无服务器（Serverless）架构的演进路径。我们不会深入探讨TCP/IP协议的端口、分段或路由机制，而是着重分析这些架构选择对服务间通信、数据一致性与弹性的影响。 第一章：微服务架构的基石与挑战 深入探讨微服务的拆分原则（如DDD领域驱动设计），服务注册与发现机制（如Consul, etcd的应用而非底层网络传输），以及如何利用API网关实现集中化的流量控制、认证和限流。重点分析服务间调用（如gRPC与RESTful API的选择）如何影响系统的整体延迟和吞吐量，但分析的侧重点在于应用层面的契约设计和容错机制，而非传输层协议的具体实现细节。 第二章：容器化与Kubernetes的统治地位 本书将容器化视为应用部署的基石，详细介绍Docker与OCI规范，但焦点在于如何利用Kubernetes（K8s）编排复杂应用。我们将详述K8s的四大核心组件（API Server, Scheduler, Controller Manager, etcd）的功能，以及如何使用Deployment、StatefulSet、DaemonSet等资源对象实现应用的声明式管理。关于网络部分，重点在于Service抽象（ClusterIP, NodePort, LoadBalancer）如何为微服务提供稳定的访问入口，以及CNI（Container Network Interface）插件（如Calico, Cilium）如何实现Pod间的隔离与策略，而非纠缠于底层IP隧道或ARP解析。 第三章：服务网格（Service Mesh）的引入与治理 随着微服务数量的激增，服务间的复杂性超出传统负载均衡器的能力范围。本章将详细介绍Istio、Linkerd等服务网格的原理，重点阐述Sidecar代理（如Envoy）如何透明地接管流量，实现熔断、重试、超时配置和金丝雀发布。我们将分析这些治理层面的策略如何提升系统的鲁棒性，但不会涉及Sidecar代理与底层网络设备进行三次握手的具体时序图。 第二部分：数据管理与状态持久化 在云环境中，数据的存储和访问模式发生了根本性的变化。本部分将侧重于数据一致性模型、数据库选型策略和数据访问的高效性，而非文件系统或网络存储设备的数据包传输顺序。 第四章：云原生数据库选型与实践 对比关系型数据库（如PostgreSQL, MySQL）与NoSQL数据库（如MongoDB, Cassandra）在云环境下的伸缩性和可用性。重点介绍NewSQL数据库（如TiDB）如何尝试结合两者的优点。分析 CAP 理论在分布式数据存储中的实际应用，例如如何通过Quorum机制保证读写一致性，以及分布式事务的解决方案（如Saga模式），而不是数据库服务器之间的TCP连接建立过程。 第五章：缓存策略与数据一致性 深入探讨Redis、Memcached等内存缓存技术在提升应用性能中的作用。内容涵盖缓存穿透、雪崩、击穿等问题的解决方案，以及如何设计有效的缓存失效策略。重点分析分布式锁的实现（如基于Redlock算法的探讨），确保在多节点环境下数据更新的原子性。 第六章：事件驱动与数据流处理 分析事件驱动架构（EDA）在解耦系统、实现最终一致性中的优势。详细介绍Apache Kafka作为高吞吐量消息队列的原理，包括其分区（Partitioning）、副本（Replication）机制，以及Producer/Consumer的提交与确认流程。讨论Streams API在实时数据处理中的应用，侧重于流处理的窗口计算与状态管理，而非消息在网络中的路由路径优化。 第三部分：运维自动化与可观测性 现代云应用要求运维具备高度的自动化和前瞻性。本部分关注如何通过工具和流程确保系统的健康运行和快速响应。 第七章：基础设施即代码（IaC）与GitOps 探讨如何使用Terraform、Ansible等工具管理云资源（如VPC、子网、安全组、负载均衡器）的生命周期，实现基础设施的快速部署和版本控制。深入解析GitOps流程，如何将Git仓库作为唯一可信的变更源，驱动ArgoCD或FluxCD对Kubernetes集群进行持续部署。重点关注配置管理与状态同步的逻辑，而非网络配置文件的手工编辑。 第八章：应用性能管理（APM）与可观测性 可观测性被视为云原生运维的核心。本章详细介绍“三支柱”——Metrics（指标）、Logs（日志）、Traces（追踪）的应用。我们将讲解Prometheus如何通过Pull模型收集时间序列数据，ELK/Loki栈如何处理海量日志，以及Jaeger/Zipkin如何实现分布式追踪。重点分析如何利用Trace Span来定位请求在微服务链中的延迟瓶颈，而不是分析请求在不同网络跳点间的TTL消耗。 第九章：混沌工程与弹性设计 介绍Netflix Chaos Monkey等工具在验证系统韧性方面的作用。讨论如何系统性地设计和执行故障注入实验，测试系统的恢复能力和容错边界。内容包括网络分区模拟、延迟注入、资源耗尽等场景，目标是验证应用代码和架构设计（如超时与重试机制）是否能在异常情况下保持服务的可用性，而不是测试网络设备对丢包的自然反应。 本书的深度在于对应用层协议、架构设计哲学和自动化工具链的掌握，是面向构建下一代大规模、高弹性云服务的技术人员的必备参考书。它假定读者对操作系统的基本原理和网络模型有基础认知，但其核心价值在于将这些基础知识转化为实际的云架构决策和工程实践。
作者简介
目录信息
Contents
Foreword
Preface
Chapter 1 Introduction And Overview
1.1 Use OfTCP/IP
1.2 Designing Applications For A Distributed Environment
1.3 Standard And Nonstandard Application Protocols
1.4 An Example Of Standard Application Protocol Use
1.5 An Example Connection
1.6 Using TELNET To Access An Altemative Service
1.7 Application Protocols And Software Flexibility
1.8 Viewing Services From The Provider's Perspeclive
1.9 The Remainder Of The Text
1.10 Summary
Chapter 2 The Client Server Model And Software Design
2.1 Introduction
2.2 Molivation
2.3 Terminology And Concepts
2.3.1 Clients And Servers
2.3.2 Privilege And Complexity
2.3.3 Standard Vs. Nonstandard Client Software
2.3.4 Pardmeterization Of Clients
2.3.5 Connectionless Vs. Connection-Oriented Servers
2.3.6 Stateless Vs. Stateful Servers
2.3.7 A Stateful File Server Example
2.3.8 .Statelessness Is A Protocol Issue
2.3.9 Servers As Clients
2.4 Summary
Chapter 3 Concurrent Processing In Client-Server Software
3.1 Introduction
3.2 Concurrency In Networks
3.3 Concurrency In Servers
3.4 Terminology And Concepts
3.4.1 The Process Concept
3.4.2 Threads
3.4.3 Programs vs. Threads
3.4.4 Procedure Calls
3.5 An Example Of Concurrent Thread Creation
3.5.1 A Sequential C Example
3.5.2 A Concurrent Version
3.5.3 Timeslicing
3.6 Diverging Threads
3.7 Context Switching And Protocol Software Design
3.8 Concurrency And Asynchronous 1/o
3.9 Concurrency Under UNIX
3.10 Execuling A Separately Compiled Program
3.11 Summary
Chapter 4 Program Interface To Protocols
4.1 Introduction
4.2 Loosely Specified Protocol Software Interface
4.2.1 Advantages And Disadvantages
4.3 Interface Functionality
4.4 Conceptua! Interface Speclfication
4.5 Implementation Of An APl
4.6 Two Basic Approaches To Network Communicatwn
4.7 The Basic 1/0 Functions Available In ANSI C
4.8 Hislory Of The UNIX Socket APl
4.9 Summary
Chapter 5 TheSocketAPI
5.1 Introduction
5.2 The History Of Sockets
5.3 Speifving A Protocol Interface
5.4 The Socket Ahstraction
5.4.1 Sockel Descriptors
5.4.2 System Data Structures For Sockets
5.4.3 Using Sockets
5.5 Specifying An Endpoint Address
5.6 A Generic Address Slructure
5.7 Functions In The Sockel API
5.7.1 The WSAStartup Function
5.7.2 The WSACleanup Function
5.7.3 The Socket Function
5.7.4 The Connect Function
5.7.5 The Send Function
5.7.6 The Recv Function
5.7.7 The Closesocket Function
5.7.8 The Bind Function
5.7.9 The Listen Function
5.7.10 The Accept Function
5.7.11 Summary Of Socket Calls Used Wilh TCP
5.8 Utility Roulines For Integer Conversion
5.9 Using Socket Culls In A Program
5.10 Symbolic Constants For Socket Call Parameters
5.11 Summary
Chapter 6 Algorithms And Issues In Client Software Design
6.1 Inlroduclion
6.2 Leurning, Atgorilhms Instead Of Delails
6.3 Client Architecture
6.4 Idenlifying The Location OfA Server
6.5 Parsing An Address Argumenl
6.6 Looking Up A Domain Name
6.7 Looking Up A Well-Known Port By Name
6.8 Port Numbers And Network Byte Order
6.9 Looking Up A Protocol By Name
6.10 The TCP Client Algorithm
