Caching 101: The Basics

Caching is one of 1999's hot technologies. Here's a look at caching fundamentals: How and why it works, comparison to alternative solutions, and how and where to deploy your cache.

by Amit Pandey
Director of NetCache, Network Appliance, Inc.

Q: Why would ISPs want to deploy web caches?

A: Web caches bring two main benefits to an ISP: improved response times and more efficient use of bandwidth.

Deploying a cache significantly reduces the "world wide wait" problem-by storing Web objects closer to end users. If the requested objects are in the cache, they get the information almost instantaneously, while requests that have to go to the origin server typically take at least several seconds to be fulfilled.

Second, caches reduce traffic and will ultimately save ISPs money by decreasing their bandwidth costs by as much as 50 percent. When users get objects from caches, they do not use as much bandwidth as if the object came from the origin server. This is crucial to ISPs as they grow their business and expand their customer base.

Q: What determines the extent of the benefits of caching?

A: The "hit rate" (requested content in cache divided by total requested content) is the primary driver both for improvements in average response time and for bandwidth savings. Hits are served instantaneously and do not consume bandwidth back to the origin server. Misses, on the other hand, take several seconds to be served and consume bandwidth.

Q: What are typical hit rates?

A: A well designed cache achieves hit rates of 30 to 60 percent of Web traffic, depending on size and diversity of the end user community, the size of the cache and the caching algorithms used. For NNTP traffic hit rates can reach 90 percent.

continued—