The transmission control protocol (TCP) is vital to communications networks. It has been in existence since the early stages of computer networking and has proved beneficial in attaining reliable communication over computer networks for years. However, with the evolution of data communications technologies, certain design characteristics of TCP also need to be changed.

TCP was originally developed for wired networks and worked well with moderate Ethernet speeds of a few Mbps. The way TCP handles network congestion and packet drops in different network environments has attracted the attention of the research community. For example, TCP assumes that the basic reason for packet drop in a network is congestion. However, in wireless networks, the reason for packet drops could be nodes moving out of communication range due to node mobility. Therefore, the way TCP handles network congestion needs to change with changing network conditions. Similarly, with the evolution of cloud computing and high-speed data center networks, the demand for high-speed and low-latency applications has grown. These networks use high-speed Gigabit Ethernet to serve such applications. It is necessary to modify the way TCP recovers from data loss due to limited-sized buffers at the switches in these data centers, since they are expected to serve applications demanding high bandwidth and low latency. Researchers have proposed changing the characteristics and behavior of the TCP protocol, giving rise to different versions of TCP. This book focuses on the behavior and analysis of TCP performance in data center networks.

In this book, the authors try to address the TCP incast problem, which degrades the performance of data center networks. This is a small book with only five chapters. In first chapter, the authors introduce the readers to cloud computing, data centers, the use of TCP in data centers, and the TCP incast problem. This chapter is easy to understand and forms the basis for the rest of the book. It emphasizes how the TCP incast problem occurs in data center networks and the need for a solution to this problem. Chapter 2 covers the design and characteristics of TCP. The information presented in this chapter is already available in the literature. Readers who already know the functionality of TCP may choose to skip this part. Chapters 3 and 4 are the core chapters of this book. In chapter 3, the authors present and simulate analytical models for the occurrence of the TCP incast problem in data center networks. In chapter 4, the authors discuss, simulate, and analyze available and proposed methods to overcome the TCP incast problem. The simulations are carried out in NS2. Chapter 5 is short, and summarizes the research work carried out by the authors and suggests future research directions.

This book precisely discusses the performance of TCP in data center networks and methods to overcome the TCP incast problem. In general, the concepts, problems, and solutions related to the performance of TCP in high-speed networks discussed in this book will be of interest to all network engineers and researchers. In particular, this book will be useful to researchers working toward enhancing the performance and efficiency of clouds and data centers designed to serve high-speed, low-latency applications. Readers can learn from the methodology adopted in this book to model and analyze the TCP incast problem. The lists of references at the end of every chapter (except for chapter 5) present good quality research work carried out to study the performance of TCP in computer networks, and should be used for further study.