CANTOR – our massive parallel pricing & risk platform

 

CANTOR is our in-house developed valuation and risk platform. It is designed from ground up with scalability, flexibility and resilience in mind.

CANTOR’s massive-parallel calculation engine is capable of utilizing multiple pools of computational resources ranging from our in-house grid to the grids of various cloud and grid providers. It supports heterogeneous operating system and hardware environments. It is currently utilising Windows and Linux grids with a mixture of Intel and IBM PowerPC 7 CPU’s. In addition CANTOR’s distributed engine is designed to maximize the utilization of the available computational, network and other resources by exploiting such techniques as data affinity, benchmark based task scheduling, dynamic adaptive load rescheduling etc.

Another important part of CANTOR is its integration and web-services engine. Its design allows for a very quick and efficient integration of different data sources (market, trade, static and reference data) and allows for seamless integration with the customer’s risk management or pricing platform.

CANTOR’s capabilities are not limited to EOD/batch type of use cases. It has been designed for intra-day and on-demand pricing and risk use as well.

Another trait of CANTOR’s design is its capability to utilize various pricing libraries and provide quick integration of models not only from our in-house proprietary library, but variety of commercial and even open source pricing libraries.