N. Carriero and D. Gelernter. The S/Net's Linda Kernel. ACM Transactions on Computer Systems, 4(2):110--129, 1986.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....discuss why they are inappropriate for use in a general communications network. We will also discuss two other replication approaches that can be alternatives to our scheme: the voting scheme and the viewstamped replication scheme. 6.1. 1 S Net Kernel The S Net kernel is described in detail in [8]. The S Net consists of several MC 68000 s with local memory, connected by a bus. The operations are executed as follows: executing out(t) causes tuple t to be broadcast to every node in the network; thus every node stores a complete copy of the tuple space. Executing in(s) triggers a local ....

Nicholas Carriero and David Gelernter. The S/Net's Linda Kernel. ACM Transaction on Computer Systems, 4(2):111-129, May 1986.

....and also more complicated, including tags, attributes, and name spaces. Furthermore, interfaces to access XML based data, such as DOM [30] are relatively complex. We chose an I O mechanism that distinguishes between storage and communications instead of a unified tuple space abstraction [9, 21, 49] because such a mechanism better reflects how applications store and communicate data. In particular, applications often modify stored data. For example, a personal information manager needs to be able to update stored contacts and appointments. Structured I O lets applications overwrite stored ....

....at the pervasive computing space. Leases have been used to control write access to cached files [22] access to remote resources in general [3] and even the lifetime of stored objects [21] Starting with Linda, tuple spaces have been used to enable coordination between loosely coupled services [9, 21, 49]. The Information Bus provides similar functionality based on a publish subscribe paradigm [34] While it nominally uses objects, exchanged information is represented by selfdescribing data objects, which are similar to tuples in one.world . Asynchronous events have been used across a wide ....

N. Carriero and D. Gelernter. The S/Net's Linda kernel. ACM Transactions on Computer Systems, 4(2):110--129, May 1986.

....Graph Structure The solution described so far is based on a centralised graph structure; this was because it was not clear whether or not a distributed graph would bring benefit to the 4.4. Distributed Graph Structure 110 system. In LIGIA, which uses the distributed central servers strategy 4 [CG86] the advantages of the distribution of the graph are clear; a distributed version of the solution described previously will avoid possible bottlenecks when several processes are modifying the graph at the same time. Yet, in some LINDA like systems processes can migrate which makes the ....

....architecture depicted in Figure 6.1 is divided into four main distinct systems: server system, client system, I O sub system and garbage collector sub system. The communication among servers occurs via sockets; the servers use distributed central servers as their tuple space distribution strategy [CG86] Each server is multithreaded a new thread is created for each new request done by a client. The client system design is composed of a multi threaded communication system plus a local tuple space manager [Row97] The latter was not included in the implementation since it would complicate ....

[Article contains additional citation context not shown here]

N. Carriero and D. Gelernter. The S/Net's Linda Kernel. ACM Transactions on Computer Systems, 4(2):110--129, 1986.

....decide what is garbage is maintained within the server in a form of a hybrid graph [8] and this paper utilises this proposal. 3 Implementation Overview 3.1 Implementation Architecture Ligia architecture is simple. The implementation uses distributed central servers as its distribution strategy [1]. Each server is multi threaded. A new thread is created for each new request done by an agent. From the point of view of the architecture, agents can be either single or multi threaded as for the servers this makes no di erence. In the implementation used in this paper, agents are ....

N. Carriero and D. Gelernter. The S/Net's Linda Kernel. ACM Transactions on Computer Systems, 4(2):110{ 129, 1986.

....a new thread is created for each new request done by an agent. In this paper an implementation with one server only is used to prove the concept. However, multiple servers can be used but where tuple spaces are stored uniquely in one of the servers a central server distribution strategy [1]. The primitives are implemented as methods of a class called ConnectionRequest. When an object of this class is created the process registration occurs with the server. When the registration is accepted, a unique name is assigned to the connection and this name is used by this object from there ....

N. Carriero and D. Gelernter. The S/Net's Linda Kernel. ACM Transactions on Computer Systems, 4(2):110-129, 1986.

....If a process tries to access a tuple space which is not in the location de ned in the handle, the Router (see below) will redirect the operation to the proper server and return an updated version of the handle. Ligia architecture uses distributed central servers as its distribution strategy [2]. Each server is multi threaded. A new thread is created for each new request done by an agent. As depicted in Figure 2, the architecture is divided into ve main components (for full description see [7] Agent Connection: This component is responsible for dealing with all communication arriving ....

N. Carriero and D. Gelernter. The S/Net's Linda Kernel. ACM Transactions on Computer Systems, 4(2):110-129, 1986.

....way as it is done for the entries. These hash codes are then used to find the shortest list in the hashCode( hashed structure. Finally the shortest list is searched for an entry matching the template. anEntry field[0] hashCode( 34 field[1] hashCode( 12 field[2] hashCode( 1 field[3].hashCode( 4321 ID = 1234 type = EntryType ID hashed size = 8 hashCode( hashed size = 8 Type lists size = 4 Same or subclass bitmap EntryTypeA EntryTypeB EntryTypeC EntryType EntryTypeA EntryTypeB EntryTypeC EntryType EntryType EntryTypeC EntryTypeB EntryTypeA add type ....

