Future of fintech is wide open
By Gary Collier
At Man AHL, we have a few core beliefs, and chief among them is that clients come first. Many firms would make the same claim. But where we differ is the approach we take to delivering a client first business ― we apply rigorous science to the financial markets.
One of the areas my team is responsible for is data infrastructure. How can we make our underlying technology faster, more efficient and easier to use? One important piece of data infrastructure that has changed significantly in recent years is the database. That’s the layer of software which stores and manages the data for an application. We live in a world where the amount of available data is increasing at an incredible rate ― an often quoted study suggests that 90 percent of the world’s data is less than two years old. Against that backdrop, a slow or difficult to use database could make a system unusable, a good database can make it unbeatable.
Two years ago we got to a point where we anticipated scalability issues with our tick store (a specialized type of data tool for storing high frequency financial services market data). It’s a crucial part of how our quantitative researchers make decisions. The old version had the potential to begin hindering their ability to do their job, so we knew we had to change. After a few trials, we created our own tick store using MongoDB. Called Arctic, the tick store is now the primary market data store for all our market data.
MongoDB is part of a new generation of databases that don’t handle data in the traditional row and column fashion used by relational databases. It looks at data as whole JSON (JavaScript Object Notation) documents, which makes things more flexible for a developer and means that the application performs well at a massive scale.
By building Arctic with MongoDB, and the open source programming language Python, we realized a 40 times cost reduction when compared to the legacy time series data store. Processing performance improved by 25 times and our disk usage reduced by around two thirds, as a result of using advanced data compression techniques. These were incredible results.
We could have stopped there, but we knew this new powerful tick store could have potential applications in industries far beyond financial services. The need to capture, store and analyze time-series data is ubiquitous in the world today, across industries, and Arctic is a general purpose solution which can store any time-series information, not just financial data. For example, it could also be deployed for a range of similarly demanding workloads from sensor readings from industrial machinery or scientific experiments, to measurement of climate data, sentiment analysis or tracking consumer behavior to drive predictive analytics. It could also be applied more generally for Internet of Things applications.
So earlier this year we decided to make it open source.
Why would we choose to give it away? It comes down to people. Despite what you may read, humans still matter in technology. To have truly state-of-the-art technology, in any industry, you need the best developers and the best can be hard to find.
The latest generation of coders was brought up in an open and collaborative world. They want to use open source software and they want to be able to contribute back to the communities that gave them so much. To be an attractive home for good developers, I believe you need to have a genuinely free and collaborative philosophy towards people and software. Making Arctic open source is a natural extension of that philosophy.
It all comes back to those core principles. We want the best people, using the best technology, to apply a data-driven, scientific approach to finance. Right now, that’s open source software and the developers who are experts in it. Regardless of the market or the model, we believe this freedom to collaborate is the most productive way to deliver value to our clients.
Our focus is not just about making money today. It’s about providing value tomorrow, next week and next year. When you’re focus is truly on client value it doesn’t make sense to hide away your team’s triumphs and failures. That’s why the future of finance, and fintech in particular, is wide open.
* * *
The writer is co-CTO of Man AHL