June 2014 Sponsored by Intel
The Ins and Outs of Infrastructure
Over the past decade, the concept of market data infrastructure has gone from a single core platform in a back room, sending data primarily to terminal displays, where sub-second latency was considered high performance, to a much broader and more complex collection of hardware, software and networks, both within firms' walls and in co-location centers and cloud farms, gobbling budgets, staff and time.
And as infrastructure becomes more complex, it has also become more critical to trading operations. Nathan Boylan, head of IT at asset manager Lord Abbett, calls infrastructure a "business-critical service" that needs continuous modernization and improvement.
Much of the modernization and improvement of recent years has been focused on low-latency technologies. But in reality, this is only one small part of the modern data infrastructure, which begins at the microprocessor level and beyond with multi-core processors, and encompasses everything from how those processors interact to how the circuitboards on which they sit connect to one another, how data is queued, cached and processed, all the way up to the fiber or microwave networks that connect data and trading platforms in a firm's office to its datacenter and to execution venues.
However, die-hard proponents of on-site infrastructure face an uphill battle as the trend of offloading infrastructure to cloud compute models and infrastructure-as-a-service providers grows. While this may not matter for vendors like Intel─who will continue to manufacture high-performance processors for computers, regardless of whether those boxes are owned by trading firms or by someone who leases virtual compute power to trading firms─it does mean that infrastructure providers and customers alike must re-think the way they provision services: Not all systems can instantly move to the cloud without some re-platforming work and its associated time and cost.
Hence, migration to cloud-based infrastructures may be the final nail in the coffin for legacy platforms. Another of those nails is the need to support an increasingly mobile workforce with mobile applications, who expect information on everything from market data to reports about the performance of that data at their fingertips as readily as their personal apps.
Ultimately, these changes may usher in the long-awaited era of true best-of-breed platforms leveraging open standards to stitch together components from multiple vendors and internal sources to deliver best performance.
Bloomberg's Gerard Francis comes on the podcast to talk about how data fragmentation issues are becoming more challenging.Subscribe to Weekly Wrap emails