Skip to Main Content
As the processing power of set-top box (STB) devices continues to grow with each new product generation and multiple service operators directly manage millions of these devices through a dedicated broadband network, we propose to harness this massive aggregate computational power to execute large-scale parallel applications. We present a prototype implementation of a heterogeneous distributed embedded system that combines a complete head-end cable system with a DOCSIS broadband network of STBs and an interoperable embedded implementation of Open MPI running on the STB operating system. We demonstrate the operations and performance of this system by reporting the experimental results that we obtained for the execution of various instances of Multiple Sequence Alignment, a parallel bioinformatics application. Based on these results we propose a new distributed embedded system architecture for highly-scalable intrusion-detection systems.