Middle Office in the Cloud
By Henrique Hablitschek, Equinix
I remember back in 2010, when I was working for a High Frequency Trading (HFT) service provider in the New York area, I researched to see if Cloud Computing could improve performance and eventually reduce costs in our business. No, at that time cloud computing was not mature enough to be useful.
About two years ago, I was indirectly involved in a project where one of the largest US banks was moving to the Cloud. Hold on, a “bank moving to the Cloud” is a very broad expression, you might say. True. In this particular project, the bank was actually redesigning their network in order to reduce the number of MPLS circuits and the Cloud was to host specific applications determined to be cost-optimized on Cloud infrastructure. Since then, I have noticed all types of financial institutions moving their Software Development, Human Resources, Regulatory Compliance, Clearing Procedures and general IT to the Cloud computing world.
Historically, financial institutions have been divided in the three areas: front-office, middle-office and back-office. The back-office or “enterprise side” of these firms is definitely moving to the cloud today. For the past of two years, financial firms have been moving the support functions of their businesses from legacy data centers to As-a-Service Providers, making hybrid cloud a reality.
So what about the front and middle-office applications? For now, I do not see front-office trading moving to the cloud. Due to regulations, proprietary applications and algorithms and an efficient price discovery mechanism, financial trading may never fully adopt Cloud as a part of the strategy on a wide scale. However, the middle-office is primed and will be the next area to move to the Cloud.
While the back-office of a financial institution provides support to the general business, the middle-office is really in place to provide support to the trading functions of a financial institution. This may involve data analytics, regulatory support or other functions that are specifically related to trading, but not directly involved in the execution of a trade. A major role of the middle-office is to reduce risk (VaR) and manage profit and loss. In order to perform these tasks, the middle-office department will require access to large amounts of existing information or historical data, a.k.a. Big Data.
Financial Big Data is a collection of historical Market Data (old stock prices from years back, the dates and times of trades, volumes of trades executed and the prices stocks were bought and sold to name a few) sourced from various asset classes and from all over the world. This broad range information could be stored in a Public Cloud where customers can scale up or down based on what they need. Not all of the historical data is relevant to a financial institution’s trading strategy so the implementation of the middle-office in the Cloud will be somewhat specific to each financial firm; keep this data, but dump that data. In this model, the Cloud Storage collocation or proximity of the data therein will play an important role before trading – risk mitigation. Mitigating risk through the use of Big Data will strengthen the relationship between the front and middle-office. When implementing such an approach, the majority of business traffic stays on private communication links (cross-connection) not accessible to cyber criminals.
There is a trend within the US Exchanges to implement the same length of cable between the Exchange’s matching engine and their customers. This architecture is equalizing the latency (or physical distance) for all players, removing one major piece of the HFT strategy. Considering that financial firms will need to access the Big Data before executing a trade, the impact of latency will involve not only that to the exchanges, but also latency to the cloud providers. The Cloud solution will be a new entrant in the financial trading environment, providing storage (Cloud Storage), database services and the processing of algorithms (Cloud Computing). The financial firms will demand state-of-the-art multi-core devices to speed up their algorithms when analyzing the large amounts of structured data and objects storage in low latency devices (in-memory cache) involved in a trade.
Conclusion: Middle-office support of trading applications utilizing ubiquitous market data to mitigate risk demands fast, but flexible storage. In cooperation with the algorithms implemented to execute trades, hybrid cloud for the middle-office is becoming a reality for the financial community. Constant computer upgrades (decreasing time to market) may eventually push trading algorithms and ultimately trade execution itself into the cloud, but today, hybrid cloud will be used to reduce pre-trade risk. The legacy financial data center continues to move to a rich interconnecting facility allowing the data to be transferred between private and public clouds, reducing latency while increasing security.
Mr. Hablitschek has 25 years of experience in the IT field, with more than 10 years in the financial services technology industry. He currently serves as Head of Capital Markets Strategy at Equinix – where he helps clients and prospects strategize deployment planning for their critical trading infrastructure around the world. Prior to working for Equinix, Mr. Hablitschek led network design and implementation for the ultra-low-latency trading platform provider, Mantara, LLC. Past employers also include Bloomberg and NYSE Technology, where Henrique spent more than 3 years engineering the SFTI network. Originally a native of Brazil Mr. Hablitschek holds a bachelor’s degree in electronic engineering, a master’s in electronic systems, as well as a post graduate degree in telecommunications.
Customers can now access an expanding library of datasets via cloud APIs.
FINRA’s Office of Financial Innovation reviewed nearly 40 broker-dealer firms.
DZ BANK, BayernLB and Deutsche Börse implemented the digital OTC derivative using DLT and cloud technology.
Firms are increasingly moving to more agile technology, microservices, and data models.
Workloads supporting retail, commercial and investment banking, and wealth management will move.