MACKLEM, R. Not Quite NFS, Soft Cache Consistency for NFS. In Proceedings of the 1994.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....address the issue of concurrent write sharing of files, and Slice as defined may provide weaker concurrent write sharing guarantees than some NFS implementations. However, the architecture is compatible with NFS file leasing extensions for consistent concurrent write sharing, as defined in NQNFS [13] and early IETF draft proposals for the NFS V4 protocol. 2.1 Related Work The Cambridge Universal File Server [5] proposed structuring a distributed file system as a separate name service and file block storage service. One system to take this approach was Swift [6] Slice is similar to Swift in ....

R. Macklem. Not quite NFS, soft cache consistency for NFS. In USENIX Association Conference Proceedings, pages 261 278, January 1994.

....back file attributes using NFS getattr and setattr operations to the directory servers. The proxy and other elements of our prototype provide no support for concurrent write sharing. However, our proposal is compatible with NFS file leasing extensions for consistent sharing, as defined in NQ NFS [18] and included in a recent IETF draft for NFS V4. 18 fileid record file uniq filelen data file Figure 2: Smallfile server data structures. 4.2 Directory Servers The Slice prototype directory servers store arbitrary collections of name entries indexed by hash chains, keyed by an MD5 ....

R. Macklem. Not quite NFS, soft cache consistency for NFS. In USENIX Association Conference Proceedings, pages 261-278, January 1994.

....in NFS implementations for concurrently shared files. We consider this to be acceptable since NFS V3 offers no firm consistency guarantees for concurrently shared files anyway. Note, however, that NFS V4 proposes to support consistent file sharing through a leasing mechanism similar to NQNFS [17]; it will then be sufficient for the proxy to propagate file attributes when a client renews or relinquishes a lease for the file. The current proxy bounds the drift by writing back modified attributes at regular intervals. Since the proxy modifies the contents of request and response packets, it ....

R. Macklem. Not quite NFS, soft cache consistency for NFS. In USENIX Association Conference Proceedings, pages 261--278, January 1994.

....and detection of write errors, force delayed writes to the server when a file is closed. This means that NFS clients get little advantage from delayed writes, and do not depend much on the update policy. More recent file service protocols, however, such as Sprite [9] Spritely NFS [15] and NQNFS [7], use explicit cache consistency protocols and so can benefit from delayed writes. 10 A BETTER UPDATE POLICY 1 10000 read response time (msec) 10 100 1000 0.1 100000 number of events 1 10 100 1000 10000 asynch write, PU, max = 5070 delayed write, PU, max = 8961 asynch write, ....

Rick Macklem. Not Quite NFS, Soft Cache Consistency for NFS. In Proc. Winter 1994 USENIX Conference, pages 261-278. San Francisco, CA, January, 1994.

....CPU, memory, and shared network interface is roughly equivalent to the price differential between IDE and SCSI storage today. Our buffer service prototype is implemented as a pair of loadable kernel modules using the stackable VFS file system interface [17] Both modules are stacked above NQNFS [16], an extended NFS file system implementation using leases [11] The server side module handles requests from clients, manages buffer server disk and memory, and issues read through and writeback requests to the backing servers. The client side module caches block location maps and redirects read ....

R. Macklem. Not quite NFS, soft cache consistency for NFS. In USENIX Association Conference Proceedings, pages 261--278, Jan. 1994.

....in which a callback scheme similar to that in Sprite [4] was added to NFS with the intention of providing strict cache consistency and improving performance by eliminating the overhead of refreshing the attribute cache with getattr. The ideas in Spritely NFS were used in modified form in NQNFS [3] which is a second protocol available from the NFS implementation of 4.4BSD. NQNFS differs from Spritely NFS in that the latter requires the server to keep state indicating the cache status of files at clients; however, NQNFS borrows the lease idea from Gray and Cheriton [2] in order to avoid ....

R. Macklem. Not Quite NFS, Soft Cache Consistency for NFS. In Proc. 1994 Winter USENIX, pages 261-- 278, January 1994.

No context found.

R. Macklem. Not Quite NFS, Soft Cache Consistency for NFS. In Proceedings of the Winter USENIX Technical Conference, pages 261--278, San Francisco, CA, January 1994.

.... and tradeoffs [4] The NQNFS (Not Quite NFS) system grants short term leases for cached objects, which increases performance by reducing both the quantity of data copied from the server to the client and the number of NFS calls that the client makes to check the consistency of their cached copies [10]. Unfortunately, adding such constructs to NFS version 2 or NFS version 3 requires non standard changes to the protocol. The NFS version 4 protocol does include a standard protocol for read leases, but has not achieved widespread deployment. Another approach to optimizing NFS read performance is ....

Rick Macklem. Not Quite NFS, Soft Cache Consistency for NFS. In Proceedings of the USENIX Winter 1994.

No context found.

MACKLEM, R. Not Quite NFS, Soft Cache Consistency for NFS. In Proceedings of the 1994.

No context found.

MACKLEM, R. Not Quite NFS, Soft Cache Consistency for NFS. In Proceedings of the 1994.

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