Load Balancing Servers, Firewalls, and Caches

Owing to their ability to scale server farms and improve the security and availability of networked applications, load balancers are expected to grow into a multibillion-dollar market in the next few years. Written from the unique perspective of an industry insider, this book explains how this critical technology can be used to solve several challenges in managing today's server farms, Web sites, firewalls, and caches.

Chandra Kopparapu offers you a close look at how load balancers work and explores the ability of load balancers to solve a multitude of network and server bottlenecks in the Internet age--from dramatic improvements in server farm scalability to removing the firewall as a network bottleneck. This book provides a technical discussion of this fast-growing market, covering the full spectrum of topics, from server and firewall load balancing to transparent cache switching to global server load balancing. In the process, he delivers insight into the way new technologies are deployed in network infrastructure and how they work. In addition, he reviews such topics as:

• The fundamentals of how load balancers work--covering packet flow through a load balancer and anatomy of a session table
• Server health checks and monitoring, load distribution methods, stateful and stateless load balancing, network address translation, and direct server return
• Advanced server load balancing topics, which ensure session persistence for shopping cart applications, content switching based on cookies or URLs, the mega-proxy problem and ways to deal with it, and handling HTTP to SSL transition for e-commerce applications
• Designing networks for high availability using load balancers in active-standby or active-active design, ways to address high availability, and handling servers with two network adapters in active-active or active-standby mode
• How to maintain Web sites at multiple data centers and direct the users to a data center that provides the best response time, or alternative data centers in case of a failure
• Removing firewalls as a single point of failure and a performance bottleneck, high availability designs for Layer 2 or Layer 3 firewalls, and stateful failover in case of a firewall or load balancer failure
• Making efficient use of caches by using a load balancer for switching based URLs, deploying caches as transparent proxy, forward proxy, or reverse proxy

About the Author

Chandra Kopparapu is Director of Product Marketing at Foundry Networks, a leading provider of high-performance end-to-end switching solutions for enterprises and service providers. Kopparapu, who has an electrical engineering degree and an MBA in marketing from UC Berkeley, worked at Compaq as Product Manager prior to working at Foundry. He has extensive experience in mission critical systems and high availability as part of his engineering work at Tandem Computers. He has also written for Network World.