The FIX Inter-Party Latency Working Group
Wed, 02 May 2012 03:41:00 GMT
An interview with Henry Young and Raymond Russell
In this interview for the HFT Review, Mike O’Hara talks to Henry Young, CEO at TS-Associates plc and Raymond Russell, CTO at Corvil, Ltd. Henry and Raymond are two of the guiding lights behind the FIX Inter-Party Latency Working Group, the group formed under the auspices of FIX Protocol Limited to develop a new set of industry standards around latency measurement.
HFT Review: Gentlemen, welcome to the HFT Review. We’re here to talk about the FIX Inter-Party Latency Working Group, so can you tell me why it was formed and what issues it is intending to address?
Henry Young: Yes, of course. The FIX Inter-Party Latency (IPL) Working Group was set up when certain key movers and shakers with FPL (FIX Protocol Limited) identified that there was a clear opportunity for some standardization, because we had a situation where there were a number of vendors of latency monitoring solutions that were not interoperable.
If for example you had an exchange and a broker who wanted to peer and exchange time-stamped observations, in order to enable either party to calculate latency between their respective systems, they would each need to purchase the same solution from the same vendor. But if one exchange was using vendor A and another exchange was using vendor B, the unfortunate broker would have to purchase both solutions.
So there was a clear need for standardization around the interoperability of latency monitoring solutions, particularly when measuring inter-party latency.
The other area that clearly required some standardization was the quoting of latency figures. Various different organizations quote latency figures, be they exchanges on their order match time or brokers quoting figures on how quick they are to market. The problem is that all these people quote numbers without really telling you what those numbers are. For example, do they include firewall latency or not? Do they include the connectivity gateway or are they simply matching engine latencies for example? And they don’t tell you under what circumstances those figures are being obtained. In some cases they don’t even tell you whether it’s a minimum, a maximum or average. Clearly there was a requirement to come up with a standard approach to where those measurements are made, in order that when organizations quote numbers, they’re quoting in the context of some standard methodology for where you make those measurements physically. That way, the numbers become a bit more comparable on an “apples to apples” basis.
So the two key initial things that the working group figured out it needed to do were to come up with a standardized reference topology so that people could make observations at comparable locations in different systems, and an interoperability protocol to enable different latency monitoring solutions to work with each other.
HFTR: Are you seeing more of a demand for this standardization from a particular type of firm, buy-side, sell-side or the vendor community, for example?
Raymond Russell: The number of participants joining the IPL Working Group from different organizations is roughly proportional to the number of those organizations, so you to tend to have more banks and brokerages than exchanges, for example.
But if you look at the really active participants in the group, you’ll see that it isn’t being driven by any one type of firm. Obviously this initiative wouldn’t be going anywhere if the vendors who provide these solutions in the market weren’t interested in supporting it, so TS-Associates are obviously key movers and Corvil support the initiative very actively as well, but you’ve also got active participation from the likes of the London Stock Exchange and New York Stock Exchange too.
HY: One of the reasons the working group was formed under the FPL organization is because the constituency of members at FPL - which is a wide selection of exchanges, brokers, vendors, buy side organizations, etc. - nicely encapsulates the constituents and sea of interest for inter-party latency monitoring. And that indeed is reflected in the membership of the working group.
If you look at how the working group breaks down, there are now 180 members but a lot of those people are not particularly active participants, they either joined up in an initial flurry of enthusiasm or they’re just there to keep tabs on what’s happening.
There is a core of maybe 10 to 15 people who dial into the calls with any degree of regularity and then amongst those people you have maybe five or six who are enthusiastic key contributors to the technical group work, of which Raymond is one of the foremost participants and I am co-Chair.
The people who are consistently enthusiastically involved tend to be those who have strong opinions and are very well informed in this space. And those individuals would be involved whichever firm they’re working for, so there’s Vitali Vinokour for example, who happens to work at a brokerage firm, and Mark Reece, who is a key contributor and who happened to be working for HSBC at the time he drafted the prototype specifications. Although Mark is no longer at HSBC, he continues to be actively involved and will probably stay involved whichever firm he works for in the future.
This is actually what’s so powerful about the IPL working group. We’ve managed to attract people from the industry who are excellently qualified to help formulate and shape what we’re doing. That’s the bottom line. It doesn’t matter whether they’re from exchanges or brokerages or buy-sides or vendors or wherever.
And the fact that there are more people from banks, brokerages and exchanges getting involved than there are from the vendor community is great because this is not just some cozy vendor jamboree. This is a genuine industry interest group that has realized this standard is in everybody’s interest, irrespective of whether the firms involved are competitors or co-operators in a wider sense. The best example of that is that two competing vendors, TS-Associates and Corvil, are co-operating so keenly to make this real.
HFTR: What kind of challenges are you facing in getting firms to accept and adopt the IPL standard? And as a standard, how easy or difficult is it to adopt? What is involved?
If you are not a registered user, click here to register for free