The software implementation of online stream processing algorithms cannot achieve the required performance due to limited computation capacity of the processing resource which is unable to handle stream data coming at a very high speed and needs to be dealt with in a certain period of time. Since, modern hardware platforms provide a high level of parallelism (highly parallel CPUs, GPUs and FPGA). The applications can be accelerated by exploiting this parallelism. The incoming stream can be processed using multiple processing elements resulting increase the performance of the algorithm significantly. The original algorithms have been implemented in Java.
1. Suitable hardware platform for the execution of the platform other than the desktop computer
2. Comparison of performance results
Java, VHDL, FPGA
Lab allocations have not been finalised