RASSD operating system

Abstract

In this work, we present an implementation of a software and hardware computational platform for a Reconfigurable Active Solid State Drive (RASSD) node that was previously introduced in Abbani et al. 2011. This platform aims to resolve the bottlenecks encountered by data-intensive applications when run on traditional computational architectures. Our platform provides a speed up of up to 154 times, when executing the K-mean algorithm on a 200 MHz pipelined streaming co-processor hardware accelerator, over a state-of-the-art Core i7-2600 running at 3.4Ghz. The software platform is called Reconfigurable Active Solid State Drive Operating System (RASSDOS), and it is intended to be a real-time operating system that runs on the MicroBlaze®, which is a 32-bit RISC soft processor core designed for Xilinx® FPGAs. The RASSDOS platform is responsible for initializing the node and configuring the various components of the system. It then controls the node, including informing the middleware of the status of the node and listening for requests made by the latter. RASSDOS provides set of services for the middleware server which hide the low-level details of the node’s hardware architecture. The services provide routines to execute the following: access the data stored on the node, receive, store and load hardware accelerators or drivelets, retrieve results obtained by running the accelerators and-or drivelets on the node and to configure it as needed.