6.11 Alloating A Socket
6.12 Choosing A Local Protocol Port Number
6.13 A Fundamenlal Problem In Choosing A Local IP Address
6.14 Connecting A TCP Socket To A Server
6.15 Commuiucating With The Server Using TCP
6.16 Reading A Response From A TCP Connection
6.17 Closing A TCP Connection
6.17.1 The Need For Partial Close
6.17.2 A Partial Close Operation
6.18 Programming A UDP Client
6.19 Coimected And Unconnected UDP Sockets
6.20 Using Connecl With UDP
6.21 Communicating With A Server Using UDP
6.22 Closing A Socket That Uses UDP
6.23 Partial Close For UDP
6.24 A Wai-ning About UDP Unreliability
6.25 Summarv
Chapter 7 Example Client Software
7.1 ntroduction
7.2 T he Imfwrtance Of Small Examples
7.3 Hiding Dpltails
7.4 An Example Pmcedure Library For Client Programs
7.5 Implementiation OfConTCP
7.6 Implementation Of ConUDP
7.7 A Procedure That Forms Connections
7.8 Using The Example Library
7.9 The OA YTIME Service
7.10 Implemenuation OfA TCP Client For DA YTIME
7.11 Reading From A TCP Connection
7.12 The TIME Service
7.13 Accessing The TIME Service
7.14 Accurale Times And Network Delays
7.15 A UDP Client For The TIME Service
7.16 The ECHO Service
7.17 A TCP Client For The ECHO Service
7.18 A UDP Client For The ECHO Service
7.19 Summary
Chapter 8 Algorithms And Issues In Server Software Design
8.1 Inlroduction
8.2 The Conceptual Senver Algorithm
Concurrent Vs. Iferative Servers
Connection-Oriented Vs. Connectionless Access
Connection-Oriented Servers
Connectionless Servers
Failure, Reliability, And Statelessness
Optimizing Stateless Servers
Four Basil' Types Of Servers
Request Processing Time
llerative Server Algorithms
An Iterative, Connection-Oriented Server Algorithm
Binding To A Well-Known Address Using INADDR_ANY
Placing The Socket In Passive Mode
Accepling Connections And Using Them
An Iterative, Connectionless Server Algorithm
Forming A Reply Address In A Connectionless Server
Concurrent Server Algorithms
Masler And Slave Threads
A Concurrent Connectionless Server Algorithm
A Concurrent, Connection-Oriented Server Algorithm
Using Separate Programs As Slaves
Apparenl Concurrency Using A Single Thread
When To Use Each Server Type
A Summary of Server Types
The Important Problem Of Server Deadlock
Alternative Implementations
Summary
Chapter 9 Iterative, Connectionless Servers (UDP)
9.1 Introduction
9.2 Creating A Passive Socket
9.3 Thread Structure
9.4 An Example TIME Server
9.5 Summary
Chapter 10 Iterative, Connection-Oriented Servers (TCP)
10.1 Introduction
10.2 Allocating A Pcassive TCP Socket
10.3 A Server For The DAYTIME Service
10.4 Thread Structure
10.5 An Example DA YTIME Server
10.6 Closing Connections
10.7 Conneclion Termination And Server Vulnerability
10.8 Summary
Chapter 11 Concurrent, Connection-Oriented Servers (TCP)
11.1 Introduction
11.2 Concurrent ECHO
11.3 Iterative Vs. Concurrent Implementations
11.4 Thread Slructure
11.5 An Example Concurrent ECHO Server
11.6 Summary
Chapter 12 Singly-Threaded, Concurrent Servers (TCP)
12.1 Introduction
12.2 Data-driven Processing In A Server
12.3 Data-Driven Processing With A Single Thread
12.4 Thread Structure OfA Singly-Threaded Server
12.5 An Example Singly-Threaded ECHO Server
12.6 Summary
Chapter 13 Multiprotocol Servers (TCP, UDP)
13.1 Introduction
13.2 The Motivation For Reducing The Number OfServers
13.3 Multiprotocol Server Design
13.4 Thread Structure
13.5 An Example Multiprotocol DAYTIME Server
13.6 The Concept Of Shared Code
13.7 Concurrent Multiprolocol Servers
13.8 Summary
Chapter 14 Multiservice Servers (TCP, UDP)
14.1 Introduction
14.2 Consolidaling Servers
14.3 A Conneclionless, Multiservice Server Design
14.4 A Connection-Oriented, Multiservice Server Design
14.5 A Concurrent, Connection-Oriented, Multiservice Server
14.6 A Singly-Threaded, Multiservice Server Implementation
14.7 Invoking Separate Programs From A Multiservice Server
14.8 Multiservice, Multiprotocol Designs
14.9 An Example Multiservice Server
14.10 Static and Dynamic Server Configuration
14.11 An Example Super Server, Inetd
14.12 Summary
Chapter 15 Uniform, Efficient Management Of Server Concurrency
15.1 Introduction
15.2 Choosing Between An Iteralive And A Concurrent Design
15.3 Level Of Concurrency
15.4 Demand-Driven Concurrency
15.5 The Cost Of Concurrency
15.6 Overhead And Delay
15.7 Small Delays Can Matter
15.8 Thread Preallocation
15.8.1 Preallocation Techniques
15.8.2 Preallocation In A Connection-Oriented Server
15.8.3 Preallocation In A Connectionless Server
15.8.4 Preallocation, Bursty Traffic, And NFS
15.8.5 Preallocation On A Multiprocessor
15.9 Delayed Thread Allocation
15.10 The Uniform Basis For Both Techniques
15.11 Combining Techniques
15.12 Summary
Chapterl6 Concurrency In Clients
16.1 Introduction
16.2 The Advantages Of Concurrency
16.3 The Motivation For Exercising Control
16.4 Concurrenl Contact With Multiple Servers
16.5 Implemenling Concurrent Clients
16.6 Singly-Threaded Implementations
16.7 An Example Concurrent Client That Uses ECHO
16.8 Execution OfThe Concurrent Client
16.9 Managing A Timer
16.10 Example Output
16.11 Concurrency In The Example Code
16.12 Summary
Chapter 17 Tunneling At The Transport And Application Levels
17.1 Introduction
17.2 Multiprotocol Environments
17.3 Mixing Network Technologies
17.4 Dynamic Circuit Allocalion
