Overview | Tutorial Descriptions | Instructors

Tuesday, February 13, 2007 Tuesday Morning Half-Day Tutorials T1 Disc Drive Technology Dave Anderson, Director, Strategic Planning, Security Solutions, Seagate; Willis Whittington, Senior Manager, Product Marketing, Enterprise Storage Division, Seagate 9:00 a.m.–12:30 p.m. This tutorial will cover several topics related to disc drives. The first section will discuss the important components of a drive, and how a drive reads and writes data. It will touch on what causes errors in a drive and how the drive is architected to recover or compensate for them. The second section will address the various types of drives and how design decisions result in drives targeted for specific applications and markets. The third section will look to the future, explaining several important areas of drive-related research and how this work will affect drives of the future. Dave Anderson is Director of Strategic Planning for Seagate and has over 20 years experience in the computer field. For the last several years he has been involved in developing the opportunities for hard drives to contribute to system security. His responsibilities have included overall interface strategy for all disc interfaces. Dave was involved in the architecture and planning of Fibre Channel during its first proposal as a disc interface. He was one of the principal architects of the disc XOR commands that are now a part of the standard SCSI interface specification and was the author and original editor of the Object based Storage Device proposal that recently became a SCSI standard. Dave was one of the original nine elected members of the SNIA (Storage Networking Industry Association) Technical Council. He was also one of the founding members of the Serial Attached SCSI working group, which defined this new interface. He has been awarded 5 patents related to disc storage. Dave is a member of ACM and the IEEE Computer Society. Willis Whittington graduated BSc and MBA in the UK and has worked in the OEM disc drive business since 1967 at corporate design centers in Glasgow, London, Paris, and Frankfurt, as well as in the USA. He has been involved in all aspects of disc drive design and development from 300 Mbyte, 600 lb monsters in the early 70's, to today's 150 GByte Enterprise class drives which fit quite nicely in a shirt pocket. He is currently Product Marketing Manager with Seagate Technology's Enterprise Storage Division in Minnesota. T2 NFSv4 and Cluster File Systems Peter Honeyman, CITI, University of Michigan 9:00 a.m.–12:30 p.m. Symmetric parallel file systems coordinate sharing in a back end to allow multiple front-end nodes to present an identical view of storage. While stateless file servers, such as NFSv3, mesh well with cluster file systems, NFSv4 has delegations, locks, and other state that must be coordinated with other NFSv4 servers and the back end. This tutorial discusses the challenges and solutions in bolting NFSv4 servers to cluster file system nodes. We begin by describing the relevant features of NFSv4 and cluster file systems. Next, we cover the major coordination issues of locking, delegation, and shares, giving special attention to fair queuing for NFSv4, NLM, and local locks. We then explore options in client migration for cluster file systems. Finally, we discuss other issues in server replication and client migration. Peter Honeyman is Research Professor of Information at the University of Michigan, where he is Scientific Director of the Center for Information Technology Integration. As an experimental computer scientist, Honeyman builds middleware for file systems, security, and mobile computing. He has been instrumental in software projects including Honey DanBer UUCP, PathAlias, MacNFS, Disconnected AFS, and WebCard, the first Internet smart card. Current work centers on CITI's open source reference implementation of NFSv4 and its extensions for high end computing. Tuesday Afternoon Half-Day Tutorials T3 Cluster Storage and File Systems Technology Brent Welch and Marc Unangst, Panasas Inc. 1:30 p.m.–5:00 p.m. To meet the demands of increasingly hungry cluster applications, cluster-based distributed storage technologies are now capable of delivering performance scaling 10's to 100's of GB/sec. This tutorial will examine current state-of-the-art high performance file systems and the underlying technologies employed to deliver scalable performance across a range of scientific and industrial applications. The first half of the tutorial provides an in-depth description of the core features common across most high-performance file systems; including details of datapath design, decoupled and scalable metadata operations, data layout techniques, failover techniques, scalable reconstruction, storage interfaces, and security. The second half describes the design trade-offs found in both open-source and commercial solutions including Lustre, GPFS, Parallel NFS and Panasas. Brent Welch is Director of Software Architecture at Panasas. Panasas has developed a scalable, high-performance, object-based distributed file system that is used in a variety of HPC environments, including many of the Top500 super computers. He has previously worked at Xerox-PARC and Sun Microsystems Laboratories. Brent has experience building software systems from the device driver level up through network servers, user applications, and graphical user interfaces. While getting his PhD at UC Berkeley, he designed and built the Sprite distributed file system. Brent participates in the IETF NFSv4 working group, and is co-author of the pNFS internet drafts that specify parallel I/O extensions for NFSv4.1. Marc Unangst is a Software Architect at Panasas, where he has been a leading contributor to the design and implementation of the PanFS distributed file system. He represents Panasas on the SPEC SFS benchmark committee, and authored draft specification documents for the POSIX High End Computing Extensions Working Group (HECEWG). Previously, Marc was a staff programmer in the Parallel Data Lab at Carnegie Mellon, where he worked on the Network-Attached Storage Device (NASD) project. He holds a Bachelors of Science in Electrical & Computer Engineering from Carnegie Mellon. T4 NFS/RDMA: A Deep Dive Tom Talpey, Network Appliance, Inc. 1:30 p.m.–5:00 p.m. The Network File System (NFS) runs over Remote Direct Memory Access (RDMA) fabrics on several platforms, bringing dramatically higher filesharing performance to reality. We will explore the open standardized protocols that make this possible, review the interconnects they operate over, and describe both client and server implementation on open source systems. We will also explore the compelling performance advantages. Tom Talpey is a Technical Director with Network Appliance, Inc. He has been involved with NFS for much of his career, most recently being principally involved with specifying, standardizing and implementing the NFS/RDMA and NFSv4 minor version 1 protocol standards.