P. Dasgupta, R. J. LeBlanc Jr. and W. F. Appelbe, `The Clouds distributed operating system: functional description, implementation details and related work', 11th IEEE Conf. on Distributed Computing Systems, 1988, pp. 2--9.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....a common process execution and communication environment. The Chorus microkernel (also termed nucleus) is the only choice applications have. In PEACE, the nucleus family presents an assortment of up to eight different members. Ra [2] is a minimal kernel for the Clouds distributed operating system [8]. The Ra kernel is designed to support the implementation of large scalable object based systems. Ra is a fairly complex minimal kernel too, implementing segment based virtual memory management and short term scheduling. At best, Ra can be compared to the PEACE nucleus instance that provides task ....

P. Dasgupta, R. J. LeBlanc Jr., W. F. Appelbe, "The Clouds Distributed Operating System: Functional Description, Implementation Details and Related Work", Proceedings of the 8th International Conference on Dis- tributed Computer Systems, pp. 2-9, IEEE, San Jose, CA (USA), June, 1988

....of the central servers. Grapevine [2] and Clearinghouse [11] are two well known systems using this approach. A few solutions in a local area network environment have used broadcast, either to broadcast a request for information, or to broadcast information updates to all 2 other machines. Clouds [5] uses the former approach for locating remote objects, and the rwhod protocol in the UNIX 1 operating system [4] uses the latter approach for managing user and machine status information. These systems work on a small scale, but broadcasting requests and updates can overwhelm machines as the ....

Partha Dasgupta, Richard J. LeBlanc Jr., and William F. Appelbe. The Clouds distributed operating system: Functional description, implementation details and related work. In Proceedings of the 8th International Conference on Distributed Computing Systems, pages 2--9, June 1988.

....the extension framework is not targeted for use by the ordinary user. Rather, it is oriented toward controlled specialization of the operating system by privileged expert users. Choices does try to address the problems of operating system debugging by providing a user level emulator. Clouds [10, 39] offers a complete object oriented operating system. All services in Clouds are offered as objects accessed through capabilities. The radical persistent object model makes no distinction between memory objects and file system. Essentially, all objects are located in a large distributed virtual ....

P. Dasgupta, R. J. LeBlank Jr., and W. F. Appelbe. The Clouds Distributed Operating System: Functional Description, Implementation Details and Related Work. In The 8th International Conference on Distributed Computing Systems Proceedings, pages 2--9. IEEE Computer Society Press, June 1988.

....1 Introduction Programming distributed memory parallel computers (DMPC) using a Shared Virtual Memory (SVM) seems to be in fashion. However, there is few such system available for DMPCs. Most of the research in this area has been done for a network of workstations such as IVY [18] Clouds [9, 21], Munin [6, 5] Memnet [10] Mach [27] and Chorus [1, 24] These implementations concern only high latency networks and do not allow the comparison of the efficiency of different strategies for parallelizing algorithms. Distributed memory parallel computers with a hypercube or 2D mesh topology ....

P. Dasgupta, R. LeBlanc, and W. Appelbe. The CLOUDS Distributed Operating System: Functional Description, Implementation Details and Related Work. In IEEE International Conference on Distributed Computing System, 1988.

....The classical object model does not specify a particular role for processes. There is a spectrum of relationships between objects and processes reflected in the concurrent object based systems described in the literature. An object may be a passive entity manipulated by one or more processes [Dasgupta 88] it may be one of many passive objects contained in a particular process [Strom 84] an object may be an active entity and be equivalent to a process [Detlefs 88] or several objects may reside in the same process and communicate among themselves and or with objects in other processes [Black ....

Partha Dasgupta, Richard J. Leblanc Jr. and William F. Appelbe. The Clouds Distributed Operating System: Functional Description, Implementation Details and Related Work. In 8th International Conference on Distributed Computing Systems, pages 2-9. San Jose CA, June, 1988.

....of a memory location m with a value v, and replaces m with another value v 0 , if and only if m contains v. We follow Herlihy s use of the compare swap in the implementation of our algorithm. The idea of replicating transactions in order to achieve fault tolerance is not new, see for example [NS89,Bir85,DLA88,Svo84]. We focus on the more recent work by Ng and Shi [NS89] In their algorithm they immediately create k replicas of the transaction in order to survive k Gamma 1 failures. Upon completion, the surviving replicas synchronize, selecting the replica with the highest 2 Note that, in this context, ....

Dasgupta, P., R. LeBlanc, and W. F. Appelbe. The Clouds Distributed Operating System: Functional Description, Implementation details, and Related Work. In Proceedings of the 8th International Conference on Distributed Computing Systems, pages 2--9, 1988.

