|
|
|
|
|
|
by hrunt
1343 days ago
|
|
|
I worked for a Usenet provider for a number of years designing and writing the software that powered their Usenet feeding infrastructure. The Usenet feed is basically a giant stream of articles, and we used queuing theory to scale out the systems that read it in, process it, and distribute it to readers. My experience with the formula's is a little different from yours. People thought what I showed them couldn't be true. "If the feed is X articles/second and system's P95 is Y, the backlog will continue to grow," would often be met with, "That can't be true ..." |
|
|