Scott DePetris, Portware
"Workflows are only getting more complex (and) there's a growing need and demand for enhanced decision support, automation and predictive analytics during trade execution."
Now that sophisticated high frequency, low latency trading and the technologies underpinning it are relatively commoditised, it seems 2015 trends will involve satisfying clients and regulators that operations are stable and reliable.
The game, said Sean Brown, president of 7ticks, is changing and an increasing number of the trading community are satisfied with being "fast enough" rather than the fastest.
7ticks is the managed services business at Interactive Data, providing ultra-low latency trading access to markets.
"Low latency is facing competition from infrastructure predictability and stability," he said. "The focus is now going to be more about transparency, consistency and reliability."
Brown added that the human element is going to be a much bigger deal. "Many firms are willing to spend more money on adding people resources to help make clients happy, rather than a faster switch."
Another dynamic expected to become more pronounced is the shift of risk to the buy side, as funds steadily take more control over their trading and incorporate tasks and functions traditionally considered sell side.
Multi-asset class complexity
Scott DePetris, president and COO of Portware, which provides broker-neutral and automated software for equities, options, futures, fixed income and FX, said that some of the largest and most progressive buy side shops are going to keep getting more automated to manage complexity across asset classes.
"Workflows are only getting more complex (and) there's a growing need and demand for enhanced decision support, automation and predictive analytics during trade execution," he said.
The increasing complexity of managing a trading desk is a reflection of market structure itself. Regulations are moving forward to boost investor confidence, shaken by what seems like constant market glitches, freezes and halts.
RegSCI, for example, enforces written procedures and audits for safe systems in the financial markets, not unlike ISO9000 best practices.
The countdown clock
In a market where speed is measured in ever smaller increments of a second, trends such as data governance and cloud adoption are going to require more attention to clock synchronisation.
Victor Yodaiken, CEO of FSMLabs, said that when the company shifted its focus from industrial uses of real-time technology to Wall Street, there was a major deficiency in time synching being reliable and documentable.
"It's pretty easy to get a piece of software that will tell you you are within a microsecond or 50 milliseconds but to actually be able to validate that, to do something when something goes wrong is hard," he said, adding that new FINRA and US CAT (consolidated audit trail) requirements will require validation and documentation at the 50 millisecond level.
Then, there are issues associated with large-scale infrastructure scattered around the globe - a complex management problem that can't use the same software and hardware designed to accommodate a shop with a couple of machines.
The problems with poor design started to pop up in major trading venues. In August 2013, for example, Deutsche Boerse's derivatives exchange Eurex halted trading for one hour due to an incorrect time sync with the system clock. Eurex has since been praised for the way it's implemented time stamping technology.
But at the time, it was just those kinds of events driving business to the vendor, Yodaiken added, as more trading outfits realise time stamp systems are inadequate.
Even with just a slight lack of synchronisation in the times of two machines, for example, an order could get confirmed before it's generated. Explain that to regulators if something goes wrong.
Meanwhile, the march of cost efficiencies and quant technology is going into the cloud as data portability makes increasingly sophisticated strategies more accessible. And cloud adoption brings up a whole new set of time sync issues because virtual machines can easily be several seconds off in the cloud.
As is often the case, problems defined at the bleeding edge, bleed through.
"What you are doing in automated trading is running a distributed transaction system…What (trading firms) have done, particularly the high frequency trading people, is they have been at the cutting edge of transaction processing but those kinds of issues that they see are not peculiar to them, they are maybe more general issues for everybody," Yodaiken said.