5 Predictions for Technology in Finance in 2013
Wed, 12 Dec 2012 16:36:16 GMT
This article was originally published at the finteligent trading technology community website, and is reproduced here with the author's permission.
This year has been a tough one for banking with fines for LIBOR abuse and Iran sanctions busting but there has been good news too. The report by the UK government into the computer based trading has been both thorough and reassuring. Computer based trading is not going away and with better analytical procedures at both execution venues and trading firms the future of technology innovation for financial trading is assured. The predicted explosion of hard wired programs in FPGAs (Field Programmable Gate Arrays) has been slow to materialise possibly because hardware engineers with finance domain specific knowledge are few and the improvements of mainstream computing from Intel have pegged FPGAs development to niche uses.
The new general Intel 26xx processor range at 2.9GHz has been shown to have the same capability of its 4.GHz cousin for trading applications when tuned to maximise its single threaded performance away from the default spaghetti code tuning.
Here are five predictions for the coming year in financial technology based on my research lab test and blog watching.
- Virtualisation for trading applications. The given wisdom that a virtualised environment can not perform as well as a native one has been demonstrated to be false. Network port binding to Virtual Machine solves both the latency issue and security through the well understood virtual LAN configurations.
- Massively parallel programming will cease to be a specialist topic with the more cores on processors. We have already seen the trend from Oracle with their T series RISC processors, it's only a matter of time before the x86 manufacturers get production level designs to challenge the GCU processor market with processors made of tens of low power cores.
- High Performance Computing modelling will become common place in Financial Services Industry. Other industries use engineering modelling techniques extensively and with the availability of faster IO and storage solutions the use of financial scenario modelling will increase.
- Hosting services will increase as point 2 and 3 kick in; the grid computing model with specialist hosting providers who will rent out time on high performance infrastructure to make scenario modelling and trade hosting a service available to more BUY side traders at a price point that was previously too high except for the SELL side firms.
- Real time controls at execution venues. Analysis of the mini flash crashes we have seen over the last year is indicating that financial trading patterns can be recognised so, controls at execution venues will pick up dangerous positive feed back loops and moderate market behaviour to prevent a flash crash.
For more news and views regarding trading technology, visit finteligent.net