Conference Report: FIXatdl Finds Favor in Sell-Side; Not So Much for OMS/EMS Vendors

While sell-side firms are eager adopters, OMS and EMS vendors have been slow to embrace an algorithmic trading language from the FIX Protocol Organisation, which helps them quickly adapt their systems to new algorithms, according to a panel at A-Team Group’s Insight Exchange conference on High Performance Technologies for Trading and Risk, held in New York this week.

The FIX algorithmic trading definition language (FIXatdl), version 1.1, debuted this past March and has attracted the support of top sell-side firms, including Bank of America Merrill Lynch, UBS, Goldman Sachs, Deutsche Bank and Nomura, said Greg Malatestinic, co-chair of the FIX Protocol Algorithmic Trading Working Group.

Malatestinic said that traditionally OMS and EMS vendors have had to develop screens that support sell-side algorithms but, in general, these screens are hard coded. FIXatdl can ease the process of creating new screen renderings because it exploits computer readable XML rather than the traditional method of rebuilding screens via detailed documentation, coding and testing.

Trading firms can get to market with new and updated algos in days rather than months, he said. However, Malatestinic said that “customisation is prevalent” among OMS/EMS vendors and they are reluctant to embrace a technology that could eliminate a professional services business opportunity for them. Not all vendors have sidestepped FIXatdl - Portware and RealTick have embraced the standard, he added. 

When asked if he welcomes FIXatdl, panelist Mark Powers, head of sales and marketing at Lime Brokerage, said: “Every broker would welcome it because of the speed to market. You won’t find anyone on my side of the bench against it.” Beyond OMS and EMS implementations, firms face a buy-or-build challenge in setting up a high-performance infrastructure to maximise their algorithmic transactions, panelists said.

The cost of building or revamping an infrastructure for state-of-the-art algorithmic trading may be overwhelming for mid-tier firms, and bulge bracket firms need to develop a quick response to high frequency trading firms, Powers said. In both cases, costly infrastructure overhauls can be a drain on IT resources and banks, especially large ones, may opt to invest in R&D for algorithms and to outsource the infrastructure.

Build or outsource, all firms face major algo infrastructure challenges, panelists said. These include the processing of high volumes of market data; the latencies created by pre-trade risk processes; order routing latencies; binary to ASCII data conversions; the Transmission Control Protocol (TCP) versus User Datagram Protocol (UDP) bottleneck; and legacy trading IT architectures.

Some issues may turn out to be more important than others, such as lost ticks of market data compared to lost orders, Malatestinic said. “You can lose sleep if you drop an order,” he noted.