Caching Servers

Hostnames are rarely accessed once. For example, if one person visits a Web site, then that person is likely to visit the same site again later. Similarly, if one site is popular, then many people are likely to request the same hostname lookup. Caching DNS servers receive and relay new requests and cache replies for faster access. Although these servers provide unauthenticated results, the results are generated much faster than if the host needed to query a root, ccTLD, SLD, and primary DNS server.

The duration for data to be held within a caching DNS server varies based on the data. Each DNS reply includes a cache timeout duration. The data should be held until the information expires. Usually the duration is between a day and a week. If the timeout value is too low, then the primary and secondary DNS servers must field more requests. In contrast, a timeout value set too large can lead to slow updates; a change to a DNS entry may take days to propagate across the Internet.

Most large companies and Internet providers operate caching DNS servers for their customers. This provides hostname resolution results faster than individual queries to the official hosting sites. In addition, caching servers remove network load by limiting the need to send requests to all of the different root, gTLD, ccTLD, SLD, and primary servers.