....are separate entities. A process can execute in several objects during its lifetime. When a process makes an invocation on another object, its execution in the object in which it is currently executing is temporarily suspended. It continues its execution in the new object. Emerald [5] and Clouds [12] use the passive object model. 35 The active object model is suited when the system supports large grain objects, whereas the passive object model is appropriate when the system supports fine grain objects. In a system that should support objects of all granularities, we need a hybrid object ....

P. Dasgupta, R. LeBlanc, and W. Appelbe, "The Clouds Distributed Operating System - Functional Description, Implementation Details, and Related Work," in IEEE 8th International Conference on Distributed Computing Systems, (San Jose), 1989.

....networking environment or when the application evolves. We are aware of three approaches, proposed by other authors, to adapt the consistency contract to application specifics. First, application programmers might choose among a range of consistency management protocols supported by the system [4, 6, 7]. Second, programmers can use basic building blocks [5] to implement easily a suitable consistency management protocol. Third, programmers can implement the best consistency contract, relying on structuring models making the consistency management protocol a first class object [11, 18] These ....

Partha Dasgupta, Richard J. Leblanc, Jr., and William F. Appelbe. The Clouds distributed operating system: Functional description, implementation details and related work. In Proceedings of the 8th International Conference on Distributed Computing Systems, pages 2--9, S. Jos'e CA (USA), June 1988. (IEEE).

....overhead. Such complexity and performance problem explain why there are very few implementations. Some object database systems have only partial implementations [1] More complete implementations can be found in some distributed operating system projects like Argus[10] LOCUS[13] Eden[16] Clouds[2] and Camelot[3] and its commercial version Encina) The overhead of nested transactions is due to log and lock inheritance along the transaction hierarchy. Log inheritance provides the atomicity property of nested transactions, i.e. the entire transaction trees. Sub transactions are not durable ....

Partha Dasgupta, Richard J. Leblanc Jr., and William F. Appelbe. The Clouds distributed operating system: Functional description, implementation details and related work. In Proc. of International Conference on Distributed Computing System, pages 2--9, San Jose, California, U.S.A, June 1988. IEEE.

....virtual memory space. This allows applications to run on different and heterogeneous virtual address spaces. The fine granularity of the objects we are willing to support leads us to take different approach than other projects with similar computational model, but tuned to coarse grained objects [18,1]. These systems do not provide a uniform object model, because the objects they support are too large to be the building blocks of applications. Among the systems that support fine grained objects, GemStone [19] and Emerald [20] are probably the ones closer to IK. GemStone provides the extensions ....

Partha Dasgupta, Richard Leblanc, and William Appelbe. The Clouds Distributed Operating System: Functional Description, Implementation Details and Related Work. In 8th Int. Conf. on Distributed Computing Systems, pages 2--9, S. Jos'e, California, USA, June 1988. IEEE. REFERENCES 19

....overhead. Such complexity and performance problem explain why there are very few implementations. Some object database systems have only partial implementations [1] More complete implementations can be found in some distributed operating system projects like Argus[11] LOCUS[14] Eden[17] Clouds[2] and Camelot[3] and its commercial version Encina) The overhead of nested transactions is due to log and lock inheritance along the transaction hierarchy. Log inheritance provides the atomicity property of nested transactions, i.e. the entire transaction trees. Sub transactions are not durable ....

Partha Dasgupta, Richard J. Leblanc Jr., and William F. Appelbe. The Clouds distributed operating system: Functional description, implementation details and related work. In Proc. of International Conference on Distributed Computing System, pages 2--9, San Jose, California, U.S.A, June 1988. IEEE.

....the extension framework is not targeted for casual use by the ordinary user. Rather, it is oriented toward controlled specialization of the operating system by privileged expert users. Choices does try to address the problems of operating system debugging by providing a userlevel emulator. Clouds [DLA88, PD88] offers a complete object oriented operating system. All services in Clouds are offered as objects accessed through capabilities. The radical persistent object model makes no distinction between memory objects and file system. Essentially, all objects are located in a large distributed virtual ....

Partha Dasgupta, Richard J. LeBlank Jr., and William F. Appelbe. "The Clouds Distributed Operating System: Functional Description, Implementation Details and Related Work." In The 8th International Conference on Distributed Computing Systems Proceedings, pp. 2--9. IEEE Computer Society Press, June 1988.

....Since MELD transactions use an optimistic algorithm, locks are not needed by transactions. The transaction manager is built on the object manager facilities for message passing between internal transaction objects , and consists of a set of methods for these objects . 2.3.4. Clouds Clouds [Dasgupta 88] is actually an object oriented operating system, rather than a language. However, Clouds incorporates a language, as well as transaction management on its objects, so we list it here under the heading of OOPL s. In an early version of Clouds, this language was Aeolus [Wilkes 86] developed by ....

