Re:A Tour Inside CloudFlare's Latest Generation Servers
by ihsw on Monday July 22, 2013 @04:05PM
(#6086144)
This sort of scaling seems to be logarithmic. 16MB is 5 orders of magnitude greater than 512KB, however the logarithmic difference is 3.46. The same increase in 5 orders of magnitude beyond 16MB is 10.39 (at 524GB) which is considerably more impressive.
That's some fancy math and it probably is completely off, but another explanation is NIC performance doesn't follow exponential scaling for unexplained reasons.
Re:A Tour Inside CloudFlare's Latest Generation Servers
by hosay123 on Monday July 22, 2013 @05:17PM
(#6086513)
The article suggests interrupt load is the major issue, although doesn't really say enough to inform us. A 12.5ms buffer means potentially just 80 wakeups/sec. They don't mention whether they use multiple hardware queues or if all interrupts hit a single core.
I'd also be interested to know if polling mode was tested with any of the cards, and why it didn't work out
Re:A Tour Inside CloudFlare's Latest Generation Servers
by eastdakota on Monday July 22, 2013 @05:56PM
(#6086725)
We use multiple hardware queues. Some details in this old blog post:
http://blog.cloudflare.com/how-the-cloudflare-team-got-into-...
We're continuing to experiment with polling-based network queues.
16MB vs 512KB is only larger? Not gigantic or even extremely large in comparison to the extremely tiny cache? Oh dear.