Capco has begun a project to re-engineer the technology infrastructure supporting its provision of managed reference data services, and plans to draw on opensource and grid technologies as well as Web Services, to improve integration with its clients’ operations. The provider’s aim is to improve automation and build in greater scalability to the platform, in order to be able to support more clients and improve the cost-effectiveness of its outsourcing proposition. The effort, which started in the new year, is being spearheaded by its recently appointed chief technology officer Richard Bemindt, who counts a stint at GoldenSource among his previous roles.
According to Capco, it currently delivers managed reference data services to 40 clients globally, and its re-engineering project has been initiated in response to increased demand for outsourced reference data management. Capco remains bullish about the business despite signs industry-wide that the data management outsourcing proposition has not lived up to expectations, with SunGard having failed to take its Banc of America Securities and Dresdner deals through to completion (Reference Data Review, December 2006) and Accenture apparently de-emphasising the business and losing key executives including Pat Tsien, Bill Cline and Neil Edelstein.
Key elements of the Capco managed reference data service include the Iverson business it acquired (Reference Data Review, February 2004) and the ING platform it brought on board last year (Reference Data Review, March 2006).
Both deals brought in technology to Capco, which has also developed further technology inhouse to support its service provision. The aim now is to build a single services oriented architecture (SOA), with the end goal being a single operating platform delivering all current and future service offerings.
Since his appointment in December, Bemindt has already begun work to expand the Capco service in a number of areas, he says, including extending the number of data feeds handled and improving automation around them. “The work will be ongoing for some period of time,” he says.
Bemindt believes his background fits him well for his task at Capco. “I spent more than 10 years at Computer Associates working on enterprise management type technologies, gaining experience in building fault tolerant and redundant systems from both a hardware and software perspective. I then went into the financial services industry, as CTO at FTI, now GoldenSource. After FTI, I became CTO/CIO of a financial services ASP, IFS, with a charter to provide processing for hedge funds. Afterward, I helped found Petal Computing, an enterprise systems play, which provided a grid computing solution, where I learned a lot about this new technology and how it fits in the enterprise systems space.” The mix of experience means he brings a diverse background to Capco, he reckons, “comprising the essential skill-sets required to deliver a managed reference data service”.
Capco currently has a variety of “well founded” technologies in place to serve its managed reference data services users, Bemindt says. “As we take on larger scale clients with more data to manage, we need to ensure all the moving pieces work together and without fault.” The more customers come on board, the more data needs to be managed, and the more data feeds need to be added, he says. “The more data there is to deliver, the more subscription information there is to manage. The databases grow very quickly, and the technology infrastructure needs to scale-up: we can ill-afford to have any bottlenecks.”
Bemindt says the platform in place today is reliable and scalable. “However, we envision that there will be a point at which – as with most software – it becomes a bit difficult to scale beyond,” he adds. “Our goal is to always stay ahead of the requirement. We will carry out load tuning and analysis of our technology to see where we can improve things. One option we may take is to take advantage of grid technologies to enable us to parallelise processing. There is plenty of hardware and software out there that will allow us to scale to a level much higher than we currently require.”
Capco has evaluated a number of technology options for its build-out, according to Bemindt, “including everything from opensource, to fully purchased products and internally developed software to facilitate the delivery of the platform”. He adds: “We will stay conformant with standards, rather than ‘painting outside the lines’, so that if we do want to replace components going forward that will be straightforward to do.”
A key aim of the re-engineering project is to enhance service delivery, Bemindt says. “We will improve the model for downstream delivery of data to our clients, and improve the integration of our data into our clients’ operations. A number of technologies I have worked with – such as Web Services and BPEL – will be useful for adding different delivery models, improving our service and reducing cost.” Capco is keen to emphasise that the transition to the new environment will be straightforward from the perspective of its clients, though it does want to encourage the take-up of new technologies where appropriate. “Anything we do in terms of improving delivery will maintain compatibility with existing capabilities – we currently deliver data in a variety of transport protocols, and will continue to do so,” Bemindt says. “We will enable our clients to pick and choose how they integrate with us. However we are keen to help our clients modernise and improve wherever possible, and we believe Web Services is the way to integrate.”
Another key aim is to “improve the cost-efficient nature of the platform we have, by introducing opensource software and new technologies for automation”, Bemindt says. “Going down the grid/parallel processing route is one way to reduce the costs of the platform, because it means we can use low cost, commodity hardware.” A further opportunity to reduce costs is to improve automation and reduce manual processes, he adds. “There are situations where data feeds from vendors come in with some corruption, or invalid values that fall outside of the norm. We do have data cleansing capabilities today. We want to tighten those up and add additional algorithms that will automatically find anomalies and report them to our operations staff so they can be fixed at minimal cost, as opposed to having to rely on manual intervention.”