....machine. Preliminary experiences and results, both in terms of performance and in terms of provided functionality, are very encouraging. Critics of Linda like system, that efficient matching is difficult [6] were not verified. Some Linda implementation hashed on the first tuple value only [6] [3] and had therefore poor matching performance in cases where the first value was the same for all tuples. Hashing on all fields, as used in this implementation, introduces some overhead at entry insertion and removal, but provides near constant time matching. Using the Harness plug in interface ....

N. Carriero and D. Gelernter. The S/Net's Linda Kernel. ACM Transactions on Computer Systems, 4(2):110--129, May 1986.

....operations to manipulate the data structures. Only these operations can be carried out upon the encapsulated data objects. The parallel processes are created through the explicit creation of sequential processes that execute in parallel with one 28 May 1997 14 another [Bal et al. 92] Linda [Carriero et al. 86] is a language which consists of four simple operators which are added to a host language to turn it into a parallel programming language. Linda processes communicate through a globally shared collection of ordered tuples these tuples are exist in a global abstract tuple space. Despite ....

....through a globally shared collection of ordered tuples these tuples are exist in a global abstract tuple space. Despite Tanenbaum s apparent enthusiasm for the object based model, the authors of Linda say the object based model is powerful and attractive, but utterly irrelevant to parallelism [Carriero et al. 86] Tanenbaum 95] 4 Conclusion Since one of the main goals of Distributed Systems is transparency, the achievement of this when DSM is implemented is only possible if the use the shared memory is completely invisible. Hellwagner [Hellwagner 1990] says that the overall objective in DSM is to ....

Nicholas Carriero, David Gelernter, The S/Net's Linda Kernel., ACM Transactions on Computer Systems, Vol. 4, No. 2, May 1986, Pages 110-129.

....the granularity of the shared memory. However, these implementations require additional input from the programmer regarding variable usage (Munin [Bennett et al., 90] Midway [Bershad et al., 93] or require the use of a specialized language or language extensions (Orca [Bal et al., 92] Linda [Carriero et al., 86] The latter implementations would appear to defeat the purpose of DSM since they require the programmer to learn a new language or use unfamiliar constructs. In addition, these implementations prevent the direct use of existing software. Much research has also been done in an attempt to ....

.... moved around the distributed system can be one of the following: fixed sized blocks of data (IVY [Li et al., 89] Treadmarks [Keleher, 94] variable sized blocks of data (Midway [Bershad et al., 93] Munin [Bennett et al., 90] or user defined structures (Orca [Bal et al., 92] Linda [Carriero et al., 86] As mentioned earlier, one of the advantages of DSM is that it makes programming on a distributed system easier and allows programmers to use a familiar programming model, the shared memory model. The notion of a variable being a distinct section of memory may be intuitive to the programmer ....

Nicholas Carriero, David Gelernter, The S/Net's Linda Kernel., ACM Transactions on Computer Systems, Vol. 4, No. 2, May 1986, Pages 110-129.

....processes to machines with low load. This is called static scheduling, because once the system makes the decision of the set of machines to receive processes, the processes must run to completion on those machines. In addition, parallel programming environments such as MPI [7] PVM [8] and Linda [9] made it easier for programmers to take advantage of transparent system clusters to solve problems in parallel. Each of these systems consisted of two components. The rst component is a programming language or language library, used by the programmer to specify a parallel computation. The second ....

....let her workstation be put to good use when it s idle, as long as she can have complete control of the machine when she is sitting in front of it. This is an example where static scheduling goes a long way, but does not quite solve the problem completely. Static scheduling systems such as Linda [9] can automatically monitor the load on all workstations, and schedule processes to idle processors when a parallel job begins. The trouble comes if Alice returns to her workstation while Bob is in the middle of running a large simulation on it. If we are using static scheduling, our options are ....

N. Carriero and D. Gelernter. The S/Net's Linda Kernel. ACM Transactions on Computer Systems, 4(2):110-129, May 1986.

....works by using a binary rewriter to inject its 3 operations into application binaries and by extending relevant operating system services across the cluster. The resulting system is exible enough to run an Oracle database. The tuplespace concept has rst been introduced in the context of Linda [10] to support parallel programming, but has subsequently been adapted to manage shared state in distributed systems. A tuplespace essentially is a collection of typed records (or objects) The three basic operations are write, which adds a tuple to the tuplespace, read, which returns a tuple ....

Nicholas Carriero and David Gelernter. The S/Net's Linda kernel. ACM Transactions on Computer Systems, 4(2):110-129, May 1986.