17.5 Encapsulation And Tunneling
17.6 Tunneling Through An IP Intemet
17.7 Applicalion-Level Tunneling Between Clients And Servers
17.8 Tunneling, Encapsulation, And Dialup Phone Lines
17.9 Summary
Chapter 18 Appiication Level Gateways
18.1 Inlroduction
18.2 Clients And Servers In Constrained Environments
18.2.1 The Reality Of Multiple Technologies
18.2.2 Computers With Limited Functionality
18.2.3 Connectivity Constraints That Arise From Security
18.3 Using Application Gateways
18.4 Interoperability Through A Mail Gateway
18.5 Implementation Of A Mail Gateway
18.6 A Comparison Of Application Gateways And Tunneling
18.7 Application Gateways And Limited Functionality Systems
18.8 Application Gateways Used For Security
18.9 Application Gateways And The Extra Hop Problem
18.10 An Example Application Gateway
18.11 Delails OfA Web-Based Application Gateway
18.12 Invoking A CGI Program
18.13 URLs For The RFC Application Gateway
18.14 A General-Purpose Application Gateway
18.15 Operation Of SLIRP
18.16 How SLIRP Handles Connections
18.17 IP Addressing And SURP
18.18 Summary
Chapter 19 External Data Representation (XDR)
19.1 Introduction
19.2 Representations For Data In Computers
19.3 The N-Squared Conversion Problem
Network Standard Byte Order
A De Facto Standard External Data Representation
XDR Datu Types
Implicil Types
Software Supporl For Using XDR
XDR Library Routines
Building A Message One Piece At A Time
Conversion Routines In The XDR Library
XDR Streams. 1/0, and TCP
Records, Record Boundaries, And Datagram 1/0
Summary
Chapter 20, Remote Procedure Call Concept (RPC)
20.1 Introduction
20.2 Remote Procedure Call Model
20.3 Two Paradigms For Building Distributed Programs
20.4 A Conceptual Modet For Conventional Procedure Calls
20.5 An Extension Of the Procedural Model
20.6 Execution Of Conventional Procedure Call And Return
20.7 The Procedural Model In Distributed Systems
20.8 Analogy Between Client-Server And RPC
20.9 Distributed Computation As A Program
20.10 Sun Microsystems' Remote Procedure Call Definition
20.11 Remote Programs And Procedures
20.12 Reducing The Number Of Arguments
20.13 Identifying Remote Programs And Procedures
20.14 Accommodating Multiple Versions OfA Remote Program
20.15 Mutual Exclusion For Procedures In A Remote Program
20.16 Communication Semantics
20.17 At Least Once Semantics
20.18 RPC Relransmission
20.19 Mapping A Remote Program To A Protocol Port
20.20 Dynamic Porl Mapping
20.21 RPC Port Mapper Algorithm
20.22 RPC Message Format
20.23 Marshaling Arguments For A Remote Procedure
20.24 Authentication
20.25 An Example Of RPC Message Representation
20.26 An Example OfAn Authentication Field
20.27 Summary
Chapter 21 Distributed Program Generation (Rpcgen Concept)
21.1 Introduction
21.2 Using Remote Procedure Calls
21.3 Programming Mechanisms To Support RPC
21.4 Dividing A Program Inlo Local And Remote Procedures
21.5 Adding Code For RPC
21.6 Stub Procedures
21.7 Mulliple Remote Procedures And Dispatching
21.8 Name Of The Client-Side Slub Procedure
21.9 Using Rpcgen To Generate Distributed Programs
21.10 Rpcgen Output And Interface Procedures
21.11 Rpcgen Input And Output
21.12 Using Rpcgen To Build A Client And Server
21.13 Summary
Chapter 22 Distributed Program Generation (Rpcgen Example)
22.1 Introduction
22.2 An Example To lllustrate Rpcgen
22.3 Diclionary Look Up
22.4 Eight Steps To A Distributed Application
22.5 Slep 1: Build A Conventional Application Program
22.6 Step 2: Divide The Program Into Two Parts
22.7 Step 3: Create An Rpcgen Specification
22.8 Step 4: Run Rpcgen
22.9 The h File Produced By Rpcgen
22.10 The XDR Conversion File Produced By Rpcgen
22.11 The Client Code Produced By Rpcgen
22.12 The Server Code Produced By Rpcgen
22.13 Step 5: Write Stub Interface Procedures
22.13.1 Client-Side Inlerface Routines
22.13.2 Server-Side Interface Routines
22.14 Step 6: Compile And Link The Client Program
22.15 Slep 7: Compile And Link The Server Program
22.16 Step 8: Start The Server And Execute The Client
22.17 Summarv
Chapter 23 Network File System Concepts (NFS)
23.1 Introduction
23.1 Remote File Access Vs. Transfer
23.3 Operations On Remote Files
23.4 File Access Among Heterogeneous Computers
23.5 Stateless Servers
23.6 NFS And UNIX File Semanlics
23.7 Review Of The UNIX File System
23.7.1 Basic Definitions
23.7.2 A Byte Sequence Without Record Boundaries
23.7.3 A File 's Owner And Group Identifiers
23.7.4 Protection And Access
23.7.5 The UNIX Open-Read-Write-Close Paradigm
23.7.6 UNIX Data Transfer
23.7.7 Permission To Search A Directory
23.7.8 UNIX Random Access
23.7.9 Seeking Beyond The End Of A UNIX File
23.7.10 UNIX File Position And Concurrent Access
23.7.11 Semantics Of Write During Concurrent Access
23.7.72 UNIX File Names And Paths
23.7.13 The UNIX tnode: Information Stored With A File
23.7.14 The UNIX Stat Operation
23.7.15 The UNIX File Naming Mechanism
23.7.16 UNIX File System Mounts
23.7.17 UNIX File Name Resolution
23.7.18 UNIX Symbolic Links
23.8 Files Under NFS
23.9 NFS File Types
23.10 NFS File Modes
23.11 NFS File Attributes
23.12 NFS Client And Server
23.13 NFS Client Operation
23.14 NFS Client And UNIX
23.15 NFS Mounts
23.16 File Handle
23.17 NFS Handles Replace Path Names
23.18 An NFS Client Under Windows
23.19 File Positioning With A Stateless Server
