SBIR Phase II: Scalable and Reliable Storage Infrastructure for Network Storage Environments

This Small Business Innovation Research (SBIR) Phase II project will build a scalable and reliable storage system for network storage environments. This outcome of this project is a revolutionary system that employs a combination of unique ideas to address the main challenges encountered in today's demanding storage environments namely scalability, availability, performance, and manageability. The ideas of this proposed solution are applied as a disk-based solution for the time-consuming network backup/restore problem. With the rapid growth of data-driven network services, traditional storage solutions are not able to keep pace with the rapidly expanding storage requirements. Unlike traditional solutions the proposed solution employs a new architecture that allows for independent and practically unlimited scalability of capacity, file access performance, and namespace access performance. The proposed product utilizes a unique, very fast coding technique called PND to ensure fast, reliable, and highly available access to data. It offers the opportunity of applying a more effective block-level edge caching technique, which enhances the performance and achieves better utilization of the valuable cache memory. It takes advantage of Data Reliability, Inc.'s innovative RAISTM storage engine to cost-effectively aggregate distributed islands of independent storage resources into a single virtual shared pool of storage. Project Phase I has clearly demonstrated the above advantages. Many applications will exploit the competitive advantages of the proposed product including Web server farms; multimedia network services, content management, document storage and delivery, digital imaging, and file transfer services. In addition, the expected solution's ideas can be expanded to build general-purpose file servers that are not subject to performance bottlenecks and capacity limitations. Therefore, these ideas will have an important impact on building next generation NAS devices. The PND technique, pioneered by this project, provides a new class of codes that are expected to result in scientific advances in coding theory. In addition, the PND technique will contribute to enhanced performance and architectures of disk arrays. Applications of PND coding in areas other than data storage include mobile communications, reliable multicasting, audio/video streaming, and digital fountain systems. The company is partnering with Jackson State University (JSU) and will offer JSU students a tremendous educational experience. Since Jackson State University is an HBCU (Historically Black College and University) in the underrepresented state of Mississippi, the project will foster continuous collaboration and will increase the participation of underrepresented and minority groups in science and technology.