4 ms·
> If all of these routers adopted this, my lowly traffic would have to read from main memory on each route lookup... yes, but current COTS hardware, via usersp
by signa11 10y ago
> If all of these routers adopted this, my lowly traffic would have to read from main memory on each route lookup...
yes, but current COTS hardware, via userspace networking can very easily saturate a 10g interface at 64b packet-size / core.
imho, it is the convergence of a large number of advances in COTS hardware over the last couple of years, that has changed the field to mostly a software problem rather than a combination of hardware + software one. when you have proprietary hardware, software tends to be more or less proprietary, which is where most of the networking vendors are currently.
a large number of projects are actively attacking this space e.g. snabb(https://github.com/lukego/snabb https://github.com/lukego/snabb), routebricks (http://routebricks.org/ http://routebricks.org/) etc.
- amazon_not 10y ago> yes, but current COTS hardware, via userspace networking can very easily saturate a 10g interface at 64b packet-size / core. Citation needed.
- feld 10y agoHeard of Netmap or DPDK?
- amazon_not 10y agoYes. As far as I know neither of them can push 15Mpps per core. Please please prove me wrong. I'd love to be.
- tyingq 10y agoIt's also comparing basically an assembled automobile with a pile of parts. I suspect there will be a well vetted router-centric distribution that leverages DPDK soon, but I don't think that exists yet.
- bogomipz 10y agoHuh? All 10 Gig NICs can and do push 14 Mpps using minimum MTU: http://www.ieee802.org/3/10G_study/public/speed_adhoc/email/msg00021.html http://www.ieee802.org/3/10G_study/public/speed_adhoc/email/... What are you talking about?
- signa11 10y ago> Citation needed. http://lmgtfy.com/?q=dpdk+15mpps http://lmgtfy.com/?q=dpdk+15mpps
- reitanqild 10y agoI hate lmgtfy used for this. Possibly more than Citation needed.
- Cerium 10y agoI wouldn't say that they easily reach 10g on small packets. I got to about 8g before I started to have trouble. It certainly takes some optimization. If I remember correctly Intel's DPDK documents warn that you only have about 80 cycles between packets.
- signa11 10y agoHmm, for our case I 'program' flows which then end up being used by per-core tasks for forwarding decisions, and with couple of tweaks etc was able to handle line rate across 4 10g interfaces for 64b packets, with one physical core per port. This is on bare metal though, so might be different experience if you are running stuff on vm's where without sr-iov you would have multiple layers of indirection + context switches before someone can actually do something meaningful to the pdu... edit: slight clarification added
- netten 10y agocan you tell me how you did this, i would really like to try this out.