23.20 Operations On Directories
23.21 Reading A Directory Slatelessly
23.22 Mulliple Hierarchies In An NFS Server
23.23 The Mount Protocol
23.24 Summary
Chapter 24 Network File System Protocol (NFS, Mount)
24.1 Inlroduction
24.2 Using RPC To Define A Protocol
24.3 Defining A Protocol With Data Structures And Procedures
24.4 NFS Conslanl, Type, And Data Declarations
24.4.1 NFSConstants
24.4.2 NFS Typedef Declarations
24.4.3 NFS Data Structures
24.5 NFS Procedures
24.6 Semantics Of NFS Operations
24.6.1 NFSPROC_NULL (Procedure 0)
24.6.2 NFSPROC_GETA TTR (Procedure 1)
24.6.3 NFSPROC_SETATTR (Procedure 2)
24.6.4 NFSPROC_ROOT (Procedure 3) [Ohsolete in NFS3]
24.6.5 NFSPROC_OOKUP (Procedure 4)
24.6.6 NFSPROC_READLlNK (Procedure 5)
24.6.7 NFSPROC_READ (Procedure 6)
24.6.8 NFSPROC_WRITECACHE (Procedure 7) [Obsolele in NFS3]
24.6.9 NFSPROC_WRITE (Procedure 8)
24.6.10 NFSPROC_REATE (Procedure 9)
24.6.11 NFSPROC_REMOVE (Procedure 10)
24.6.12 NFSPROC_RENAME (Procedure 11)
24.6.13 NFSPROC_LINK (Procedure 12)
24.6.14 NFSPROC_SYMUNK (Procedure 13)
24.6.15 NFSPROC_MKDIR (Procedure 14)
24.6.16 NFSPROC_RMD1R (Procedure 15)
24.6.17 NFSPROC_READDlR (Procedure 16)
24.6.18 NFSPROC_STATFS (Procedure 17)
24.7 The Mount Protocol
24.7.1 Mount Constant Definitions
24.7.2 Mount Ty'pe Definitions
24.7.3 Mount Data Structures
24.8 Procedures In The Mount Protocol
24.9 Semantics of Mount Operations
24.9.1 MNTPROC_ULL (Procedure 0)
24.9.2 MNTPROC_MNT (Procedure 1)
24.9.3 MNTPROC_DUMP (Procedure 2)
24.9.4 MNTPROC_MNT (Procedure 3)
24.9.5 MNTPROC_UMNTALL (Procedure 4)
24.9.6 MNTPROC_EXPORT (Procedure 5)
24.10 NFS And Moimt Authentication
24.11 Changes In NFS Version 3
24.12 Summarv
Chapter 25 A TELNET Client (Program Structure)
25.1 Introduction
25.2 Overview
25.2.1 The User's Terminal
25.2.2 Command And Control Information
25.2.3 Tenninals, Windows, and Files
25.2.4 The Need For Concurrency
25.2.5 A Thread Model For A TELNET Client
25.3 A TELNET Client Algorithm
25.4 Keyboard 1/0 In Windows
25.5 Global Variables Used For Keyboard Control
25.6 Initializing The Keyboard Thread
25.7 Finite Stale Machine Specification
25.8 Embedding Commands In A TELNET Data Stream
25.9 Option Negotiation
25.10 Request/Offer Symmetry
25.11 TELNET Character Definitions
25.12 A Finite State Machine For Data From The Server
25.13 Transitions Among States
25.14 A Finite State Machine Implementation
25.15 A Compact FSM Representa'tion
25.16 Keeping The Compact Representation At Run-Time
25.17 Implementation OfA Compact Representation
25.18 Building An FSM Transition Matrix
25.19 The Socket Output Finite State Machine
25.20 Definitions For The Socket Output FSM
25.21 The Option Subnegotiation Finite State Machine
25.22 Definitions For The Option Subnegotiation FSM
25.23 FSM Initialization
25.24 Arguments For The TELNET Client
25.25 The Heart Of The TELNET Client
25.26 TELNET Synchronization
25.27 Handling A Severe Error
25.28 Implemenlation Of The Main FSM
25.29 A Procedure For Immediate Disconnection
25.30 Abort Procedure
25.31 Summary
Chapter 26 A TELNET Client (Implementation Details)
26.1 Introduction
26.2 The FSM Action Procedures
26.3 Recording The Type Of An Option Requesl
26.4 Performing No Operation
26.5 Responding To WILLWONT For The Echo Oplion
26.6 Sending A Response
26.7 Responding To WILUWONT For Unsupported Options
26.8 Responding To WLLWONT For The No Go-Ahead Option
26.9 Generating DO/DONT For Binary Transmission
26.10 Responding To DO/DONT For Unsupporled Options
26.11 Responding To DO/DONT For Transmit Binary Option
26.12 Responding To DO/DONT For The Terminal Type Option
26.13 Option Subnegoliation
26.14 Sendmg Terminal Type Information
26.15 Tennincning Suhnegotiation
26.16 Sending A Character To The Server
26.17 Displaying Incoming Data On The User's Terminal
26.18 Writing A Block Of Data To The Server
26.19 Interacting With The Local Client
26.20 Responding To lllegat Commands
26.21 Scripting To A File
26.22 Implementation Of Scripting
26.23 Initialzation Of Scripting
26.24 Collecting Characters OfThe Script File Name
26.25 Opening A Script File
26.26 Terminating Scripting
26.27 Printing Status Information
26.28 Summarv
Chapter 27 Porting Servers From UNIX To Windows
27.1 Introduction
27.2 Operating in Background
27.3 Shared Descriptors And Inherilance
27.4 The Controlling TTY
27.5 Working Directories
27.6 File Creattion And Umask
27.7 Process Groups
27.8 Descriptors For Slandard 1/0
27.9 Mutual Exclusion For A Server
27.70 Recording A Process ID
27.11 Waiting For A Child Process To Exit
27.12 Using A Syslem Log Facility
27.12.1 Generating Log Messages
27.13 Miscellaneous Incompatibilities
27.14 Summarv
Chapter 28 Deadlock And Starvation In Client-Server Systems
28.1 Introduction
28.2 Definition Of Deadlock
28.3 Difficulty Of Deadlock Detection
28.4 Deadlock Avoidance
28.5 Deadlock Between A Client And Server
28.6 Avoiding Deadlock In A Single Interaction
28.7 Starvalion Among A Set Of Clients And A Server
28.8 Busy Connections And Starvation
28.9 Avoiding Blocking Operations
28.10 Threads, Connections, And Other Limits
28.11 Cycles Of Clients And Servers
28.12 Documenting Dependencies
28.13 Summary
Appendix 1 Functions And Library Routines Used With Sockets
Appendix 2 Manipulation Of Windows Socket Descriptors
Bibliography
Index
· · · · · · (收起)
Foreword
