Notes from the Consultant’s Jungle

TelX hosted a financial services roundtable event at their Financial Business Exchange (FBX) data center in Clifton, New Jersey. The panel of industry experts (and TelX customers) was moderated by the TABB Group’s Kevin McPartland, and introduced by TelX CEO Eric Shepcaro.

Mr. Shepcaro introduced the event by giving his own personal background about why we find ourselves hosting and attending an event like this. In other words, what is it that has created this focus on the financial services vertical by so many data center providers with property in locations such as New York City, Chicago, Toronto, and London? One issue clearly is the drive toward automation (of trading and other financial transactions). Electronic automation that is commonplace today was not thought of as recently as ten years ago, even as consumer-focused online trading services were emerging. Changes in technology are also driving this interest. Technology changes at the network level, server level, and even at the silicon level are impacting what is happening within this vertical and within the data center itself. Changes from the regulatory perspective are also in play, and of course the rebounding of the global economy is fueling momentum in the financial services vertical.

Speed matters in this business, whether through increased processing throughput or reduced network latency. Indeed because of skyrocketing quantities of market data and changes in equity markets, we’re already talking about nanoseconds in the data center, when yesterday we spoke of microseconds of network latency and sub-twenty millisecond WAN latencies between New York and Chicago. Shepcaro sees the network as a distributed IT utility, and TelX itself as a dynamic platform integrating communications through data centers.

The five-member panel at the roundtable was introduced by Kevin McPartland of the TABB Group . The panel included Eric Shepcaro, and thought leaders from Ballista, ACTIV Financial, Direct Edge, and RCN Metro Optical Networks. Mr. McPartland offered a sober perspective on the quest for ever lower latency by suggesting that there are 200 or so firms for whom ultra low latency is critical, but there is an abundance of opportunity in the single digit millisecond range as well.

When asked about trends through 2012 in the data center space, the panel’s response was predictably centered on greater power density and support for low latency network communications facilitated by the ecosystem of business partners collocating adjacent to one another in the data center. There was some creative discussion about the possibility of cloud-based utility computing services that can be economically shared by business areas that are time-of-day dependent (and do not share the same time of day, that is), but the notions of regulatory compliance and security project significant complexity on this idea. A second departure from the norm came from Mr. Shepcaro’s commentary on the way thin clients and mobile devices are making an impact because of greater security provided by placing the data processing in the consolidated data center, and the lower cost of these devices over the more conventional laptop.

The issue of fair access within the data center was raised, referring to possible legislation that would seek to ensure that no customer within a data center is at an unfair latency/proximity position with respect to other tenants. It was generally dismissed though, that while such legislation is likely, the target is not yet clear and the bar will likely be set at a level that’s easily accommodated by providers.

Regarding the value of low latency, and indeed whether this value can be quantified, is a difficult topic. Surely there are many firms that can quantify the value of lost opportunities if a particular transaction is not the first one hitting the state of the market. However, the race as currently defined doesn’t seem to really have a finish line. For example, if my investment of capital sets a new ultra low latency bar of say, 10 microseconds, this advantage lives only as long as it takes the next guy to match my innovation. Sure, there is money to be made in the mean time, but the race course resembles a hamster wheel rather than a road map. It’s important too, to be mindful of the overall holistic latency of the business process. It’s great if I am fastest at getting a trade to the floor but if my trade is using old market data I’ve missed some opportunity.

The panel discussion wrapped up with some talk about global markets (this was a US-centric event, and the bulk of the commentary assumed a world with only Chicago and New York as its cities.) In the context of this panel and its host, the landscape outside of the USA is complicated by the number of important global markets including London, Frankfurt, Hong Kong, et. al.) Interestingly, Toronto did not come up at all in these discussions even though there are very exciting developments taking place there now.