Part 4 of 4 [Don concludes this series by looking at CEP applications beyond the financial markets]
A Few Examples Beyond Capital Markets Financial services is fundamentally about information. As a result, there is probably no end to the amount of use cases which will ultimately take advantage of CEP. At a minimum, wherever there are large amounts of fast moving data, and there is a need to understand things in tighter and tighter intervals leading up to real time information, CEP based applications will emerge.
One we have already seen is Treasury Liquidity Management, where there is a growing push in the market to manage bank-wide liquidity on an intra-day (real-time) basis. In fact, this is one of the pioneer applications where CEP has been used. Several banks are using the Aleri Liquidity Management System which is based on the Aleri Streaming Platform, proving the bank with a holistic view of positions in real-time.
There are implications in terms of managing down the required collateral posted at the central bank, diminished requirements for middle office intervention, increased customer servicing capabilities, and a drastically increased ability to understand and manage liquidity risk.
The characteristics of this use case are very, very similar to those involved with managing order books, where positions are updated in real-time, and silo data is also further aggregated for an enterprise (or market-wide) view of positions. Another example would be fraud analysis and detection, which could entail fraudulent credit card usage, or perhaps anti-money laundering efforts.
In this case, there are similarities to the market surveillance efforts discussed above. And the last example would be data orchestration and synchronization. This is a bit different, in that it involves maintaining an understanding of both new information coming into the bank, like with market data feeds, as well as an understanding of the sate of various silo systems, much like with liquidity, and then applying processing logic to determine where, when, and how to move certain data or update certain silos.
A good example of data orchestration would be the management and distribution of security master file information. This may seem quite straightforward. In practice, however, it is anything but. When silos exist in different countries with different laws governing what and how data may be treated, and the silos themselves are normally recipients of the data, but at times can be contributors of new data, then the ability to apply very complex processing logic can bridge the gap between the need to centrally manage and act of the same information while accommodating the very complex nuances of a globally distributed operation. Again, CEP with full state and very robust characteristics can provide an ideal platform to address such business issues.
Beyond Financial Services
While CEP feels like it is specific to the financial markets, at the end of the day, it is most certainly horizontal technology. The success and endurance of this technology may well be a function of it’s broader market acceptance in the long run. In fact, the fist commercial instance of CEP was really within Bell Labs in 1998 with its “Sunrise Project”.
While it has come a long way since then, CEP remains well suited for many Telco applications, like real-time billing, real-time mediation, and more. There are also a multitude of uses within the government, with the most obvious one being homeland security. Not that we want this too closely associated with the negative image of illegal wiretapping and espionage, the fact of the matter is that CEP is exceptionally well suited for any problem where very large amounts of fast moving data need to be analyzed in real-time.
That can be everything from packages moving through UPS or Fedex, to shampoo sales at Walgreens, to cars passing through the high speed RFID based tollways on highways today, to a vast number of use cases yet to be defined, yet undoubtedly soon to reveal themselves. One really interesting aspect to this is that CEP solutions offer an SOA-style approach to integrating with other applications.
As such, CEP applications such as the ones discussed here can be implemented with minimal intrusion to the existing silos. This means that big gains can be made in a short time frame with minimal cost. Some people are reluctant to believe this. It’s as if you were used to distance running with a portable cassette player.
It held limited music, was bulky, had limited battery life, and was fragile to the point that if you dropped it, you expected it not to work. Then you were handed an ipod Shuffle, with 20 times the music capacity and battery life at 1/30th the size and would likely still work if you dropped it on the ground and ran over it.
That’s counter-intuitive for those familiar with only the cassette player. Likewise, if your reference point is a big trading system, or a corporate CRM or ERP system, you expect months or years of implementation, tens of millions of dollars, and a project filled with setbacks and disappointments. They just are not an easy thing. Yet, CEP is a bit like the ipod Shuffle, where the gains are a bit hard to get your arms around when your point of reference are these large, monolithic projects where a hardened architecture is imposed.
There is no doubt that also trading is being enhanced by CEP. There is no loss of good stories about how this technology can help algo traders. And everyone likes to talk about how a new technology can help someone make millions, and we all know that prop traders are at the forefront of the new Wild West that we have come to know as the world of electronic, multi-asset, multi-geography trading.
It’s glamorous. It’s cool. But it’s a very, very small part of a much larger and ever growing picture that is Complex Event Processing. So stay tuned, the best is yet to come![tags]cep,complex event processing,aleri[/tags]