Preface
Chapter 1 Introduction And Overview
1.1 Use OfTCP/IP
1.2 Designing Applications For A Distributed Environment
1.3 Standard And Nonstandard Application Protocols
1.4 An Example Of Standard Application Protocol Use
1.5 An Example Connection
1.6 Using TELNET To Access An Altemative Service
1.7 Application Protocols And Software Flexibility
1.8 Viewing Services From The Provider's Perspeclive
1.9 The Remainder Of The Text
1.10 Summary
Chapter 2 The Client Server Model And Software Design
2.1 Introduction
2.2 Molivation
2.3 Terminology And Concepts
2.3.1 Clients And Servers
2.3.2 Privilege And Complexity
2.3.3 Standard Vs. Nonstandard Client Software
2.3.4 Pardmeterization Of Clients
2.3.5 Connectionless Vs. Connection-Oriented Servers
2.3.6 Stateless Vs. Stateful Servers
2.3.7 A Stateful File Server Example
2.3.8 .Statelessness Is A Protocol Issue
2.3.9 Servers As Clients
2.4 Summary
Chapter 3 Concurrent Processing In Client-Server Software
3.1 Introduction
3.2 Concurrency In Networks
3.3 Concurrency In Servers
3.4 Terminology And Concepts
3.4.1 The Process Concept
3.4.2 Threads
3.4.3 Programs vs. Threads
3.4.4 Procedure Calls
3.5 An Example Of Concurrent Thread Creation
3.5.1 A Sequential C Example
3.5.2 A Concurrent Version
3.5.3 Timeslicing
3.6 Diverging Threads
3.7 Context Switching And Protocol Software Design
3.8 Concurrency And Asynchronous 1/o
3.9 Concurrency Under UNIX
3.10 Execuling A Separately Compiled Program
3.11 Summary
Chapter 4 Program Interface To Protocols
4.1 Introduction
4.2 Loosely Specified Protocol Software Interface
4.2.1 Advantages And Disadvantages
4.3 Interface Functionality
4.4 Conceptua! Interface Speclfication
4.5 Implementation Of An APl
4.6 Two Basic Approaches To Network Communicatwn
4.7 The Basic 1/0 Functions Available In ANSI C
4.8 Hislory Of The UNIX Socket APl
4.9 Summary
Chapter 5 TheSocketAPI
5.1 Introduction
5.2 The History Of Sockets
5.3 Speifving A Protocol Interface
5.4 The Socket Ahstraction
5.4.1 Sockel Descriptors
5.4.2 System Data Structures For Sockets
5.4.3 Using Sockets
5.5 Specifying An Endpoint Address
5.6 A Generic Address Slructure
5.7 Functions In The Sockel API
5.7.1 The WSAStartup Function
5.7.2 The WSACleanup Function
5.7.3 The Socket Function
5.7.4 The Connect Function
5.7.5 The Send Function
5.7.6 The Recv Function
5.7.7 The Closesocket Function
5.7.8 The Bind Function
5.7.9 The Listen Function
5.7.10 The Accept Function
5.7.11 Summary Of Socket Calls Used Wilh TCP
5.8 Utility Roulines For Integer Conversion
5.9 Using Socket Culls In A Program
5.10 Symbolic Constants For Socket Call Parameters
5.11 Summary
Chapter 6 Algorithms And Issues In Client Software Design
6.1 Inlroduclion
6.2 Leurning, Atgorilhms Instead Of Delails
6.3 Client Architecture
6.4 Idenlifying The Location OfA Server
6.5 Parsing An Address Argumenl
6.6 Looking Up A Domain Name
6.7 Looking Up A Well-Known Port By Name
6.8 Port Numbers And Network Byte Order
6.9 Looking Up A Protocol By Name
6.10 The TCP Client Algorithm
6.11 Alloating A Socket
6.12 Choosing A Local Protocol Port Number
6.13 A Fundamenlal Problem In Choosing A Local IP Address
6.14 Connecting A TCP Socket To A Server
6.15 Commuiucating With The Server Using TCP
6.16 Reading A Response From A TCP Connection
6.17 Closing A TCP Connection
6.17.1 The Need For Partial Close
6.17.2 A Partial Close Operation
6.18 Programming A UDP Client
6.19 Coimected And Unconnected UDP Sockets
6.20 Using Connecl With UDP
6.21 Communicating With A Server Using UDP
6.22 Closing A Socket That Uses UDP
6.23 Partial Close For UDP
6.24 A Wai-ning About UDP Unreliability
6.25 Summarv
Chapter 7 Example Client Software
7.1 ntroduction
7.2 T he Imfwrtance Of Small Examples
7.3 Hiding Dpltails
7.4 An Example Pmcedure Library For Client Programs
7.5 Implementiation OfConTCP
7.6 Implementation Of ConUDP
7.7 A Procedure That Forms Connections
7.8 Using The Example Library
7.9 The OA YTIME Service
7.10 Implemenuation OfA TCP Client For DA YTIME
7.11 Reading From A TCP Connection
7.12 The TIME Service
7.13 Accessing The TIME Service
7.14 Accurale Times And Network Delays
7.15 A UDP Client For The TIME Service
7.16 The ECHO Service
7.17 A TCP Client For The ECHO Service
7.18 A UDP Client For The ECHO Service
7.19 Summary
Chapter 8 Algorithms And Issues In Server Software Design
8.1 Inlroduction
8.2 The Conceptual Senver Algorithm
Concurrent Vs. Iferative Servers
Connection-Oriented Vs. Connectionless Access
Connection-Oriented Servers
Connectionless Servers
Failure, Reliability, And Statelessness
Optimizing Stateless Servers
Four Basil' Types Of Servers
Request Processing Time
llerative Server Algorithms
An Iterative, Connection-Oriented Server Algorithm
Binding To A Well-Known Address Using INADDR_ANY
Placing The Socket In Passive Mode
Accepling Connections And Using Them
An Iterative, Connectionless Server Algorithm
Forming A Reply Address In A Connectionless Server
Concurrent Server Algorithms
Masler And Slave Threads
A Concurrent Connectionless Server Algorithm
A Concurrent, Connection-Oriented Server Algorithm
Using Separate Programs As Slaves
Apparenl Concurrency Using A Single Thread
When To Use Each Server Type
A Summary of Server Types
The Important Problem Of Server Deadlock
Alternative Implementations
Summary
