Storage System Architectures
CSE226 Winter 2007

Text

The textbooks (recommended but not required) are:

• "High Performance Mass Storage and Parallel I/O." by Hai Jin, Toni Cortes, Rajkumar Buyya. Wiley InterScience.

• "Algorithmic Studies in Mass Storage." by C.K. Wong. Computer Science Press.

• "Course Reader." The set of notes is available at Soft Reserves, Student Center A.

Reading List  (partial)

Readings for material will be posted as we go along. We will have at least one guest lecture during the quarter.

• Disk Modeling
Ruemmler & Wilkes, An Introduction to Disk Drive Modeling, IEEE Computer 1994, 17-28.
Anderson, Dykes & Reidel, More than an interface-- SCSI vs. ATA, FAST 2003, 245-257.
Worthington, Ganger, Patt & Wilkes, On-Line Extraction of SCSI Disk Drive Parameters, Sigmetrics 1995,146-156.
• Disk Performance
Ng, Impact of Disk Technology: Performance Issues,IEEE Computer, May 1998, 75-81.
Response time modeling, Course notes, 6-7.
RAIDframe: A rapid prototyping tool for RAID systems, August 1996, www.pdl.cmu.edu.
• Disk Arrays
Patterson, Gibson, Katz, A case for redundant arrays of inexpensive disks (RAID), Sigmod, 1988, 109-116.
Colarelli, Grunwald, Massive arrays of idle disks for storage archives, Supercomputing, 2002, 1-11.
• Information Dispersal
Rabin, Efficient dispersal of information for security, load balancing and fault tolerance, Journal of the ACM, 1989, 335-348.
• Reliability
Reliability Analysis, notes pages 17-20.
Probabilistic Reliability: An Engineering Approach, Shooman, Krieger Publishing, 1990 (reference only.)
Schroeder, Gibson, Disk failures in the real world: what does a MTTF of 1,000,000 hours mean to you?, FAST'07
• Declustering i
Holland, Gibson, Parity declustering: for continuous operation in redundant disk arrays, ASPLOS 1992, pages 23-35.
Alvarez et. al., Tolerating multiple failures in RAID architectures with optimal storage and uniform declustering, ISCA 1997, pages 62-72.
• Declustering ii
Alvarez et. al., Tolerating multiple failures in RAID architectures with optimal storage and uniform declustering, ISCA 1997, pages 62-72.
Alvarez et. al., Declustered disk array architectures with optimal and near-optimal parallelism, ISCA 1998, pages xx-xx.
• Declustering iii
Alvarez et. al., Declustered disk array architectures with optimal and near-optimal parallelism, ISCA 1998, pages xx-xx.
Schwarz et. al., Permutation Development Data Layout (PDDL), HPCA 1999, pages 214-217.
• Distributed File System
Weil et. al., CRUSH: Controlled, Scalable, Decentralized Placement of Data, SC 2006.
Weil et. al., Ceph: A scalable, high-performance distributed file system, OSDI 2006, pp 307- .
• Data Layouts
Honicky and Miller, Replication Under Scalable Hashing: A Family of Algorithms for Scalable Decentralized Data Distribution, IPDPS 2004.
• Google file system
Ghemawat et al., The google file system, SOSP 2003, pp 29-43.
• GPFS
Schmuck and Haskin, GPFS: A shared-disk file system for large computing clusters, FAST 2002, pp 231-244.
• Virtual file system
Lee and Thekkath, Petal: distributed virtual disks, ASPLOS 1996, pp 84-96.
• Data Layouts
Theodoris et al., B: Disk array data layout tolerating multiple failures, MASCOTS 2006, pp 21-32.
• Reliability revisted
• Peer-to-Peer