The cost of technology to optimise trading infrastructure is going down, but that doesn’t mean demand for increasingly low-latency solutions is going down too. A panel at A-Team Group’s recent TradingTech Summit in New York City discussed the issues of optimisation for high performance under the auspices of moderator Greg Piesco, associate partner at Citihub Consulting.
Opening on the cost of technology, and how consumers are responding, the panel went on to discuss the issue that whenever lower latency is mentioned, there is always an impetus to build a new network, but how much will it cost?
As trading organisations continue to call for lower latency solutions and the name White Rabbit – a time-deterministic, low-latency Ethernet-based network which enables transparent, sub-nanosecond accuracy timing distribution – pops up repeatedly, low latency technology providers are under pressure to create more competitive technologies at lower costs.
Jason England, head of capital markets networks and hosting at TD Securities, said: “Technology for performance is getting much cheaper, we don’t buy anything less than a ten-gig circuit for anything now. Within the metro area it’s about one thousand dollars a month.”
Raymond Russell, co-founder and chief technology officer at Corvil, added: “Technology is constantly getting cheaper, but there’s still a competitive pressure to make sure you’re keeping up with your competitors, so there’s a certain steady state on the price you have to pay, but there’s no let up on the pressure on that. We see our customers operating in a way that’s ever more cost conscious, and one of the moves that we see is that people are consolidating and sharing infrastructure both internally and externally – it’s cheaper, but it’s still not cheap technology. We’re trying to sweat the most out of technology to meet the needs of business, but also to provide an economic price point.”
Cost-conscious firms are scrutinising where they spend as the cost of being a major player in the trading game has increased significantly overall in terms of infrastructure. In order to manage costs, Eric Powers, head of infrastructure at HC Technologies, proposed a model for charging underlying teams. He said: “In a smaller organisation, we’re very conscious about individual trading strategies being aligned directly with all the infrastructure they’re using. We look at the costs in alignment with trading teams to see, for example, if team one is more successful and profitable than team two.”
The discussion moved on to the granularity of timestamping based on three key areas: time, implementation and how protocols function. Steve Williams, general manager at Fixnetix, warned that if clock drift occurs, analysis can be off the mark, and if core systems are implemented incorrectly, data could be inaccurate. He said that if timestamping is not implemented properly in an overarching style that allows firms to connect to all the elements in their world and sync them, there will be clock drift. England added: “With MiFID II and the CAT, I would say that our cost to measure and be compliant, and process alerts and understand how things are drifting, is more expensive than the underlying system.”