J. Israel, J. Mitchell, H. Sturgis, "Separating Data from Function in a Distributed File System", Proc. 2nd Int'l Symp. on Operating Systems, IRIA, 1978; to appear in D. Lanciaux, ed., Operating Systems, North Holland.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....of specific citations throughout the report, devote Chapter 7 entirely to discussion of related work. Here is a short summary of some of the more closely related research. Our locking method is an extension of the well known two phase locking protocol [EGLT76] used in many systems (e.g. Gray78, IMS78, LeLann81] Our object state restoration (transaction undo) algorithm is also based on methods used in many systems (e.g. Gray78, IMS78, SMB79, Paxton79, LeLann81] The distributed transaction management algorithm is original, except in its incorporation of a two phase commit protocol ....

....more closely related research. Our locking method is an extension of the well known two phase locking protocol [EGLT76] used in many systems (e.g. Gray78, IMS78, LeLann81] Our object state restoration (transaction undo) algorithm is also based on methods used in many systems (e.g. Gray78, IMS78, SMB79, Paxton79, LeLann81] The distributed transaction management algorithm is original, except in its incorporation of a two phase commit protocol [Gray78, Gray80a, Paxton79, IMS78, Reed78, LeLann81, Lindsay79, HS80] The distributed deadlock detection algorithm is an extension of work ....

[Article contains additional citation context not shown here]

J. Israel, J. Mitchell, H. Sturgis, "Separating Data from Function in a Distributed File System", Proc. 2nd Int'l Symp. on Operating Systems, IRIA, 1978; to appear in D. Lanciaux, ed., Operating Systems, North Holland.

....Process table 4 Level 3 2 1 0 A more complex constellation of processes exists in FileSuites, which constructs a single replicated file from a set of representative files, each containing data from some version of the replicated file. The representatives are stored in a transactional file system [11], so that each one is updated atomically, and each carries a version number. For each FileSuite being accessed, there is a monitor that keeps track of the known representatives and their version numbers. The replicated file is considered to be updated when all the representatives in a write quorum ....

Israel, J.E., Mitchell, J.G., and Sturgis, H.E. Separating data from function in a distributed file system. Second Int. Symposium on Operating Systems, IRIA, Rocquencourt, France, Oct. 1978.

....as a concurrency control method for our implementation. Three ways to serialize the concurrent access to shared data are: ffl Locking Using this approach, a server sets a lock on each data item of a file before it is accessed on behalf of the first transaction which wants to access it [24]. Only the transaction which locked the data item (a lockable unit) can access it. Other transactions must wait until the data item is unlocked. The data item is unlocked when the transaction is completed (committed or aborted) ffl Optimistic concurrency control Kung and Robinson [26] have ....

....detection algorithms in the distributed system have been proved incorrect [20] On the other hand, the lock timeouts are a method for resolution of deadlocks that is commonly used in file servers [12] for example, Tandem. Therefore based on our requirement and the result of an analysis of XDFS [24], we propose to use timeout to resolve the deadlock in the RHODOS system. If not detected, a deadlocked and or permanently blocked transaction could take a long time. A transaction can also take long time if it is nested or highly computation oriented. This implies that until it is proved that a ....

Israel, J.E., Mitchell, j.G., and Sturgis, H.E. "Separating data from function in a distributed file system", Operating Systems: Theory and Practice, (Ed. Lanciaux, D.), North-Holland, Amsterdam, pp. 17-27, 1978.

....requirements and recommendations for the design and evaluation of trusted recovery mechanisms and assurance. The reader who is unfamiliar with the notion of system recovery and security modeling required of B3 and Al systems may find it useful to refer both to the recovery literature (such as [1, 5, 14 16, 20 23, 25, 27]) and the security literature (such as [3,11, 26, 29] cited in this guide. 1.3 SCOPE Trusted recovery refers to mechanisms and procedures necessary to ensure that failures and discontinuities of operation don t compromise a system s secure operation. The guidelines for trusted recovery ....

....with respect to the set of expected failures. Most commercially available I O hardware satisfies this requirement. This scheme, called shadowing, was conceived by Lampson and Sturgis in 1976, was reported in detail in [21] and was implemented in several distributed storage systems at Xerox PARC [20, 23, 25, 27]. 4.3.3.2 Logging Mechanisms which assure that a TCB primitive exhibits transaction atomicity and which are based on logging assume that nonvolatile storage exists which can be made reliable enough to survive all expected failures, e.g. by storage duplexing. With logging, each update of a TCB ....

Israel, J., J. Mitchell, and H. Sturgis, "Separating Data from Function in a Distributed File System," Proceedings of the Second International Symposium on Operating Systems, IRIA, Rocquencourt, France, October 1978.

.... that the transaction is committed by all servers even in the case of a server crash in the middle of the transaction [Coulouris and Dollimore 88] but in general, this method works and has been implemented in a number of systems (to my knowledge, first in the Xerox Distributed File System [Israel et al. 78] An earlier version of Hyper G also adopted this method [Kappe 93] However, the multi server transaction has scalability problems in certain situations. When for some reason many servers (say, 1000) decide to refer to a specific document (e.g. by pointing a link to it or by replicating it) ....

Israel, J. E., Mitchell, J. G., Sturgis, H. E.: "Separating Data from function in a Distributed File System". In Lanciaux, D. (editor), Operating Systems: Theory and Practice, 17--27. North-Holland, Amsterdam (1978).

No context found.

Israel, J.E., Mitchell, J.G. and Sturgis, H.E. Separating data from function in a distributed file system. Proc. Second International Symposium on Operating Systems, IRIA, Rocquencourt, France, Oct 1978.

No context found.

Israel, J.E., Mitchell, J.G. and Sturgis, H.E. Separating Data from Function in a Distributed File System, in Operating Systems: Theory and Practice. (Ed. Lanciaux D.), North-Holland, pp. 17-27

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