....providers and requesters in face of the decoupled nature of open distributed service markets. In addition to the autonomy of providers and requesters, decoupling refers to the fact, that they are located in different address spaces and that they possibly act in disjoint time intervals. In Linda [CAGE86] this is called space and time decoupling. It is important to note that services may be shared among different users and that the relationship between an imported type and an exported type may be a polymorphic one. The former supports resource sharing and the cooperation via shared objects, the ....

N. Carriero and D. Gelernter. The S/Net's Linda Kernel. ACM Transactions on Computer Systems, 4(2):110--129, May 1986.

....the list of tuple space nodes to provide a wider distribution of tuples. The addition of the new tuple to each tuple space causes any blocked threads to be awakened so that they can try to match against the new tuple. 6 Discussion Distributed Linda can be implemented using either replication [2, 6, 14] or hashing [4] Some designs also include the notion of dynamic migration: if a set of tuples is being consumed on a particular node, it is more efficient if they are stored directly on that node. The hash based approach implements tuple space as a distributed hash table in which one or more ....

Nicholas Carriero and David Gelernter. The S/Net's Linda kernel. ACM Transactions on Computer Systems, 4(2):110--129, May 1986.

....[6] For example, Mach uses RPC for all of its operating system services, and Camelot and Avalon use RPC in transaction based reliable distributed systems. The second mechanism is Linda tuple space, a form of high level shared memory that can be provided in both distributed and parallel systems [8]. Remote Procedure Call The client server model is an extremely successful method of structuring distributed systems and fits naturally into ML. A server defines its exported interface via a signature, so that the unit of distribution is the ML module. Constructing distributed systems becomes no ....

Nicholas Carriero and David Gelernter. The S/Net's Linda kernel. ACM Transactions on Computer Systems, 4(2):110--129, May 1986.

....decision. Known architectures for distributed implementation of Linda like systems include full replication of a repository to all nodes, no replication with a single, centralized repository, no replication of repositories in a ring like organization ( Pol95] or a partial replication as in [CG86] As long as there is one defined node that follows a gateway protocol to other sub PageSpaces, our architecture supports all of them. In fact, the current Jada implementation uses a centralized or fully replicated repository, whereas Laura implements a partial replication scheme. For a gateway, ....

Nicholas Carriero and David Gelernter. The S/Net's Linda Kernel. ACM Transactions on Computer Systems, 4(2):110--129, 1986. A-17

....in which information is retrieved by content rather than by address. This property, together with the temporal and spatial decoupling inherent in the abstraction, makes it especially easy for use by application programmers. Linda implementations are available on a number of different architectures [12, 9, 14] and for a number of different languages [27, 21, 19] Linda has been used for many applications, including VLSI design, oil exploration, pharmaceutical research, fluid flow systems design, and financial analysis [10] Despite these advantages, one significant deficiency of Linda as originally ....

Nicholas Carriero and David Gelernter. The S/Net's Linda kernel. ACM Transactions on Computer Systems, 4(2):110--129, May 1986.

....held and operations on them suffice for the implementation. For a distributed implementation, the communication costs for replication and distributed search have to be balanced for efficiency. We chose a replication scheme similar to the one originally proposed for the S NET Linda implementation ([4]) The idea is to replicate elements only partially on subsets of all nodes. Within this subset, the search for an element can be performed locally. In contrast to a full replication, the costs for removing a replica decrease. In order to find an element from all subsets, only one node per ....

Nicholas Carriero and David Gelernter. The S/Net's Linda Kernel. ACM Transactions on Computer Systems, 4(2):110--129, 1986.

No context found.

N. Carriero and D. Gelernter. The S/Net's Linda Kernel. ACM Transactions on Computer Systems, 4(2):110--129, 1986.

No context found.

N. Carriero, D. Gelernter, The S/Net#s Linda Kernel, ACM Transactions on Computer Systems 4 (2) (1986) 110-- 129.

No context found.

Nicholas Carriero and David Gelernter. The S/Net's Linda Kernel. ACM Transactions on Computer Systems, 4(2):110--129, 1986.

No context found.

N. Carriero and D. Gelernter. The S/Net's Linda Kernel. ACM Transactions on Computer Systems, 4(2):110--129, 1986.

No context found.

Nicholas Carriero and David Gelernter. The S/Net's Linda Kernel. ACM Transactions on Computer Systems, 4(2):110--129, 1986.

No context found.

Nicholas Carriero and David Gelernter. The S/Net's linda kernel. ACM Transactions on Computer Systems, 4(2):110--129, 1986.

No context found.

Nicholas Carriero and David Gelernter. The S/Net's Linda Kernel. ACM Transactions on Computer Systems, 4(2):110--129, 1986.

No context found.

N. Carriero and D. Gelernter, `The S/Net's Linda kernel', ACM Transactions on Computer Systems, 4, (2), 110--129 (1986).

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