A. Birrell, G. Nelson, S. Owicki, and E. Wobber. Network objects. Technical report, SRC Research Report 115, Digital Systems Research Center, 1995.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....reference counting algorithm is the most suitable for a port over XML protocol. The issues to consider are the number of messages to be exchanged when contracts are communicated between agents, and the directions of these messages [32] Class Loading Java RMI [25] extends Network Objects [6] by its ability to load code dynamically. Such facility is used by SoFAR to load the code of ontology terms dynamically, when received by agents. Let us remind the reader that this code is generated by the ontology compiler and provides important methods for matching terms and for visitor ....

A. Birrell, G. Nelson, S. Owicki, and E. Wobber. Network Objects. Technical Report 115, Digital Systems Research Center, Feb. 1994.

....for roaming users. Unfortunately, the storage capacity of these devices is small compared to fixed installations, so mobile users generally rely on connectivity with their home network for storage of larger amounts of data, e.g. by using a distributed file system like AFS [8] Coda [3] or Echo [1]. This dependence on a remote system introduces reliability problems (e.g. network unreachable) quality of service problems (e.g. jitter in audio or video) and security problems (e.g. eavesdropping) that have to be solved. The problems of reliability and quality of service are greatly reduced ....

A. D. Birrell, A. Hisgen, T. Mann C. Jerian, and G. Swart. The echo distributed file system. SRC Research Report 111, Digital Systems Research Center, 1993.

....(OODP) is a generalization of the client server model, because objects encapsulate an internal state and make it accessible through a well defined interface. A distributed system ensures complete encapsulation and method calls are a convenient place to insert the required communication [Birrell 93] 3.3.1 Programming in PVM and MPI The programming model behind PVM and MPI has simplified message passing programming since it harness the existing hardware and software to build and run parallel programs. While this is by itself a step forward, because no special hardware and 8 software are ....

Birrell, A., Nelson, G., Owicki, S., and Wobber, E. Network Objects. Technical Report 115. Digital Systems Research Center, Palo Alto, CA, February 1994.

....sharing of documents and parts of documents. 4.6.1 Global versus Context Dependent Naming If names are used for communication about objects, then they must either be valid in a global context or they must be accompanied by a representation of the context in which they are to be resolved. Birrell [21] suggests a global hierarchic name space concept originally developed for a distributed file system. The goal of global naming is to ensure that a specific name will have the same meaning to other users and programs in the same distributed system. The name space is a tree of labeled arcs with a ....

Birrell, A. D., Hisgen, A., Jerian, C., Mann, T., and Swart, G. The Echo Distributed File System. Tech. rep., Digital Systems Research Center, Palo Alto, 1993.

....using small files, with little concurrent sharing seems to be the standard workload assumed by most research file system designs. Other studies, even those that draw data from many different environments [7] confirm that this workload is very common. One notable exception is Echo [6]. Through the use of read and write tokens, Echo provides replicated service and really does serialize reads and writes. Every reader sees the most recent data. This clean and strong guarantee comes at the cost of high complexity. Implementing any level of consistency in a replicated system ....

T. Mann, A. Hisgen and G. Swart. An Algorithm for Data Replication. Technical Report 46, Digital Systems Research Center, June 1989.

....for the file system prototype. Bill Schilit s labor produced the file traces used in the trace driven experiments described in this dissertation. vi 1 1 Introduction Distributed file systems are a fundamental structure of distributed computing, and much attention has been focused on their design [3, 9, 13, 18, 29, 30, 31, 44, 51, 53]. But one important new design point has not yet been explored thoroughly: how to support mobile clients. Portable workstations are becoming increasingly common, and people who use these machines should not be forced to accept inferior performance or usability. This leaves an open question for ....

....and the file is in its cache, it delivers its value to the reader. Thereafter, strict read can be performed by direct client to client communication, reverting to the usual method only when the PCW loses its CT. Write Write Sharing and File Locking. Unlike systems such as Deceit [53] and Echo [30], we do not use the concept of a write token in our design. This 36 means that two PCWs may simultaneously modify their cached copies of a file, and this will be detected as a conflict during update propagation. Although we have not provided explicit support for file locking, it would be ....

[Article contains additional citation context not shown here]

T. Mann, A. Hisgen, and G. Swart. An Algorithm for Data Replication. Technical Report 46, Digital Systems Research Center, June 1989. 119

....1 Introduction This work investigates how to maintain replicas in a distributed file system, especially one supporting mobile clients. While the topic of replica management within file systems has received so much attention that one might think there is no design point left unconsidered [3, 16, 14, 2, 9, 18, 1, 10, 4], the notion of mobile clients is a new development that alters operating circumstances and therefore suggests new designs. The idea of mobile file service clients stems from the exploding popularity of portable computers. We argue that existing file replica management schemes would not cope well ....

....caching, it is possible for 1USR clients to make irreconcilable updates to a file without any knowledge that they are doing so. For example, two clients may both strictly read a file and then make conflicting updates into their caches. 1SR is not attainable using our design (or most others) Echo [18] is an experiment in providing 1SR via a file system. Division of read into loose read and strict read ensures that the entire cost of establishing 1USR consistency is charged to the process (reader) that demands the consistent value. This approach is in contrast to the other systems we know of, ....

[Article contains additional citation context not shown here]

T. Mann, A. Hisgen, and G. Swart. An Algorithm for Data Replication. Technical Report 46, Digital Systems Research Center, June 1989.

....one, given the likely marriage between two exploding trends: portable computers and wireless networks. In a previous paper [14] we have argued that client mobility is a major new development that requires re thinking file system design, and that existing approaches to replica management (e.g. [7, 13]) would not cope well with mobile clients. In that paper we proposed an alternative: a lazy server based update scheme and a new service interface that allows applications to select strong or weak consistency semantics on each particular read call. Our earlier work addressed only some of the ....

T. Mann, A. Hisgen, and G. Swart. An Algorithm for Data Replication. Technical Report 46, Digital Systems Research Center, June 1989.

No context found.

A.D. Birrell, A. Hisgen, C. Jerian, T. Mann, and G. Swart. The Echo distributed file system. Technical Report 111, Digital Systems Research Center, September 1993.

No context found.

A. Birrell, G. Nelson, S. Owicki, and E. Wobber. Network objects. Technical report, SRC Research Report 115, Digital Systems Research Center, 1995.

No context found.

A.D. Birrell, A. Hisgen, C. Jerian, T. Mann, and G. Swart. The Echo distributed file system. Technical Report 111, Digital Systems Research Center, September 1993.

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