Great article! Simple to digest. Can you give us an idea of the scale of requests and latency that these root servers typically deal with? If there are just 13 in the world, I'm curious as to what sort of mechanisms are in place to ensure they don't get overwhelmed.
Great article! Simple to digest. Can you give us an idea of the scale of requests and latency that these root servers typically deal with? If there are just 13 in the world, I'm curious as to what sort of mechanisms are in place to ensure they don't get overwhelmed.
Great article! Simple to digest. Can you give us an idea of the scale of requests and latency that these root servers typically deal with? If there are just 13 in the world, I'm curious as to what sort of mechanisms are in place to ensure they don't get overwhelmed.
hey Shalini, good question! I should've instead written, '13 root server clusters'.
Here are some ways the root servers scale:
- they have replicas: 1,936 root server instances across the world
- they route the requests to the closest available server (via anycast routing)
- caching values at different client levels before the root server reduces the traffic
I found a site to track root servers: https://root-servers.org/
The client request uses UDP for low overhead and fast response. Hope it helps!
Thank you so much Neo!