Chapter 9 Iterative, Connectionless Servers (UDP)
9.1 Introduction
9.2 Creating A Passive Socket
9.3 Thread Structure
9.4 An Example TIME Server
9.5 Summary
Chapter 10 Iterative, Connection-Oriented Servers (TCP)
10.1 Introduction
10.2 Allocating A Pcassive TCP Socket
10.3 A Server For The DAYTIME Service
10.4 Thread Structure
10.5 An Example DA YTIME Server
10.6 Closing Connections
10.7 Conneclion Termination And Server Vulnerability
10.8 Summary
Chapter 11 Concurrent, Connection-Oriented Servers (TCP)
11.1 Introduction
11.2 Concurrent ECHO
11.3 Iterative Vs. Concurrent Implementations
11.4 Thread Slructure
11.5 An Example Concurrent ECHO Server
11.6 Summary
Chapter 12 Singly-Threaded, Concurrent Servers (TCP)
12.1 Introduction
12.2 Data-driven Processing In A Server
12.3 Data-Driven Processing With A Single Thread
12.4 Thread Structure OfA Singly-Threaded Server
12.5 An Example Singly-Threaded ECHO Server
12.6 Summary
Chapter 13 Multiprotocol Servers (TCP, UDP)
13.1 Introduction
13.2 The Motivation For Reducing The Number OfServers
13.3 Multiprotocol Server Design
13.4 Thread Structure
13.5 An Example Multiprotocol DAYTIME Server
13.6 The Concept Of Shared Code
13.7 Concurrent Multiprolocol Servers
13.8 Summary
Chapter 14 Multiservice Servers (TCP, UDP)
14.1 Introduction
14.2 Consolidaling Servers
14.3 A Conneclionless, Multiservice Server Design
14.4 A Connection-Oriented, Multiservice Server Design
14.5 A Concurrent, Connection-Oriented, Multiservice Server
14.6 A Singly-Threaded, Multiservice Server Implementation
14.7 Invoking Separate Programs From A Multiservice Server
14.8 Multiservice, Multiprotocol Designs
14.9 An Example Multiservice Server
14.10 Static and Dynamic Server Configuration
14.11 An Example Super Server, Inetd
14.12 Summary
Chapter 15 Uniform, Efficient Management Of Server Concurrency
15.1 Introduction
15.2 Choosing Between An Iteralive And A Concurrent Design
15.3 Level Of Concurrency
15.4 Demand-Driven Concurrency
15.5 The Cost Of Concurrency
15.6 Overhead And Delay
15.7 Small Delays Can Matter
15.8 Thread Preallocation
15.8.1 Preallocation Techniques
15.8.2 Preallocation In A Connection-Oriented Server
15.8.3 Preallocation In A Connectionless Server
15.8.4 Preallocation, Bursty Traffic, And NFS
15.8.5 Preallocation On A Multiprocessor
15.9 Delayed Thread Allocation
15.10 The Uniform Basis For Both Techniques
15.11 Combining Techniques
15.12 Summary
Chapterl6 Concurrency In Clients
16.1 Introduction
16.2 The Advantages Of Concurrency
16.3 The Motivation For Exercising Control
16.4 Concurrenl Contact With Multiple Servers
16.5 Implemenling Concurrent Clients
16.6 Singly-Threaded Implementations
16.7 An Example Concurrent Client That Uses ECHO
16.8 Execution OfThe Concurrent Client
16.9 Managing A Timer
16.10 Example Output
16.11 Concurrency In The Example Code
16.12 Summary
Chapter 17 Tunneling At The Transport And Application Levels
17.1 Introduction
17.2 Multiprotocol Environments
17.3 Mixing Network Technologies
17.4 Dynamic Circuit Allocalion
17.5 Encapsulation And Tunneling
17.6 Tunneling Through An IP Intemet
17.7 Applicalion-Level Tunneling Between Clients And Servers
17.8 Tunneling, Encapsulation, And Dialup Phone Lines
17.9 Summary
Chapter 18 Appiication Level Gateways
18.1 Inlroduction
18.2 Clients And Servers In Constrained Environments
18.2.1 The Reality Of Multiple Technologies
18.2.2 Computers With Limited Functionality
18.2.3 Connectivity Constraints That Arise From Security
18.3 Using Application Gateways
18.4 Interoperability Through A Mail Gateway
18.5 Implementation Of A Mail Gateway
18.6 A Comparison Of Application Gateways And Tunneling
18.7 Application Gateways And Limited Functionality Systems
18.8 Application Gateways Used For Security
18.9 Application Gateways And The Extra Hop Problem
18.10 An Example Application Gateway
18.11 Delails OfA Web-Based Application Gateway
18.12 Invoking A CGI Program
18.13 URLs For The RFC Application Gateway
18.14 A General-Purpose Application Gateway
18.15 Operation Of SLIRP
18.16 How SLIRP Handles Connections
18.17 IP Addressing And SURP
18.18 Summary
Chapter 19 External Data Representation (XDR)
19.1 Introduction
19.2 Representations For Data In Computers
19.3 The N-Squared Conversion Problem
Network Standard Byte Order
A De Facto Standard External Data Representation
XDR Datu Types
Implicil Types
Software Supporl For Using XDR
XDR Library Routines
Building A Message One Piece At A Time
Conversion Routines In The XDR Library
XDR Streams. 1/0, and TCP
Records, Record Boundaries, And Datagram 1/0
Summary
Chapter 20, Remote Procedure Call Concept (RPC)
20.1 Introduction
20.2 Remote Procedure Call Model
20.3 Two Paradigms For Building Distributed Programs
20.4 A Conceptual Modet For Conventional Procedure Calls
20.5 An Extension Of the Procedural Model
20.6 Execution Of Conventional Procedure Call And Return
20.7 The Procedural Model In Distributed Systems
20.8 Analogy Between Client-Server And RPC
20.9 Distributed Computation As A Program
20.10 Sun Microsystems' Remote Procedure Call Definition
20.11 Remote Programs And Procedures
20.12 Reducing The Number Of Arguments
20.13 Identifying Remote Programs And Procedures
20.14 Accommodating Multiple Versions OfA Remote Program
20.15 Mutual Exclusion For Procedures In A Remote Program
20.16 Communication Semantics
