CFM's strategic migration to AWS for scalable quantitative trading operations
Context
Capital Fund Management (CFM) is one of the largest hedge funds in France, with more that 240 collaborators in Paris, London, New York, Sydney and Tokyo and more than $10 billion managed funds. A pioneer in the field of quantitive trading applied to capital markets, their methodology relies on statistically robust analysis of terabytes of financial data for asset allocation, trading decisions and automated order execution. This approach requires large amounts of computing power.
CFM's strategy to reinforce their leadership in the markets by increasing drastically the size of its activities and geographical presence has led to a saturation of their on-premises HPC clusters. Hence, CFM wants to be able to scale-out to AWS whenever their on-premises resources are exhausted, with the aim of increasing the availability of computing resources and the dynamisation of the research of new investment strategies.
Objectives
- Migrate daily production simulations to AWS, increasing its number while keeping a strict running time cap
- Migrate complex portfolio optimization processes to AWS, liberating on-premises resources and reducing the running time
- Protect CFM's IP by implementing strict security measures
- Reduce costs for less frequently run workloads by leveraging AWS on-demand hyper-scaling capabilities
Results
- Daily simulations are able to accomodate an increased number of processes while being executed in the time it takes to run the longest process
- Portfolio optimization is able to burst to AWS elegible tasks, reducing the total running time by 80%
- The on-premises HPC cluster is no longer saturated
- Costs are proportional to actual usage and no longer dependent on peak load provisioning