Qube Research & Technologies Delivers Investment Strategies Faster Using AWS

Business is constantly evolving for Qube Research & Technologies (QRT). The company develops statistical and quantitative techniques and programs and applies them to third-party and proprietary datasets to identify and develop potentially profitable trading and investment strategies. These strategies—known as alphas—are the key to QRT’s success. “To get ahead, we are always refining our processes to search for and create new alphas,” says Nick Harris, the company’s chief technology officer.

The alpha discovery process was challenging, though, because QRT was limited by an on-premises environment that supported its high-performance computing (HPC) data analysis applications. “We had a finite number of servers to use in our data centers, and we couldn’t do all the data processing we needed to effectively,” Harris says. “We had machines running around the clock, but we still couldn’t support the growth in data coming in from different sources. Our applications source and analyze data around the clock.” Without the ability to use compute capacity on demand, QRT lacked the agility it needed to keep pace with data growth. “We essentially couldn’t find or create new alphas fast enough for the business,” says Harris.

“We can integrate trading data from more data sources each day by scaling our HPC system on demand on AWS.”

Nick Harris, Chief Technology Officer, Qube Research & Technologies

Running Financial Trading Applications on AWS

QRT found the solution to its problems in the Amazon Web Services (AWS) Cloud. “We knew the cloud would provide the on-demand capabilities we needed, and AWS was far above any competitors in the marketplace when it came to the technology,” says Jon Fautley, systems and cloud engineer at QRT.

QRT chose to run its HPC platform on the AWS Cloud, leveraging Amazon Elastic Compute Cloud (Amazon EC2) for compute and Amazon Simple Storage Service (Amazon S3) as its bulk data storage platform, which enables the company to retain large volumes of data and access it when required for research activities. “Amazon S3 gives our research team the ability to be unconstrained by limited data sources where we would have to restrict our historical data to a subset of what is generated due to escalating storage costs,” Harris says. “We can use Amazon S3 and its different storage classes to store data for future use with a significant savings over traditional enterprise storage.” QRT also uses Amazon Relational Database Service (Amazon RDS) for the overall database environment. “Right away, AWS worked well for us and helped us accelerate our research,” says Harris.

After experiencing the benefits of faster analysis on AWS, QRT then chose to take advantage of Amazon Elastic File System (Amazon EFS) for scalable file storage for specific financial trading applications. “A number of our internal financial trading applications rely on having access to an underlying internal file system, and they are written in a specific way and can’t be modified,” says Harris. “Amazon EFS gives us a massively parallel, high-performance file system. It provides transient data storage for sharing data between HPC compute nodes, as well as stores results and enables them to be processed by users on their desktops.”

Increasing Profits by Creating More Alphas

Leveraging the agility and scalability of AWS, QRT has increased the amount of data it can analyze. As a result, it can produce more-detailed, higher-quality fund returns. “We can integrate trading data from more data sources each day by scaling our HPC system on demand on AWS,” Harris says. “Also, using Amazon EFS, we can make the data more widely available to researchers in an easily accessible file system. We no longer need to process and synchronize the data back to an on-premises solution.”

This agility and scalability have helped QRT discover more alphas. “We can run more data through our systems and generate more results,” Harris says. “And for any given alpha, we look at how an investment strategy responded to market conditions in the past, so we can extrapolate that to how we think it will perform in the future. Ultimately, more alphas means more profit because we can deploy new strategies we feel will generate the most return.”

Delivering Investment Strategies in Weeks Instead of Months

QRT is getting new investment strategies into production faster using the AWS Cloud. “Having access to on-demand compute and file storage means our research teams can iterate faster and test new things faster,” says Harris. “For example, we can test new services or features on high-performance GPUs without investing in hardware and maintenance. We have optimized the investment strategy discovery process by running on AWS. Previously, getting a new strategy from inception to production could take up to six months because we were bound by our compute requirements. This process takes only a few weeks on AWS.”

Relying on Amazon EFS, QRT can access a cloud-based file system that does not require application changes. “We can use our existing code base while moving the business forward by taking advantage of massively parallel compute along with a file system that matches its performance,” says Harris. “This ultimately gives us a competitive advantage because we can get to market faster.”

Additionally, the organization has built a research platform that enables the company to bring on more researchers faster. “Using AWS services, we have created a centralized research platform,” says Harris. “It helps researchers get onboarded quickly and gives them more compute capacity than their desktop could provide. Then, when they want to take advantage of parallel compute capabilities, that is available to them instantly, on a self-service basis, without us needing to provision hardware. It’s another example of how AWS gives us a competitive edge.”

Learn More

Learn more about high-performance computing on AWS.