yea it was real easy we just ported over our code from our outdated virtex-5 what do i have a ( LISP ) when i talk lol m - the guy with the v-7 farm floating in the atlantic - you know i just realized i might be talking to my competition are you just on the equities side?
Not my area, but fair enough if you are porting code onto FPGA - you are ahead of some IBs if doing this anyway.
we are a bd for many years and trade thru prime banks. we think we are definitively off the hook as far as technology. we monitor latency globally. also we have a retail software product that hopefully will become it's own virtual global exchange.
I was talking from an inhouse IB perspective and only for equities in EMEA. Nothing to do with me, I've never been the one personally making 6-7 figures per day from one market. What may surprise some people is that there aren't massive armies working on this. Most deployments will involve two small teams - Market Links (server tech, exchange colo, filling in massive account/config forms), and the Strategy Team (AST) who deploy the strategies and monitor messenging latency and report on the PnL. Do you use SolarFlare equipment or can say what your half round trip micros is regarding messaging?
Joke... It will always take energy to do something whether it be place a trade or move Newton's apple from point A to point B. No matter what there will always be latency involved - human beings can't avoid latency and in the world we live in human beings control the bots. Anything designed to cut latency is a scam. Colo is the only thing you can do - do you really think that they are going to build a massive server farm in the middle of the ocean that far away from the real trading floors? Why not just draw a straight line and run fiber conduit?
Of course he meant it as a joke, I don't know enough about FPGA to get the punchline though considering nobody has ever beaten the latency systems i'm familiar with, which don't use FPGA at all and only rely on the latest networking gear to have lowest latency.
This is a nice pleasant thread. Colo seems logical as Winston points out. Having information handy is important. Why centralize? Winston, my thanks also for sizing up the participants here. StrategyC, if you please, could you suggest how many exchanges around the globe allow for the daily extraction level per account (6 to 7 digits) you mentioned? My estimate is 85 and I sense there is also a range of market operating rules that would be cause for tweaking a central extraction operation. I also agree with On.Target's appraisal of the effect of HFT (a positive smoothing effect)(but, there seems to be an odd harmonic in the space) and the uncomplex view that there is a lot of time for taking profit segments in an unfettered effective and efficient manner. Finally, I empathize with those whose views are similar to those of Winston on the derrogatory nature of the development of instruments that effectively destroyed the future of the members of the public who played by the rules all their lives. I certainly hope the new markets of the future have the trustworthiness required to protect non participants from the reach of these new types of markets. It is the responsibility of today's non-financial industry types (such as myself (we)) to continue to extract and divert out (money velocity concept) of the markets as much capital as the capacity (the market capital flow rate) of these markets will allow.
85 appropiate markets, that's a lot - I don't know how many appropiate global markets there are. I was talking more about one major EU market with MTF support that was added much later than US alternative venues (ECNs). These systems do not need central global connectivity unless doing something that requires that.
Thank you so much. I wasn't too clear about the nature of the markets. What I meant to inquire about was how many sizable markets. I threw in global not meaning to imply the market had to be global. It could be local and sizable and allow for distributed planning allocations. Who would want to move money across borders since the problems to be solved are already localized? In this thread, there is performance information. The dimensions of the operator's performance can be such that it is possible to say that an operation can be grown to trade in a variety of markets using a given number of accounts in each market. My orientation has always been to the concept of FPGA's. Now they are coming into physical existence. The colo paired with using FPGA's at a location seems like a nice future for decentralized operations around the globe. Winston, to me, is the type of professional best positioned to be a key support.
Hmm, i thought you were meant to be a "master" of the markets? What you can see is what everyone else can see, so why are you asking stupid questions about globalisation and max extration of ticks from markets? Or are you just trying to take the piss and attempting sarcasm? Why don't you tell us all why you thought there are are 85 live global markets that have enough volume for us all to steal from, and have HFT taking place on them. I hope you can give a proper answer to this and not one that involves you talking pure shite. PS: Jack - do you know what MTF means?