....as objects themselves. In this last case, it also uses the object manager. Table 3 1 summarizes the main features of each object system s type manager. The transaction server (Camelot [Spector 88] the byte server (ObServer II [Hornick 87] and the object oriented operating system (Clouds [Dasgupta 88] do not have type managers. The type manager in Mneme [Moss 88] would be part of its back end, and Basic Mneme has no type manager in its simple back end. The only type management in Argus [Liskov 88] is CLU s runtime type management; object type definitions are not stored persistently. ....

Partha Dasgupta, Richard J. Leblanc Jr. and William F. Appelbe. The Clouds Distributed Operating System: Functional Description, Implementation Details and Related Work. In 8th International Conference on Distributed Computing Systems, pages 2-9. San Jose CA, June, 1988.

....is implemented by combining a caching server and a replication server. Our research of object stacking has been motivated by the following three technologies: 1) Object based distributed operating systems. Most of the modern distributed operating systems are constructed as object based systems [1,2,4,5,7,12,22]. Object stacking can be used in these systems. These systems consist of clients, servers and a distributed kernel. Servers manage objects, such as files, processes, or directories, and provide services to clients. The distributed kernel provides physical I O and IPC (interprocess communication) ....

P.Dasgupta, R.J.LeBlanc,Jr. and W.F.Appelbe: "The Clouds Distributed Operating System: Functional Description, Implementation Details and Related Work", Proc. IEEE 8th Intl. Conf. on Distributed Computing Systems, pp.2-9, 1991.

....it is necessary to define a generic architecture that supports a large spectrum of existing object oriented models. Existing distributed objectoriented systems can be roughly divided in two trends: 1. Systems that do not provide a uniform object model, such as Argus [19, 20, 21] Clouds V2 [14], Eden [3, 11, 18] Hermes [12] and SOS [24, 25, 26] 2. Systems that do provide a uniform object model, such as Amber [13] Emerald [9, 10, 16] and Guide [7, 8, 17] Systems with a non uniform object model are typically designed for specialized distributed applications with, for example, strong ....

Partha Dasgupta, Richard J. Leblanc, Jr., and William F. Appelbe. The Clouds distributed operating systems: Functional description, implementation details and related work. In Proc. 8th Int. Conf. on Distributed Computing Systems, pages 2--9, S. Jos'e CA (USA), June 1988. (IEEE).

....on machines with different data representations, the programmer is responsible for ensuring that data is properly translated before being written to or read from a shared memory location. 2. 3 Object based Systems A number of object based distributed systems have been developed in recent years [4, 10, 12]. Typically, such systems encapsulate data and operations in abstract objects. These objects can then be migrated and or replicated among various machines within an interconnection. The object paradigm can be used to avoid some of the thrashing and false sharing problems of DSM. Because objects ....

Dasgupta, P., et. al., "The Clouds Distributed Operating System: Functional Description, Implementation Details and Related Work", Proceedings 8th International Conference on Distributed Computing Systems, June, 1988, pp. 2-9.

.... to research efforts resulting in Distributed Shared Memory (DSM) libraries layered on top of the networking and virtual memory systems [6,7,14,19,23] They have also resulted in the development of distributed object oriented systems offering user programs network wide access to shared services [12,18,32] as well as object oriented concurrent programming layers on top of shared and distributed memory machines [10,21,33,36] Our contribution to the emerging field of heterogeneous parallel programming is the development of the Indigo communication library. Using the small set of calls offered by ....

Partha Dasgupta, Richard J. LeBlanc, and William F. Appelbe. The Clouds distributed operating system: A functional description, related work and implementation details. In Proceedings of the Eighth International Conference on Distributed Computing Systems, June 1988.

....under the direction of the Endicott Programming Laboratory. an overview of the ARCADE distributed environment and the features which distinguish it from related work. 2. 1 Clouds Clouds is an object oriented distributed operating system developed at the Georgia Institute of Technology [9] [18] A primary goal of the Clouds project is to provide a highly reliable distributed computing facility for a set of general purpose computers connected by a local area network. It is based on the object thread paradigm. In Clouds, objects are the primary abstraction for data storage, while ....

....vary widely from one system to another. A very rough comparison, however, shows that ARCADE s performance figures are in the same approximate range as those for the Argus and Clouds distributed systems. For those systems, typical remote operations require on the order of 40 to 300 milliseconds [9] [16] Figure 5. Performance of Remote Transfer Operations 4.8 Operating System Services The ARCADE architecture is only a platform for an operating system. It does not include normal operating system services, such as a user interface or a file system. In order to evaluate ARCADE as a ....

