Aloke Gupta and W. K. Fuchs. Garbage collection in a distributed objectoriented system. IEEE Transactions on Knowledge and Data Engineering, 5(2), April 1993.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....method at that time will be started. Each ProcessManager records the times of the Wakeup events it has scheduled so that multiple Wakeup events for the same simulation time can be avoided. 4.4. 3 Garbage Collection In APOSTLE, garbage collection is implemented using reference count techniques [Gupta 93] The events IncRefCount and DecRefCount are used to increment and decrement the refer Figure 4.7: The ProcessManager s main processing loop. void ProcessManager: performProcessing( do runnableMethod = false; execute all methods for the current simulation time runMethods( ....

A. Gupta & W. K. Fuchs, "Garbage Collection in a Distributed ObjectOriented System", IEEE Trans. on Knowledge and Data Engineering, April 1993, 5(2):257-265.

....in ML programs [24] In distributed object databases, reclaiming distributed cyclic garbage will be a significant concern. While our current work does not directly address this issue, there is a large body of previous work that has explored reclaiming distributed cyclic garbage (e.g. see [17], 26] Our work can be extended with these techniques. Partition based collection algorithms are an extension of generational algorithms used in primary memory garbage collection. In a 1977 dissertation, Bishop investigated the possibility of garbage collection in a very large address space ....

....of database connectivity. Fig. 15. Total storage required as a function of maximum allocated storage. 170 IEEE TRANSACTIONS ON KNOWLEDGE AND DATA ENGINEERING, VOL. 10, NO. 1, JANUARY FEBRUARY 1998 While there has been some work done in handling distributed cyclic garbage in distributed systems [17], 26] which can be applied to partitioned collection, previous work in partitioned collection has maintained that cross partition cycles will probably not be a problem [33] We have seen, however, that even small increases in connectivity can produce significant amounts of distributed cyclic ....

A. Gupta and W.K. Fuchs, "Garbage Collection in a Distributed Object-Oriented System," IEEE Trans. Knowledge and Data Eng., vol. 5, no. 2, pp. 257--265, Apr. 1993.

....backward references are calculated using these tables and local tracing. The overhead of calculating backward references is heavy, and it is very hard to preserve safety and completeness in the presence of concurrent mutators and garbage collectors. Migration [Bishop, 1977; Shapiro et al., 1990; Gupta, 1993] was also suggested to collect cyclic garbage. In this algorithm, all objects on a garbage cycle are migrated to a single system, and are collected during local garbage collection. This approach is not suitable for large scale distributed systems, because some objects may not be migrated, and some ....

Aloke Gupta, W. Kent Fuchs, Garbage collection in a distributed object-oriented system, IEEE Transaction on Knowledge and data engineering, vol. 5, no. 2, April 1993, pages. 257-265.

....with e, while b 2 will not be migrated. This is desirable because b 2 is likely to be live. b b2 site T e e2 b: S 1 e: S 20 inrefs Figure 6.7: Objects traced from a suspected inref are batched for migration. 6.2. 1 Where to Migrate Some previous schemes migrate objects to a fixed dump site [GF93] but this can be a performance or fault tolerance bottleneck in a large system. Other schemes migrate objects to sites that refer to them. To ensure that objects in a cycle converge on the same site instead of following each other in circles, sites are totally ordered and migration is allowed in ....

A. Gupta and W. K. Fuchs. Garbage collection in a distributed object-oriented system. IEEE Transactions on Knowledge and Data Engineering, 5(2), 1993.

....self referential garbage structures that cross partition boundaries) Our collectors do not reclaim these garbage structures, although our future work will investigate this phenomenon in more detail. While there has been some work done in handling distributed cyclic garbage in distributed systems [12], which can be applied to partitioned collection, previous work in partitioned collection has maintained that crosspartition cycles will probably not be a problem [23, 25] We have seen, however, that even small increases in connectivity can produce significant amounts of distributed cyclic ....

