Opening Cross: Don't Get Too Attached to Data ‘Furniture'

After four years in the same apartment, my wife and I are finally moving to new digs more befitting a married couple. However, for better or worse, the move is forcing us to be critical about our assets, and to part ways with furniture and other items that won't fit with our new décor, or simply won't fit into the new space.

The market data industry is also dealing with separation-forced or otherwise. For example, McGraw-Hill last week split its Standard & Poor's business into two units: its ratings agency, which will retain the S&P name, and McGraw-Hill Financial, a new unit that will comprise its indexes business, Valuations and Risk Strategies (VARS) division, its MarketScope Advisor research platform and Capital IQ subsidiary.

The new unit will be headed by Lou Eccleston, a veteran of S&P, Thomson Financial and Bloomberg, whose history suggests the vendor wants to take on Thomson Reuters, Bloomberg and Interactive Data, while lending more weight to Capital IQ's efforts to displace other vendors such as FactSet.

This could be one reason for move, says Adam Honoré, research director at Aite Group - or that McGraw-Hill is positioning the data group for a sale to another vendor looking to strengthen its offering.

Officials were unavailable to elaborate on the strategy behind the decision last week, but a company statement says the move will enable the S&P division to "focus on creating enhanced credit risk benchmarks and research for the expanding global markets in the new and evolving regulatory environment."

Bob McDowall, research director at TowerGroup, says this sounds like S&P's way of ensuring that any stringent regulation of credit rating agencies is not applied across its other business lines, and that he wouldn't be surprised to see other ratings agencies with data divisions, such as Fitch Solutions and Moody's Analytics, review their structures to ensure these areas are sufficiently segregated from their ratings business.

For others, separation is a good thing. For firms trying to disentangle legacy data infrastructures to adopt new technologies, the open-source Market Data Abstraction Layer from the Collaborative Software Initiative could prove a godsend. Instead of forcing firms to unpick tedious interfaces, MDAL acts as a layer between the old and the new, enabling firms to introduce new features and content as the business needs them, while allowing their IT groups as long as they need to extract the legacy platforms.

While focusing on this project, CSI will not expand its range of feed handlers, though there is clear demand for open-source versions, as firms typically face a choice: buy a vendor solution that may not best suit your needs, or build your own-with the associated maintenance costs.

"A lot of handlers today are for-purpose rather than generic-for example, a feed handler built for ultra-low latency... may not suit other parts of the business," says Emilio Mercado, director of financial services product strategy at CSI. "You can knock out a line handler quite quickly, but it can become an expensive proposition when you have 50 to 100 handlers to support."

Mike Dunne, chief technology officer of Activ Financial, which is porting its feed handlers onto FPGA cards in the vendor's hardware-accelerated MPU appliance, says that a trading firm can build handlers tailored specifically for its particular needs that will be faster than a generic vendor solution. However, he adds that firms concerned with speed will inevitably have to adopt hardware solutions, which are more complex, and which the firm may not have the required expertise to build in-house, and also that firms must adopt fit-for-enterprise solutions rather than focusing solely on speed.

But even with convincing latency and throughput metrics, it can still be hard to pry firms from their stuff when it comes time to pack up and move.