via Slashdot : "Scientists at the San Diego Supercomputer Center found that 98% of the DNS queries at the root level are unnecessary." At first glance this seems like there is something wrong. 98% is a high percentage of unnecessary hits. But after I rooted through the oh-so-witty discussion on meaningless stats at the beginning of the Slashdot discussion, I came across some insightful posts talking about why this stat might be so high. First off, DNS Servers below the root level cache (save) the successful queries they perform so that if someone looks up the same address again they can respond quicker and save bandwidth. They hold onto these saved entries for some period of time (hours? days? weeks?) before purging and querying again. Queries that are unsuccessful are POSSIBLY cached (not sure about this) but I doubt it. Which means that on the whole the most common type of queries to hit the root servers will be the unsuccessful ones. The successful ones were likely already cached further down the chain. If you combine this with the fact that most modern browsers' URL area doubles as a search field, which means that simple searches are likely first passed through a DNS query, then you can see why even more unsuccessful queries see the light of day. And even further, with the garbage email addresses that people type into online registration forms and the non-existent return email addresses that are bundled into spam, each of these have to query the DNS records to figure out where the mail should go resulting in even more unsuccessful queries hitting the root servers. Mix into that script kiddies and their DOS attacks and the load gets higher. So taking all of that into account it actually isn't that surprising that the vast majority of hits on the root servers are unsuccessful and that the vast majority of the valid queries are being returned from caches. Impressive stats aren't often as impressive once thought through a little.