Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

F-Stack – A network development kit with high performance based on DPDK

f-stack.org

Post reply on HN
Print view View original

11–20 of 32 posts

Re: F-Stack – A network development kit with high performance based on DPDK

#11

Post by touisteur »

ozgrakkurt wrote: post #8
nurumaik wrote: post #3

Earlier quoted context omitted.

Anything more innovative happened since (honestly curious)?

You can apparently do 100gbit/sec on a single thread over ethernet with io uring.

Recently did 400gb/s on a single core / 4x100gb nics (or just the one 400g nic, too) with dpdk. Mind you it's with jumbo frames and constant packet size for hundreds of mostly synchronized streams... You won't process each packet individually, mostly put them in queues for later batch-process by other cores. Amazing for data acquisition applications using UDP streams.

I keep watching and trying io_uring and still can't make it work as fast with simple code as consistently for those use cases. AF_XDP gets me partly there but then you're writing ebpf... might as well go full-dpdk.

Maybe it's a skill issue on my part, though. Or just a well-fitting niche.

Re: F-Stack – A network development kit with high performance based on DPDK

#12

Post by lossolo »

touisteur wrote: post #11
ozgrakkurt wrote: post #8

Earlier quoted context omitted.

You can apparently do 100gbit/sec on a single thread over ethernet with io uring.

Recently did 400gb/s on a single core / 4x100gb nics (or just the one 400g nic, too) with dpdk. Mind you it's with jumbo frames and constant packet size for hundreds of mostly synchronized streams... You won't process each packet individually, mostly put them in queues for later batch-process by other cores. Amazing for data acquisition applications using UDP streams. I keep watching and trying io_uring and still can…

Any numbers for io_uring with 4x100gb nics in your tests?

Re: F-Stack – A network development kit with high performance based on DPDK

#14

Post by Kamillaova »

ozgrakkurt wrote: post #8
nurumaik wrote: post #3

Earlier quoted context omitted.

Anything more innovative happened since (honestly curious)?

You can apparently do 100gbit/sec on a single thread over ethernet with io uring.

Of course, but when working directly with the NIC, such speeds can be achieved with smaller packets, getting even closer to the linerate.

Re: F-Stack – A network development kit with high performance based on DPDK

#15

Post by baruch »

antoinealb wrote: post #5
js4ever wrote: post #4

Earlier quoted context omitted.

I don't think so, but my guess is raw performance rarely matters in the real world. I once explored this, hitting around 125K RPS per core on Node.js. Then I realized it was pointless, the moment you add any real work (database calls, file I/O, etc.), throughput drops below 10K RPS.

The goal of this kind of system is not to replace the application server. This is intended to work on the data plane where you do simple operations but do them many time per second. Think things like load balancers, cache server, routers, security appliances, etc. In this space Kernel Bypass is still very much the norm if you want to get an efficient system.

We do storage systems and use DPDK in the application, when the network IS the bottleneck it is worth it. Saturating two or three 400gbps NICs is possible with DPDK and the right architecture that makes the network be the bottleneck.

Re: F-Stack – A network development kit with high performance based on DPDK

#16

Post by ozgrakkurt »

touisteur wrote: post #11
ozgrakkurt wrote: post #8

Earlier quoted context omitted.

You can apparently do 100gbit/sec on a single thread over ethernet with io uring.

Recently did 400gb/s on a single core / 4x100gb nics (or just the one 400g nic, too) with dpdk. Mind you it's with jumbo frames and constant packet size for hundreds of mostly synchronized streams... You won't process each packet individually, mostly put them in queues for later batch-process by other cores. Amazing for data acquisition applications using UDP streams. I keep watching and trying io_uring and still can…

Sounds super cool but dpdk sounds like it won't be worth the difficulty from what I read so far.

I also want to get into socket io using io_uring in zig. I'll try to apply everything I found in liburing wiki [0] and see how much I can get (max hardware I have is 10gbit/s).

Seems like there is: - multi-shot requests - register_napi on uring instance - zero copy receive/send. (Probably won't be able to get into it)

Did you already try these or are there other configurations I can add to improve it?

[0]: https://github.com/axboe/liburing/wiki/io_uring-and-networki...

Re: F-Stack – A network development kit with high performance based on DPDK

#17

Post by gonzopancho »

When originally published they wouldn’t even ack Patrick Kelsey, the author of libuinet or that they had forked libuinet.

Now they say this: “Thanks to libplebnet and libuinet this work became a lot easier.”

F-stack is literally forked libuinet using DPDK instead of netmap.

The net-net is that Kelsey took his work private and tencent isn’t advancing the work.

Back in the day I was sponsoring work on libuinet in order to move enough of the kernel needed for a security appliance to libuinet to underpin a performance improvement for pfsense.

Then Tencent did what they did, Patrick reacted as he did and that was over.

We pivoted to VPP. But back in 2016 it also needed a lot of work.

Re: F-Stack – A network development kit with high performance based on DPDK

#18

Post by gonzopancho »

eqvinox wrote: post #9
antoinealb wrote: post #5

Earlier quoted context omitted.

The goal of this kind of system is not to replace the application server. This is intended to work on the data plane where you do simple operations but do them many time per second. Think things like load balancers, cache server, routers, security appliances, etc. In this space Kernel Bypass is still very much the norm if you want to get an efficient system.

> In this space Kernel Bypass is still very much the norm if you want to get an efficient system. Unless you can get an ASIC to do it, then the ASIC is massively preferrable; just the power savings generally¹ end the discussion. (= remove most routers from the list; also some security appliances and load balancers.) ¹ exceptions confirm the rule, i.e. small/boutique setups

ASICs require years to develop and aren’t flexible once deployed

Re: F-Stack – A network development kit with high performance based on DPDK

#19

Post by yxhuvud »

nurumaik wrote: post #3
alexdns wrote: post #2

It was considered innovative when it was first shared here eight years ago.

Anything more innovative happened since (honestly curious)?

Well, io_uring came along and removed a lot of the incentive.

Re: F-Stack – A network development kit with high performance based on DPDK

#20

Post by Eduard »

Hacker News is the kind of place where you can have _this_ submission (PRC-sponsored Tencent-owned network devkit) on the front page next to a submission about how PRC-sponsored cybercrime group Salt Typhoon pwned 'nearly every American': https://news.ycombinator.com/item?id=45074157
Post reply on HN