Stories
Slash Boxes
Comments

SoylentNews is people

posted by janrinok on Tuesday January 24 2017, @11:27PM   Printer-friendly
from the fast,-faster,-fastest dept.

We have to trade faster!

Solarflare Communications is an unheralded soldier in the eternal war on latency. With its founding in 2001, Solarflare took on the daunting raison d'être of grinding down latency from one product generation to the next for the most latency-sensitive use cases, such as high frequency trading. Today, the company has more than 1,400 customers using its networking I/O software and hardware to cut the time between decision and action. In high frequency trading, the latency gold standard is 200 nanoseconds. If you're an equity trader using a Bloomberg Terminal or Thomson Reuters Eikon, latency of more than 200 nanoseconds is considered to be shockingly pedestrian, putting you at risk of buying or selling a stock at a higher or lower price than the one you saw quoted. Now, with its announcement of TCPDirect, Solarflare said it has cut latency by 10X, to 20-30 nanoseconds.

[...] The CTO of an equity trading firm, who agreed to talk with HPCwire's sister [publication] EnterpriseTech anonymously, said his company has been a Solarflare customer for four years and that its IT department has validated Solarflare's claims for TCPDirect of 20-30 nanoseconds latency. He regards Solarflare as a partner that allows his firm to focus on core competencies, rather than devoting in-house time and resources to lowering latency. "It used to be the case that there weren't a lot of commercial, off-the-shelf products applicable to this space," he said. "If one of our competitors wanted to do something like this for competitive advantage, Solarflare can do it better, faster, cheaper, so they're basically disincentivized from doing so. In a sense this is leveling the playing field in our industry, and we like that because we want to do what we're good at, rather than spending our time working on hardware. We're pleased when external vendors provide state-of-the-art technology that we can leverage."


Original Submission

 
This discussion has been archived. No new comments can be posted.
Display Options Threshold/Breakthrough Mark All as Read Mark All as Unread
The Fine Print: The following comments are owned by whoever posted them. We are not responsible for them in any way.
  • (Score: 1) by LordByronStyrofoam on Wednesday January 25 2017, @12:02AM

    by LordByronStyrofoam (5232) on Wednesday January 25 2017, @12:02AM (#458339)

    Signals traverse copper cable at about .64c (where c is the speed of light). That's about 5-8nS per foot.
    So they are placing their facility within 4 feet of the exchange's boxes? If not, I think I'm calling bullshit.

  • (Score: 2) by Scruffy Beard 2 on Wednesday January 25 2017, @12:08AM

    by Scruffy Beard 2 (6030) on Wednesday January 25 2017, @12:08AM (#458341)

    Light moves at about 1 ft per ns. So it is actually about 1.5ft/ns in a transmission line.

    • (Score: 1) by Scruffy Beard 2 on Wednesday January 25 2017, @12:24AM

      by Scruffy Beard 2 (6030) on Wednesday January 25 2017, @12:24AM (#458346)

      about 8inches per ns in a transmission line.

      Who uses feet an inches anymore?

      • (Score: 1) by Scruffy Beard 2 on Wednesday January 25 2017, @12:28AM

        by Scruffy Beard 2 (6030) on Wednesday January 25 2017, @12:28AM (#458350)

        The (nano) seconds per foot in contrast with meters per second is apparently confusing.

        • (Score: 3, Informative) by bob_super on Wednesday January 25 2017, @12:36AM

          by bob_super (1357) on Wednesday January 25 2017, @12:36AM (#458353)

          Obviously, since you typed that it would be faster in a transmission line (1.5 ft/ns) than in the void (1 ft/ns).

          I use 180ps/in in typical PCB stackup, but metric everywhere else

        • (Score: 3, Touché) by takyon on Wednesday January 25 2017, @12:41AM

          by takyon (881) <{takyon} {at} {soylentnews.org}> on Wednesday January 25 2017, @12:41AM (#458356) Journal

          You require more thinking-typing-submit latency. :^)

          --
          [SIG] 10/28/2017: Soylent Upgrade v14 [soylentnews.org]
  • (Score: 3, Informative) by FatPhil on Wednesday January 25 2017, @08:45AM

    by FatPhil (863) <reversethis-{if.fdsa} {ta} {tnelyos-cp}> on Wednesday January 25 2017, @08:45AM (#458427) Homepage
    They're not measuring network transmission overhead, they're measuring PHY->userspace overhead.

    "TCPDirect attacks this network stack overhead problem with a “slimmer” kernel bypass architecture in which the TCP/IP stack resides in the address space of the user-mode application. ... it’s a use-case that requires only a limited number of connections. ... removes the need to copy packet payloads as part of sending or receiving."

    No need to threaten to call bullshit, you could instead just read TFA.
    --
    Great minds discuss ideas; average minds discuss events; small minds discuss people; the smallest discuss themselves
  • (Score: 5, Touché) by Shimitar on Wednesday January 25 2017, @08:45AM

    by Shimitar (4208) on Wednesday January 25 2017, @08:45AM (#458428) Homepage

    Come on guys... This is simply amazingly funny! Really, what kind of unit of measures you use?
    I can understand feed or pounds in the "real" world, if you where used to them... But, honestly, when i see "c" mixed with feets and "nano" attached to inches... i really start laughing!

    --
    Coding is an art. No, java is not coding. Yes, i am biased, i know, sorry if this bothers you.
  • (Score: 1) by khallow on Wednesday January 25 2017, @11:20PM

    by khallow (3766) on Wednesday January 25 2017, @11:20PM (#458719) Journal

    So they are placing their facility within 4 feet of the exchange's boxes?

    At this point, I wouldn't be surprised if in the near future someone is putting their facility inside the exchange's boxes - perhaps even on the same board. The whole mess will probably break down once HFT is renting space on the chip, but there's still a ways to go there.