Swiss Re Migrates Risk Model to the Cloud with MATLAB Parallel Server
There are internal security restrictions that prevent Swiss Re from using the public cloud in certain ways. Public IP addresses, for example, are not allowed in the setup. So, Swiss Re worked with MathWorks engineers to customize the deployment process and make the whole thing work.
Key Outcomes
- Millions of large-scale simulations ran with improved scalability and availability
- Cluster deployment times reduced from days to minutes
- Costs reduced by hundreds of thousands of dollars
As one of the world’s leading providers of reinsurance, insurance, and other forms of insurance-based risk transfer, Swiss Re relies on a sophisticated capital adequacy model to comply with Swiss Solvency Test and Solvency II requirements, as well as business steering. Developed in MATLAB®, the model is part of a complex technology ecosystem that Swiss Re has operated in a private cloud for several years. The company uses MATLAB Production Server™ and MATLAB Parallel Server™ to run simulation workloads on a computing cluster with more than 280 CPUs and 5 TB of RAM.
As part of a larger cloud transformation journey currently underway, Swiss Re is migrating its capital adequacy model to Microsoft® Azure® while continuing to use MATLAB, MATLAB Production Server, and MATLAB Parallel Server. The Integrated Risk IT team worked with MathWorks to customize reference architecture templates for Azure, helping Swiss Re maximize throughput, minimize cloud costs, and enable on-demand scaling as they run business-critical profit-and-loss simulations in the public cloud.
