navigating the technology demands of sound counterparty risk management
DESCRIPTION
My talk given at PRMIA (Professional Risk Manager's International Association) event in Singapore, September 2012TRANSCRIPT
The views and opinions expressed in this presentation are
those of the speaker’s and do not necessarily reflect the
official policy or position of Microsoft
PROCESS
Modeling
Automation
Governance
PEOPLE
Awareness
Experience
Culture
TECHNOLOGY
Price / Performance
Data Management
Reliability & Scalability
Scenario Generation
Instrument Valuation
Portfolio Aggregation
Credit Value Adjustment
Trades Path Steps # of Obsrv
HVaR 99% 10,000 250 1 2,500,000
MCVaR 99% 10,000 1000 1 10,000,000
PFE 99% 10,000 1000 100 1,000,000,000
Trade Data Reference Data
Market Data Collateral & Netting
Ad-hoc Analysis
Risk Reports
Excel
Data Quality &
Transformation
High Performance
Computing Grid
(CVA Calculation)
Capacity Burst
into Cloud
Client Drivers
ETL Load
Interface
Support/Patching
Corporate Backup
Solution
CONTROL
RACKDATA RACK
Control Node
Management Node
Landing Zone
Backup Node
On-premise
Desktop User
HPC Head Node
Broker Node(s)
HPC Cluster
Desktop Compute Cloud via
Idle Win 7 Workstation Cores
Azure Compute Instances
Azure Compute Proxies
Company decision makers were so impressed, that
they asked our risk managers to recalculate
every simulation that they had run over the
previous year in just two weeks
Hein Brat
Program Manager, ING Insurance Benelux
• Monte Carlo Risk simulations were time consuming and infrastructure intensive
• The bank initiated project on the Microsoft platform with Numerix, a leading
risk solution provider, to burst computations from the High Performance
Compute Server to the Microsoft Azure cloud
• Calculations that took 300 minutes on a PC, took 30 minutes with six Azure
nodes or 15 minutes with 20 Azure nodes