Wang, R. Y., Anderson, T. E., and Dahlin, M. D. (1998a). Experience with a Distributed File System Implementation. Technical Report UCB/CSD 98/986, University of California at Berkeley.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....tasks that are designed to occur during otherwise idle time. For example, in many systems, there are a variety of support tasks that scan large portions of disk contents, such as report generation, RAID scrubbing, virus detection, and backup. Another set of examples is the many defragmentation [15, 29] and replication [18, 31] techniques that have been developed to improve the performance of future accesses. A third set of examples is anticipatory disk activities such as prefetching [7, 11, 13, 19, 27] and prewriting [2, 4, 8, 10] Using simulation, our previous work explored two specific uses ....

R. Y. Wang, T. E. Anderson, and M.D. Dahlin. Experience with a distributed file system implementation. Technical Report CSD 98 986. University of California at Berkeley, January 1998.

....added to disk firmware for particularly large customers. In addition to the handful of real examples, there have been a number of research efforts that propose or would benefit from more expressive storage interfaces. For example, some have proposed to dynamically place data near the disk head [12, 54] or to piggyback write backs on rotational delays [4] Perhaps the clearest examples are recent proposals for object based storage [19, 18, 35] and those for Active disks [1, 41, 25] Object based storage defines a new storage interface that is much like a file system with a flat namespace; this ....

....on overall positioning overheads (seek time plus rotational latency) 49, 24] However, this may require that freeblock scheduling decisions be made by disk firmware. Fommately, the increasing processing capabilities of disk drives [1, 18, 25, 41] make advanced on drive storage management feasible [18, 54]. 5.3 Applications of Free Bandwidth Freeblock scheduling is a new tool, and we expect that system designers will find many unanticipated uses for it. This section describes some of the applications we see for its use. 13 Scanning applications. In many systems, there are a variety of support ....

[Article contains additional citation context not shown here]

Randolph Y. Wang, Thomas E. Anderson, and Michael D. Dahlin. Experience with a distributed file system implementation. Technical Report CSD-98-986. University of California at Berkeley, January 1998.

....file system the coherency protocol is necessarily a distributed algorithm, at least requiring communication and coordination between the client and the server. The protocol must deal with non atomic operations, out of order and lost messages, and machine or network failures. Several researchers [Chandra96, Chandra97, Wang98] provide examples of coherency protocols that are affected by these issues; we summarize the problems here. Ideally, coordination between caches would be an atomic operation; that way there would be no doubt about the resulting states of the two caches. In reality it is difficult to make ....

....Teapot, for example, produces output that can be verified by the Murphi [Dill92] verification system. Murphi performs an exhaustive search of the state machine s state space, and detects deadlock. Researchers have reported on the usefulness of Murphi in verifying protocols specified in Teapot [Chandra97, Wang98]. Murphi s input is a program that implements the state machine, and consists of data types, global variables, procedures, a set of rules, a set of invariants, and the initial values for the global variables (the machine s initial state) The values of the global variables as the Murphi program ....

R. Y. Wang, T. E. Anderson, and M. D. Dahlin. Experience with a Distributed File System Implementation. http://www.cs.berkeley.edu/~rywang/papers/spe98.ps, 1998.

....logging, it can lose track of free blocks in a crash, necessitating an occasional garbage collection scan to find them again. We are unable to compare the performance of their system with ours at present, as performance numbers for their file system layer are not available. The xFS file system [1, 36] comes closest in spirit to Frangipani. In fact, the goals of the two systems are essentially the same. Both try to distribute the management responsibility for files over multiple machines and to provide good availability and performance. Frangipani is effectively serverless in the same sense ....

Randy Wang, Tom Anderson, and Mike Dahlin. Experience with a distributed file system implementation. Technical report, University of California, Berkeley, Computer Science Division, June 1997.

No context found.

Wang, R. Y., Anderson, T. E., and Dahlin, M. D. (1998a). Experience with a Distributed File System Implementation. Technical Report UCB/CSD 98/986, University of California at Berkeley.

No context found.

R. Y. Wang, T. E. Anderson, and M. D. Dahlin. Experience with a distributed file system implementation. Technical Report CSD--98--986. University of California at Berkeley, January 1998.

No context found.

Randolph Y. Wang, Thomas E. Anderson and Michael D. Dahlin, "Experience with a Distributed File System Implementation", Computer Science Division, University of Califomia at Berkeley.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC