search menu icon-carat-right cmu-wordmark

A System Architecture for Processing Flows

In this presentation, Raj Srinivasan proposes a clustering architecture and demonstrates its implementation for commercial applications.

Software Engineering Institute


The simple architecture presented has the following features:

  • Scalable for CPU performance
  • Scalable for specialized coprocessor functions
  • Scalable NPU capabilities
  • Linearly scalable by choosing the appropriate mode of operation for NPU
  • A standard application environment that promotes ease of maintenance and application portability
  • Manageability
  • In addition, this architecture has been implemented for commercial applications in a highly integrated 2U chassis.

The next version of this product is under development and will incorporate more powerful processors, utilizing the best of SMP and Clustering features in CPUs, a faster backplane, and an NPU that can support 10 GBPS operation.

Part of a Collection

FloCon 2006 Collection

This content was created for a conference series or symposium and does not necessarily reflect the positions and views of the Software Engineering Institute.