hannesgredler

3 ms·

1.
by hannesgredler 10y ago
netflix caches are serving millions of subscribers using a thousands to ten-thousands of prefixes, well below 100K. their caches are highly regionalized so its no practical problem.
2.
by hannesgredler 10y ago
Its a 3TBits/s chipset from Broadcom https://www.broadcom.com/products/ethernet-communication-and...
3.
by hannesgredler 10y ago
we started integrating it in March. Arguable the compile chain is a bit heavy weight, but the code is very structured, easy to extend and a clean architecture. - Dave Barach and the VPP crew were very quick answering questions we have. BTW
4.
by hannesgredler 10y ago
thanks, i have taken the liberty to C&P your explanation into the blog.
5.
by hannesgredler 10y ago
history traffic patterns show clearly a rising in-equality/ power-law distribution. so unless that multi-year (decade) long trend reverses, forwarding lookup hierarchies are going to work even better.
6.
by hannesgredler 10y ago
where theory hits reality is when those whiteboxes and their routing-stacks encounter the 600K routes coming from 40 different exits. still no viable solution out there today- quagga, bird, any ?
7.
by hannesgredler 10y ago
good guess :-) - our MLP (minimum lovable product) is BGP and IS-IS along with a VPP based software forwarding module.
8.
by hannesgredler 10y ago
apologies ! - get them constantly wrong in german and so do i in english.
9.
by hannesgredler 10y ago
your reasoning is correct (the path latency gets higer) its just that you got the final numbers wrong ;-) on modern software forwarding core (fd.io/VPP and DPDK) you can forward with a sub 100us latency. so your total latency ends up b