Dasgupta, P., Leblanc, R. and Appelbe, W. "The Clouds Distributed Operating System: Functional Description, Implementation Details and Related Work", in Proceedings of the 8th International Conference on Distributed Computing Systems, IEEE, San Jose, CA, June, 1988, pp. 2-9.

.... Since differences in remote to local access times are even more pronounced in distributed machines like sets of workstations connected via high speed networks [55] notions of distributed objects have been a topic of research for such systems for quite some time, as evidenced by work on Clouds [69] at Georgia Tech, on Chorus [197] and on fragmented objects [227] in France. 3.3 Synchronization When multiple cooperating processes execute simultaneously, synchronization primitives are needed for concurrency control. When multiple processes share an address space, synchronization is required ....

Partha Dasgupta, Richard J. LeBlanc Jr., and William F. Appelbe. The clouds distributed operating system: Functional description, implementation details and related work. In Proceedings of the 9th International Conference on Distributed Computing Systems, San Jose, CA., pages 2--9. IEEE, June 1988.

....networking environment or when the application evolves. We are aware of three approaches, proposed by other authors, to adapt the consistency contract to application specifics. First, application programmers might choose among a range of consistency management protocols supported by the system [4, 6, 7]. Second, programmers can use basic building blocks [5] to implement easily a suitable consistency management protocol. Third, programmers can implement the best consistency contract, relying on structuring models making the consistency management protocol a first class object [11, 19] These ....

Partha Dasgupta, Richard J. Leblanc, Jr., and William F. Appelbe. The Clouds distributed operating system: Functional description, implementation details and related work. In Proceedings of the 8th International Conference on Distributed Computing Systems, pages 2--9, S. José CA (USA), June 1988. (IEEE).

....was proposed by Redell [19] and has been implemented in IBM s system 38 [20] The main reason is to ease sharing of objects between processes, and in a parallel system this is a particularly important consideration. This has led to more recent single address space operating systems such and Clouds [21] and Psyche [22] In Angel we extend the consistent address space to span multiple PEs in a distributed memory system, maintaining coherence where necessary using DSM techniques. 1 We assume the emergence of 64 bit processors to make this viewpoint feasible. There are several reasons to expect ....

P. Dasgupta, R. LeBlanc, and W. Appelbe, "The Clouds distributed operating system: functional description, implementation details and related works," in International Conference on Distributed Computing Systems, 1988.

....latter decision provides us with a rich set of ready made tools for perusal and management of long lived segments. 3.6.3. Software Implementation of Large Address Spaces Many distributed systems provide large name spaces with software interpretation. Some provide names for heavyweight objects [1, 16]; others for communication ports [14, 42] still others for mappable segments [33] In any case, objects in the distributed name space cannot be accessed via hardware addressing modes (except perhaps with a temporary mapping) and must generally be treated differently from addressable local ....

P. Dasgupta, R. J. LeBlanc, Jr. and W. F. Appelbe, "The Clouds Distributed Operating System: Functional Description, Implementation Details and Related Work," Proceedings of the Eighth International Conference on Distributed Computing Systems, 13-17 June 1988, pp. 2-9.

....different approaches taken towards customization in fault tolerant computing. A large number of other projects are not addressed in detail here, including ADS [IM84] Amoeba [RST89, KT91] AMp and xAMp [VRB89, RV91, RV92] Argus [Lis85, Lis88] Avalon [DHW88] Chorus [BFG 85] Clouds [DLA88, DLAR91] Delta 4 [PSB 88, BHV 90, Pow91] Saturne [DFF 90] Totem [AMMS 93, AMMS 95, MMSA 96] Transis and Lansis [ADKM92b, ADKM92a] and the V kernel [CZ85] 2.1.1 Static Systems 2.1.1.1 Isis The Isis system is a toolkit developed to support construction of distributed ....

P. Dasgupta, R. LeBlanc, and W. Appelbe. The Clouds distributed operating system: Functional description, implementation details and related work. In Proceedings of the 8th International Conference on Distributed Computing Systems, pages 2--9, Jun 1988.

No context found.

P. Dasgupta, R. J. LeBlanc Jr. and W. F. Appelbe, `The Clouds distributed operating system: functional description, implementation details and related work', 11th IEEE Conf. on Distributed Computing Systems, 1988, pp. 2--9.

No context found.

Dasgupta, P., et. al., "The Clouds Distributed Operating System: Functional Description, Implementation Details and Related Work", Proceedings of the 8th International Conference on Distributed Computing Systems, June, 1988, pp. 2-9.

First 50 documents

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