Complex Event Processing - It’s More Than Algo Trading
By Don DeLoach, President and CEO, Aleri
Part 2 of 4
[Continuing on from Don’s introduction to this subject, he begins to drill down on applications that can leverage CEP]
Not unlike any emerging modality in computing, market acceptance tends to determine the “must have” capabilities for the market participants. This was true with relational databases regarding stored procedures and triggers and row level locking; it is true of televisions supporting HDTV, and it is true for digital audio support of MP3 formats. Notwithstanding the bluster of any particular vendor with larger than reality self perception, it is the market, and only the market, that determines standards and direction.
In the case of CEP, the market is beginning to reveal itself regarding these basic capabilities, and is well illustrated in the use-cases that stretch across the trade life-cycle beyond simply supporting algo trading. Here are some examples …
Market Data Enrichment
As markets become more competitive there is a growing trend toward taking data direct from the source rather than through an intermediary to reduce latency. At the same time, as trading strategies become more sophisticated, there is a need for more “intelligent” market data. This entails everything from price quality assessment and cleansing to value added field calculation that might be a part of an existing market data feed but absent from direct data, to multi-stream aggregation.
While this is hardly a glamorous use of CEP, the technology provides an elegant way of easily checking and enhancing the incoming data streams at high rates, from multiple sources, all without adding material latency. CEP is a particularly attractive enabling technology when you consider the rate of change in change sources, formats and the input needs of trading applications. Unlike a hard-coded “value added data server”, this functionality can be implemented rapidly on a CEP platform and then easily modified as landscape evolves.
Pricing/Auto quoting/Market Making
There is no doubt that the acceleration of the trade cycle brought on by electronic trading and the vast increase in the volume of data in the market has taxed the upper limits of any human based market making activity. A trader on the phone quoting currency pairs is no longer a viable option for establishing a market in currencies, much less most other instruments. While the human mind generally takes somewhere in the neighborhood of 250 milliseconds to respond to an event, trades today happen at millisecond speed. In 250 milliseconds, the market can come and go, with the human trader left behind. CEP is exceptionally well suited to perform this task, from absorbing multiple prices from the market at any given point in time, calculating corresponding offer prices, and offering them back out into the market. CEP can enable market making with very low latency, yet exceptionally large reach. The human trader quoting three pairs with a 10 second delay becomes an auto quoting engine quoting 100 pairs with a one millisecond delay.
Order Book Management/Market Depth Analysis
With the growing competition amongst exchanges and the increase in ATS’s and ECN’s, there is increasing complexity and fragmentation of liquidity in the market. The ability to manage multiple order books, reconstruct a virtual book of the entire market, and have the ability to execute rules against full market depth is an increasing demand of the market, and one well served by complex event processing. This, in particular, caters not only to those platforms where state is managed effectively, but requires a CEP implementation that can process incoming data not just as new elements (a new order added to the book) but as changes to existing data elements (an order cancellation or change to order size).
[To be continued - more applications to be discussed in part 3 next week]
See Aleri at:
Technorati Tags: cep, complex event processing, aleri