20.17 At Least Once Semantics
20.18 RPC Relransmission
20.19 Mapping A Remote Program To A Protocol Port
20.20 Dynamic Porl Mapping
20.21 RPC Port Mapper Algorithm
20.22 RPC Message Format
20.23 Marshaling Arguments For A Remote Procedure
20.24 Authentication
20.25 An Example Of RPC Message Representation
20.26 An Example OfAn Authentication Field
20.27 Summary
Chapter 21 Distributed Program Generation (Rpcgen Concept)
21.1 Introduction
21.2 Using Remote Procedure Calls
21.3 Programming Mechanisms To Support RPC
21.4 Dividing A Program Inlo Local And Remote Procedures
21.5 Adding Code For RPC
21.6 Stub Procedures
21.7 Mulliple Remote Procedures And Dispatching
21.8 Name Of The Client-Side Slub Procedure
21.9 Using Rpcgen To Generate Distributed Programs
21.10 Rpcgen Output And Interface Procedures
21.11 Rpcgen Input And Output
21.12 Using Rpcgen To Build A Client And Server
21.13 Summary
Chapter 22 Distributed Program Generation (Rpcgen Example)
22.1 Introduction
22.2 An Example To lllustrate Rpcgen
22.3 Diclionary Look Up
22.4 Eight Steps To A Distributed Application
22.5 Slep 1: Build A Conventional Application Program
22.6 Step 2: Divide The Program Into Two Parts
22.7 Step 3: Create An Rpcgen Specification
22.8 Step 4: Run Rpcgen
22.9 The h File Produced By Rpcgen
22.10 The XDR Conversion File Produced By Rpcgen
22.11 The Client Code Produced By Rpcgen
22.12 The Server Code Produced By Rpcgen
22.13 Step 5: Write Stub Interface Procedures
22.13.1 Client-Side Inlerface Routines
22.13.2 Server-Side Interface Routines
22.14 Step 6: Compile And Link The Client Program
22.15 Slep 7: Compile And Link The Server Program
22.16 Step 8: Start The Server And Execute The Client
22.17 Summarv
Chapter 23 Network File System Concepts (NFS)
23.1 Introduction
23.1 Remote File Access Vs. Transfer
23.3 Operations On Remote Files
23.4 File Access Among Heterogeneous Computers
23.5 Stateless Servers
23.6 NFS And UNIX File Semanlics
23.7 Review Of The UNIX File System
23.7.1 Basic Definitions
23.7.2 A Byte Sequence Without Record Boundaries
23.7.3 A File 's Owner And Group Identifiers
23.7.4 Protection And Access
23.7.5 The UNIX Open-Read-Write-Close Paradigm
23.7.6 UNIX Data Transfer
23.7.7 Permission To Search A Directory
23.7.8 UNIX Random Access
23.7.9 Seeking Beyond The End Of A UNIX File
23.7.10 UNIX File Position And Concurrent Access
23.7.11 Semantics Of Write During Concurrent Access
23.7.72 UNIX File Names And Paths
23.7.13 The UNIX tnode: Information Stored With A File
23.7.14 The UNIX Stat Operation
23.7.15 The UNIX File Naming Mechanism
23.7.16 UNIX File System Mounts
23.7.17 UNIX File Name Resolution
23.7.18 UNIX Symbolic Links
23.8 Files Under NFS
23.9 NFS File Types
23.10 NFS File Modes
23.11 NFS File Attributes
23.12 NFS Client And Server
23.13 NFS Client Operation
23.14 NFS Client And UNIX
23.15 NFS Mounts
23.16 File Handle
23.17 NFS Handles Replace Path Names
23.18 An NFS Client Under Windows
23.19 File Positioning With A Stateless Server
23.20 Operations On Directories
23.21 Reading A Directory Slatelessly
23.22 Mulliple Hierarchies In An NFS Server
23.23 The Mount Protocol
23.24 Summary
Chapter 24 Network File System Protocol (NFS, Mount)
24.1 Inlroduction
24.2 Using RPC To Define A Protocol
24.3 Defining A Protocol With Data Structures And Procedures
24.4 NFS Conslanl, Type, And Data Declarations
24.4.1 NFSConstants
24.4.2 NFS Typedef Declarations
24.4.3 NFS Data Structures
24.5 NFS Procedures
24.6 Semantics Of NFS Operations
24.6.1 NFSPROC_NULL (Procedure 0)
24.6.2 NFSPROC_GETA TTR (Procedure 1)
24.6.3 NFSPROC_SETATTR (Procedure 2)
24.6.4 NFSPROC_ROOT (Procedure 3) [Ohsolete in NFS3]
24.6.5 NFSPROC_OOKUP (Procedure 4)
24.6.6 NFSPROC_READLlNK (Procedure 5)
24.6.7 NFSPROC_READ (Procedure 6)
24.6.8 NFSPROC_WRITECACHE (Procedure 7) [Obsolele in NFS3]
24.6.9 NFSPROC_WRITE (Procedure 8)
24.6.10 NFSPROC_REATE (Procedure 9)
24.6.11 NFSPROC_REMOVE (Procedure 10)
24.6.12 NFSPROC_RENAME (Procedure 11)
24.6.13 NFSPROC_LINK (Procedure 12)
24.6.14 NFSPROC_SYMUNK (Procedure 13)
24.6.15 NFSPROC_MKDIR (Procedure 14)
24.6.16 NFSPROC_RMD1R (Procedure 15)
24.6.17 NFSPROC_READDlR (Procedure 16)
24.6.18 NFSPROC_STATFS (Procedure 17)
24.7 The Mount Protocol
24.7.1 Mount Constant Definitions
24.7.2 Mount Ty'pe Definitions
24.7.3 Mount Data Structures
24.8 Procedures In The Mount Protocol
24.9 Semantics of Mount Operations
24.9.1 MNTPROC_ULL (Procedure 0)
24.9.2 MNTPROC_MNT (Procedure 1)
24.9.3 MNTPROC_DUMP (Procedure 2)
24.9.4 MNTPROC_MNT (Procedure 3)
24.9.5 MNTPROC_UMNTALL (Procedure 4)
24.9.6 MNTPROC_EXPORT (Procedure 5)
24.10 NFS And Moimt Authentication
24.11 Changes In NFS Version 3
24.12 Summarv
Chapter 25 A TELNET Client (Program Structure)
25.1 Introduction
25.2 Overview
25.2.1 The User's Terminal
25.2.2 Command And Control Information
25.2.3 Tenninals, Windows, and Files
25.2.4 The Need For Concurrency
25.2.5 A Thread Model For A TELNET Client
25.3 A TELNET Client Algorithm
25.4 Keyboard 1/0 In Windows
25.5 Global Variables Used For Keyboard Control
25.6 Initializing The Keyboard Thread
25.7 Finite Stale Machine Specification
25.8 Embedding Commands In A TELNET Data Stream
25.9 Option Negotiation
25.10 Request/Offer Symmetry
25.11 TELNET Character Definitions
25.12 A Finite State Machine For Data From The Server
