P. Dasgupta et al. The design and implementation of the Clouds distributed operating system. Computing Systems, 3(1), Winter 1990.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....memory. Our work is orthogonal, because we are examining the low level mechanisms that are necessary to manage resources in a Java operating system. Many research projects have explored operating systems issues within the context of programming languages. For example, Argus [33] and Clouds [15] explored the use of transactions within distributed programming languages. Other important systems that studied issues of distribution include Eden [2] Emerald [9] and Amber [11] These systems explored the concepts underlying object migration, but did not investigate resource management. ....

P. Dasgupta et al. The design and implementation of the Clouds distributed operating system. Computing Systems, 3(1), Winter 1990.

....system s nodes. Ideally, such a place ment must be made so as to optimize the two conflicting criteria that are the number of message exchanges and the processors loads. In general, processor allocation policies offered by existing distributed object based systems such as Argus [34] Clouds [21], Doom [36] Emerald [39] and Guide [31] are simple. The node to be used for executing a method is selected among one of the following: object s storing node, calling node, and node explicitly mentioned by the programmer. A study has been undertaken in our group in order to integrate a ....

P. Dasgupta, R. C. Chen, S. Menon, M. P. Pearson, R. Ananthanrayanan, R. J. LeBlanc, M. Ahamad, U. Ramachandra, W. F. Appelbe, J. M. Bernab'eu-Auban, P. W. Hutto, M. Y. A. Khalidi, and C. J. Wilkenoh. The design and implementation of the Clouds distributed operating system. Computing Systems, 4(3):243--275, 1990.

....RPC, and included additional data transfer optimizations. This makes it difficult to isolate the benefits of the improved control transfer. Object oriented systems have traditionally distinguished between active and passive objects, corresponding to static and migrating thread models[9] Clouds[16] exemplifies a passive object (migrating thread) model, while Emerald[5] as we do, provides both active and passive objects support for both styles of execution. Chorus[26] can use only thread switching between user level tasks, but between tasks running in the kernel s protection domain it ....

Partha Dasgupta et al. The design and implementation of the Clouds distributed operating system. Computing Systems, 3(1), Winter 1990.

....was kept sequentially consistent [16] Although Ivy performed well on some applications, it performed poorly for programs with significant degrees of sharing, which limited its applicability. Since this time, the driving goal of DSM research has been to improve performance by any means necessary [1, 2, 4, 7, 9, 11, 13, 14, 21]. The performance of DSM systems is dominated by two issues: i) the communication overhead required to maintain consistency and (ii) the computation overhead required to service remote requests and compute the information needed to maintain consistency. Over the years, there has been extensive ....

....Over the years, there has been extensive research aimed at dramatically reducing the communication overhead required to keep DSM managed data consistent. These efforts have included: ffl exploiting object encapsulation to give programmers the ability to precisely specify the coherence unit [1, 2, 7, 9], ffl allowing programmers to explicitly lock data to specific nodes and thereby control at a fine grain when data is communicated [9, 11] ffl supporting multiple programmer selected coherence protocols [4] ffl supporting multiple concurrent writers to eliminate false sharing [4, 14] and ....

[Article contains additional citation context not shown here]

P. Dasgupta, et al., The design and implementation of the Clouds distributed operating system. Computing Systems Journal, 3, Winter 1990.

....developments in operating systems as well as in programming languages and databases. One emerging model for the support of co operative distributed applications is that of a distributed shared universe organized as a set of passive objects (active agents are define outside of objects) Bal 92] [Dasgupta 90], Liskov 92] In this paper we report on our experience in designing, implementing, and using a system to support such a model. Our goal is to provide an efficient platform for a family of objectoriented languages such as Guide (a language designed by our group [Krakowiak 90] and a persistent ....

