By Michel Finzi
Without competition there would be no need for innovation. Innovation results in new products, new functionalities and new markets, all of which force industry participants to make important decisions, such as which markets to operate in, which business opportunities to pursue, grow or exit, and which competitive forces to react to. Most businesses today are bombarded with a seemingly endless stream of options that need to be constantly evaluated in order to move forward. The phrase “the only constant is change” may be overused, but that doesn’t make it any less true.
In parallel, many of us have become accustomed to getting ‘more for less’, which has now has become something of an expectation in today’s markets. When evaluating any buying decision, two key elements we often look at first are perceived improvement in performance (the ‘more’) and reduction in cost (the ‘less’). And why not? These two factors are essential when making such choices. The tricky bit, however, is that the initial analysis of these two inter-related topics can often focus on just the tip of this iceberg, the seemingly innocuous visible portion above the surface. But we all know that much more lies beneath the surface.
Performance is not just about latency
Tackling the performance part first. In the world of high reliability and high performance trading infrastructure, the most visible part of the performance/cost iceberg is generally latency and at Celoxica, this question comes up all the time. It is all too common that firms considering the deployment of new trading technology for processing market data, performing risk checks or routing orders to liquidity centers, tend to focus on the speed at which all of this occurs. However, even making the very broad assumption that everyone uses a consistent measurement approach, focusing on just this one ‘headline’ number is misleading because it will not cover the other important attributes that will ultimately distort the cost side of the equation substantially and very likely not result in improved overall performance.
To be absolutely clear, we are not minimizing the importance of latency statistics or downplaying latency’s importance in any performance evaluation process. Far from it. Celoxica has always delivered – and will continue to deliver – ultra-low latency performance reliably and deterministically across markets, geographies and asset classes. We are simply sharing that in our experience, for many firms, while this is a critical component of the decision-making process, as they dig in, latency is not the only true measure of performance.
Many other factors lie just beneath the surface. When considering new ways to cope with capacity and performance, Celoxica finds time and again, that it is equally essential to consider factors such as scalability across the enterprise, integration with consuming applications, adaptability of a solution for different business use cases and coping with change over time, scaling the solution across markets, geographies and asset classes and how well the technology is supported in mission-critical situations. Said another way, what is the benefit of owning a high performance sports car if one lives in a crowded city, with bumpy roads and with no place to park it?
It is basic human nature to confront any question or topic by relating it to something we have experienced in the past. The same can be said for the criteria often used to help in the decision-making process, which are often shaped by external forces or the industry specific chatter ‘du jour’. This approach will not necessarily yield the best results. Firms that regularly perform ‘mark to markets’ and consider the broader implications a change could have on their organization will be best equipped to cope with the future decisions they will have to make. Such firms will be able to properly calculate the performance/cost equation for themselves. Failure to do so is only postponing the inevitable.
Celoxica is always happy to engage with clients and prospects in these broader discussions and we encourage any firm considering new technologies to obligate their vendors to do the same. Doing so up front will save time, money and effort.
In future blog posts we will look at the cost side of the performance/cost equation, so check back soon.
Read the original post on The Trading Mesh.