Aloke Gupta and W. Kent Fuchs. Garbage collection in a distributed object-oriented system. IEEE Transactions on Knowledge and Data Engineering, 5(2):257--265, April 1993.

....self referential garbage structures that cross partition boundaries) Our collectors do not reclaim these garbage structures, although our future work will investigate this phenomenon in more detail. While there has been some work done in handling distributed cyclic garbage in distributed systems [16], which can be applied to partitioned collection, previous work in partitioned collection has maintained that cross partition cycles will probably not be a problem [28, 31] We have, however, seen that even small increases in the connectivity of the database can produce significant amounts of ....

Aloke Gupta and W. Kent Fuchs. Garbage collection in a distributed object-oriented system. IEEE Transactions on Knowledge and Data Engineering, 5(2):257--265, April 1993.

....object stores, even small amounts of uncollected garbage can cause significant storage loss over time. The problem can be solved either by using a complementary marking scheme, or by migrating objects so that cyclic garbage ends up in a single node and is collected by the local collector [Bis77, SGP90, GF93]. Migration based schemes, like distributed reference counting, are decentralized and fault tolerant: The collection of a cycle requires the cooperation of only those nodes that contain it, and progress is made even if other nodes or other parts of the network fail. Existing migration schemes have ....

....the cooperation of only those nodes that contain it, and progress is made even if other nodes or other parts of the network fail. Existing migration schemes have some practical problems. Most schemes migrate all locally unreachable objects, although they may be reachable from roots on other nodes [Bis77, GF93]. Migration of live objects is undesirable because it wastes processor and network bandwidth. Worse, it interferes with object placement and load balancing. This paper presents simple techniques to avoid unnecessary migration. We use a heuristic to detect objects that are highly likely to be ....

A. Gupta and W. K. Fuchs. Garbage Collection in a Distributed Object-Oriented System. IEEE Transactions on Knowledge and Data Engineering, Vol. 5, No. 2, April 1993.

....Department of Defense, monitored by the Office of Naval Research under contract N00014 91 J 4136. using a complementary marking scheme to collect cyclic garbage [Ali84, JJ92, LQP92] or by migrating objects so that cyclic garbage ends up in a single node and is collected by the local collector [Bis77, SGP90, GF93]. The advantage of migration is that, like distributed reference counting, it is decentralized and fault tolerant. The collection of a cycle requires the cooperation of only those nodes that contain it, and progress can be made even if other nodes or other parts of the network fail. Therefore, ....

....objects to migrate between nodes. However, existing migration schemes have some practical problems. First, they tend to migrate live objects along with garbage. Most schemes migrate locally unreachable objects, either immediately [Bis77, Bag91] or if the objects are not used for some time period [GF93]. In a persistent store, however, live objects may not be accessed for long periods (say, weeks or months) so even systems that wait will migrate live objects. Migration of live objects is undesirable because it wastes processor and network bandwidth. Also, it interferes with load balancing. For ....

[Article contains additional citation context not shown here]

A. Gupta and W. K. Fuchs. Garbage Collection in a Distributed Object-Oriented System. IEEE Transactions on Knowledge and Data Engineering, Vol. 5, No. 2, April 1993.

....of Washington, Mach [45] at Carnegie Mellon University, V [46] at Stanford University, Locus [47] at the University of California, Los Angeles, and Clouds [48] at the Georgia Institute of Technology. Other projects in distributed systems have evolved algorithms for distributed object management [49, 50, 51]. 1.3 The X Window System The X Window System [12, 13, 14, 15] has emerged as the de facto standard in the Unix workstation domain. It was developed in the early 1980s as part of project Athena at MIT. Portability was an early design concern in X Window since MIT was using a diverse group of ....

A. Gupta and W. Fuchs, "Garbage collection in a distributed object-oriented system, " IEEE Transactions on Knowledge and Data Engineering, vol. 5, April 1993.

No context found.

Aloke Gupta and W. K. Fuchs. Garbage collection in a distributed objectoriented system. IEEE Transactions on Knowledge and Data Engineering, 5(2), April 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