P. Dasgupta, R.C. Chen, S. Menon, M.P. Pearson, U. Ananthanarayanan, U. Ramachandran, M. Ahamad, R.J. LeBlanc, W.F. Appelbe, J.M. Bernabeu-Auban, P.W. Hutto, M.Y.A. Khalidi, C.J. Wilkenloh. The Design and Implementation of the Clouds Distributed Operating System, Computing Systems , vol. 3, 1, pp. 11-46

....combine the best features of shared memory and distributed memory machines. They support the convenient shared memory programming model on distributed memory hardware, which is more scalable and less expensive to build. However, although many DSM systems have been proposed and implemented (e.g. [3, 5, 8, 10, 12, 21, 22]) achieving good performance on DSM systems for a sizable class of applications has proven to be a major challenge. This challenge can be best illustrated by considering how a conventional DSM system is implemented [21] The global shared address space is divided in virtual memory pages. The ....

....a thread attempts to write to replicated data, a message is transmitted to invalidate all other copies of the data. The thread that generated the miss blocks until all invalidation messages are acknowledged. This single owner consistency protocol is typical of what existing DSM systems provide [10, 12, 21], and is what we use exclusively to represent a conventional DSM system in our performance evaluation. Once read only data has been initialized, no further updates occur. Thus, the consistency protocol simply consists of replication on demand. A runtime error is generated if a thread attempts to ....

[Article contains additional citation context not shown here]

P. Dasgupta, R.C. Chen, S. Menon, M. Pearson, R. Ananthanarayanan, U. Ramachandran, M. Ahamad, R. LeBlanc Jr., W. Applebe, J.M. Bernabeu-Auban, P.W. Hutto, M.Y.A. Khalidi, and C.J. Wileknloh. The design and implementation of the Clouds distributed operating system. Computing Systems Journal, 3, Winter 1990.

....an SDSM system was Ivy[32] It uses a single writer, invalidation based coherence protocol, with virtual memory pages as the unit of coherence. The problem of multiple writers to different parts of a shared page, called false sharing, caused frequent and costly page movement between hosts. Clouds [14] reduced the effects of false sharing by giving the programmer the ability to lock a data object for the duration of its use. Mirage [20] similarly automatically locked pages for short periods of time to provide the same effect without programmer intervention. These software DSM systems ....

P. Dasgupta, R.C. Chen, S. Menon, M. Pearson, R. Ananthanarayanan, U. Ramachandran, M. Ahamad, R. LeBlanc Jr., W. Applebe, J.M. Bernabeu-Auban, P.W. Hutto, M.Y.A. Khalidi, and C.J. Wileknloh. The design and implementation of the Clouds distributed operating system. Computing Systems Journal, 3, Winter 1990.

....and a limited pointer directory based (NEWS) protocol for a mesh interconnect. Rather than storing exact pointer the copy set only tells the probable existence of the copies of a data page in one of the four directions at a mesh point (similar to Li s probable owner concept) Clouds Clouds [29, 30] is an object based distributed operating system. In [73] Ramachandran et al. have proposed implementation of distributed shared memory in the Clouds kernel. This work is based on Li s work but provides additional efficient support for synchronisation by integrating synchronisation with ....

P. Dasgupta, R. Chen, S. Menon, M. Pearson, R. Ananthanarayanan, U. Ramachandran, M. Ahamad, R. L. Jr., W. Applebe, J. M. Bernabeu-Auban, P. Hutto, M. Khalidi, and C. J. Wileknloh. The Design and Implementation of the Clouds Distributed Operating System. In Computing Systems Journal, Winter 1990.

....communicate has been established as a natural interface for parallel programming. Distributed systems do not support global memory in hardware, and hence, this feature has to be implemented in software. While systems built around Distributed Shared Memory (DSM) like IVY [34] Munin [9] Clouds [16, 20, 18], Mether NFS [35] provide a more natural programming model, they still suffer from the high cost of distributed synchronization and the inability to provide suitable fault tolerance. A mature system that uses a variant of the DSM concept is Linda [14] Instead of having a global address space, ....

P. Dasgupta, R. C. Chen, S. Menon, M. P. Pearson, R. Ananthanarayanan, U. Ramachandran, M. Ahamad, R. J. LeBlanc, W. F. Appelbe, J. M. Bernab'eu-Aub'an, P. W. Hutto, M. Y. A. Khalidi, and C. J. Wilkenloh. The design and implementation of the Clouds distributed operating system. Computing Systems, 3, 1990.

Documents 51 to 59  Previous 50

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