Dave Fellinger: Reducing Human Intervention in the Maintenance of Mass Storage Systems


Abstract

The bandwidth of servers has increased at the rate of Moore’s law over the last ten years and current interfaces have enabled communication and data sharing at the same rate. Infiniband coupled clusters served by multicore systems have become commodity standards. At the same time the random access bandwidth of disk drives has remained static over the same period.

The result of this dramatic mismatch is that storage systems have grown in size and complexity so that the aggregate performance of a large number of disk drives is used to satisfy the demands of clustered systems. The problem posed by this mismatch is that a large number of mechanical devices must be used to efficiently serve even the most conservative cluster architectures. In general the service requirements of I/O systems greatly exceed the requirements of the cluster considering the quantity of elements of each system and the mechanical uncertainties of the storage system. This problem is exacerbated by the current market demand for the most cost effective disk drives available lowering the overall system reliability. A storage architecture will be discussed that limits these service requirements and the resultant human interface.

About the speaker

Dave Fellinger is the Chief Technical Officer [CTO] at DataDirect Networks, Inc., the world’s leading provider of networked storage and cluster solutions for high performance and eResearch computing environments and requirements - and one of the key architects of DataDirect Networks’ Silicon Storage Appliance (S2A) storage controller technology. Dave has over thirty years of experience in engineering including film systems, ASIC design and development, GaAs semiconductor manufacture, RAID and storage systems, and video processing devices. His experience includes serving as Vice President of Engineering for Ultimatte Corporation, where his products won several awards including an Oscar and an Emmy. Dave attended Carnegie-Mellon University (Electrical Engineering and Physics) and holds numerous patents in electrical engineering, optics, motion control, video processing, and pattern recognition.