25.13 Transitions Among States
25.14 A Finite State Machine Implementation
25.15 A Compact FSM Representa'tion
25.16 Keeping The Compact Representation At Run-Time
25.17 Implementation OfA Compact Representation
25.18 Building An FSM Transition Matrix
25.19 The Socket Output Finite State Machine
25.20 Definitions For The Socket Output FSM
25.21 The Option Subnegotiation Finite State Machine
25.22 Definitions For The Option Subnegotiation FSM
25.23 FSM Initialization
25.24 Arguments For The TELNET Client
25.25 The Heart Of The TELNET Client
25.26 TELNET Synchronization
25.27 Handling A Severe Error
25.28 Implemenlation Of The Main FSM
25.29 A Procedure For Immediate Disconnection
25.30 Abort Procedure
25.31 Summary
Chapter 26 A TELNET Client (Implementation Details)
26.1 Introduction
26.2 The FSM Action Procedures
26.3 Recording The Type Of An Option Requesl
26.4 Performing No Operation
26.5 Responding To WILLWONT For The Echo Oplion
26.6 Sending A Response
26.7 Responding To WILUWONT For Unsupported Options
26.8 Responding To WLLWONT For The No Go-Ahead Option
26.9 Generating DO/DONT For Binary Transmission
26.10 Responding To DO/DONT For Unsupporled Options
26.11 Responding To DO/DONT For Transmit Binary Option
26.12 Responding To DO/DONT For The Terminal Type Option
26.13 Option Subnegoliation
26.14 Sendmg Terminal Type Information
26.15 Tennincning Suhnegotiation
26.16 Sending A Character To The Server
26.17 Displaying Incoming Data On The User's Terminal
26.18 Writing A Block Of Data To The Server
26.19 Interacting With The Local Client
26.20 Responding To lllegat Commands
26.21 Scripting To A File
26.22 Implementation Of Scripting
26.23 Initialzation Of Scripting
26.24 Collecting Characters OfThe Script File Name
26.25 Opening A Script File
26.26 Terminating Scripting
26.27 Printing Status Information
26.28 Summarv
Chapter 27 Porting Servers From UNIX To Windows
27.1 Introduction
27.2 Operating in Background
27.3 Shared Descriptors And Inherilance
27.4 The Controlling TTY
27.5 Working Directories
27.6 File Creattion And Umask
27.7 Process Groups
27.8 Descriptors For Slandard 1/0
27.9 Mutual Exclusion For A Server
27.70 Recording A Process ID
27.11 Waiting For A Child Process To Exit
27.12 Using A Syslem Log Facility
27.12.1 Generating Log Messages
27.13 Miscellaneous Incompatibilities
27.14 Summarv
Chapter 28 Deadlock And Starvation In Client-Server Systems
28.1 Introduction
28.2 Definition Of Deadlock
28.3 Difficulty Of Deadlock Detection
28.4 Deadlock Avoidance
28.5 Deadlock Between A Client And Server
28.6 Avoiding Deadlock In A Single Interaction
28.7 Starvalion Among A Set Of Clients And A Server
28.8 Busy Connections And Starvation
28.9 Avoiding Blocking Operations
28.10 Threads, Connections, And Other Limits
28.11 Cycles Of Clients And Servers
28.12 Documenting Dependencies
28.13 Summary
Appendix 1 Functions And Library Routines Used With Sockets
Appendix 2 Manipulation Of Windows Socket Descriptors
Bibliography
Index
· · · · · · (收起)
读后感
用户评价
从专业深度来看，我必须说，这本书的论述层次是相当扎实的，它绝非市面上那些浮于表面的“速成指南”。它似乎是为那些不满足于仅仅“会用”网络协议，而渴望“精通”其底层机制的工程师准备的。作者对路由协议——特别是 OSPF 和 BGP 的讲解，达到了教科书级别的严谨。在讨论 OSPF 的 LSA 类型的层次时，它没有停留在对 RFC 中定义的简单复述，而是深入剖析了每种 LSA 在不同区域类型中扮演的具体角色，以及它们如何协同工作来维持整个路由域的一致性，这对于理解大型企业网络或 ISP 网络的稳定性至关重要。更让我印象深刻的是，书中对流量工程和 QoS 策略的探讨，这部分内容在很多基础读物中往往是一笔带过。但在这里，作者详细拆解了 DiffServ 模型的实现原理，并结合实际案例说明了如何通过 DSCP 标记来区分关键业务流，这显示出作者拥有丰富的实战经验。阅读这些章节时，我感觉自己不再是单纯在学习“技术规范”，而是在学习如何运用这些规范去解决真实世界中复杂的网络性能优化问题，这种实用性是无价的。
这本书的封面设计充满了复古的魅力，那种深蓝色的背景搭配着烫金的字体，一下子就把我的思绪拉回到了那个网络技术刚刚兴起的黄金年代。拿到手里的时候，就能感觉到纸张的厚重和质感，这让我对书中的内容充满了期待。我本来以为这会是一本枯燥的技术手册，但翻开目录后发现，作者的叙事方式非常巧妙。他没有直接堆砌那些晦涩难懂的协议栈定义，而是选择了一个更具故事性的切入点，比如从早期的 ARPANET 讲起，一步步构建起我们今天所依赖的庞大网络体系。这种编排方式对于我这种初学者来说简直是福音，它让我能够在一个更宏观的视角下去理解各个技术点之间的逻辑关系，而不是像看字典一样零散地记忆知识。尤其是一些关于早期网络拓扑和路由选择的描述，作者运用了生动的比喻，让那些抽象的概念瞬间变得鲜活起来。我特别喜欢其中穿插的一些历史轶事，它们不仅仅是背景知识，更是理解技术演进脉络的钥匙。比如，关于 TCP 拥塞控制算法的几次重大迭代，作者没有仅仅罗列公式，而是探讨了每一次修改背后的驱动力和实际应用中的挑战，这使得我对“为什么是这样设计”有了更深层次的理解。整本书的阅读体验非常流畅，让人有种“停不下来”的感觉，仿佛在跟随一位经验丰富的向导，深入探索这个复杂而迷人的数字世界。
对于寻求动手实践的读者而言，这本书的配套资源和案例分析环节是其最大的亮点之一。作者深知，网络技术是“做出来”而不是“读出来”的。书中大量的配置示例，特别是针对主流厂商设备（虽然没有直接提及品牌名称，但其结构很容易映射到行业标准设备上）的配置片段，清晰地展示了如何将理论转化为实际操作。更妙的是，很多配置后面都附带了“验证”步骤，即使用户在自己的实验环境中模拟配置后，也清楚地知道应该用哪些命令（如 `show ip route` 或特定的抓包分析工具的输出结果）来确认配置是否生效且正确。这种“理论—配置—验证”的闭环学习模式极大地增强了读者的信心。我个人尤其欣赏它在故障排除章节的处理方式。作者没有提供标准的故障排除流程清单，而是设计了一系列“症状描述”，然后引导读者像侦探一样，通过分析 ARP 缓存、ICMP 响应时序和 TTL 值的变化，逐步锁定问题的根源。这种引导式的、启发性的教学方法，远比死记硬背“先查 L1，再查 L2……”的死板步骤来得有效和深刻。
这本书的排版和内文的细节处理，显示出出版方在专业书籍制作上的匠心。字体选择上，正文采用了一种非常清晰易读的宋体变种，而在关键的技术术语和代码示例部分，则果断切换为等宽字体，并且用加粗或斜体进行了明确区分，这极大地减轻了阅读时的视觉疲劳。更值得称赞的是图表的质量。我阅读过很多技术书籍，很多插图都模糊不清，或者为了节省篇幅而过度简化，导致关键信息丢失。然而，这本书中的每一张网络结构图、数据流示意图，都精确到了像素级别，线条清晰，色彩搭配得当，即便是那些复杂的七层协议模型图，也能让人一眼看清数据包的封装与解封装过程。特别是关于数据链路层帧结构和 IP 头部解析的插图，作者甚至细心地用不同颜色标注了各个字段的含义，这种细致入微的设计，让我不再需要频繁地在文字和图表之间来回跳转查找，极大地提升了学习效率。对于需要对照 RFC 标准进行深入研究的读者来说，书中对关键字段的十六进制表示和二进制位的解析也做得非常到位，这些微小的细节，共同铸就了一本真正服务于读者的高质量工具书。
这本书的写作风格透露出一种独特的、近乎哲学思辨的视角，这让我对网络的看法产生了微妙的转变。作者似乎非常热衷于探讨“分布式系统的本质”这一宏大命题，并将其与 TCP/IP 协议的设计哲学紧密结合起来。他没有把协议栈视为一堆静态的代码规则，而是将其描绘成一个在持续博弈、不断妥协中达成的“君子协定”。例如，在讨论 TCP 的三次握手与四次挥手时，他不仅解释了 ACK 包和 FIN 包的流转，更深入地探讨了这种设计如何巧妙地平衡了“连接的可靠性”与“系统资源的释放效率”之间的矛盾。这种“平衡的艺术”贯穿全书，使得原本冰冷的技术术语染上了一层人文色彩。在谈到 IP 地址的分配和子网划分时，作者没有仅仅停留在数学计算层面，而是将其置于全球资源管理的宏大背景下，讨论了 IPv4 地址枯竭的必然性和向 IPv6 迁移的紧迫性，这让读者在学习具体技术的同时，也能培养起对整个互联网生态的责任感和前瞻性思维。这种将技术与更广阔的工程伦理、管理视野相结合的叙事，远超出了我对一本纯技术书籍的期待。
