A. Bernstein. Output Guards and Nondeterminism in "Communicating Sequential Processes". ACM Transactions on Programming Languages and Systems, 2(2):234--238, Apr. 1980.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....a strictly deterministic implementation. Use of such a deterministic matching scheme allows each process to select the same guard to execute, independently. The mechanism defined here differs from that of CSP in two regards. First, output commands may be placed in guards. Hoare [8] and Bernstein [3] discuss the desirability of this. To simplify implementation , such a facility was not originally included in CSP. Second, an otherwise guard h as been added, which is similar to a feature in the Ada SELECT statement. Lastly, note that distributed deadlock detection is easily accomplished. ....

....mutual exclusion in a distributed system is proved. We happily note that their solution can be viewed as an optimization of a distributed semaphore based solution to the critical section problem. Other implementations of the nondeterministic message passing mechanism in CSP are described in [3, 21, 22]. Each uses a different .mechanism to allow processes to compute independently an ordering on the triples in our Com set. 7. CONCLUSIONS To date, numerous language proposals have appeared that include message passing facilities with which process synchronization can be implemented. We have ....

BERNSTEIN, A.J. Output guards and nondeterminism in "Communicating Sequential Processes." ACM Trans. Program. Lang. Syst. 2, 2 (Apr. 1980), 234-238.

.... mixed choice for channel based synchronous communication, as exemplified with Concurrent ML [Rep91] and Facile [TLK96] Despite Palamidessi s impossibility result, there also exist algorithms for the distributed implementation of such languages, and in particular of mixed choice, e.g. by Bernstein [Ber80], Buckley and Silberschatz [BS83] and Knabe [Kna93] which all have been proven although rather informally to be correct or, at least, to be deadlockfree. The question arises how these practically satisfactory implementations relate to Palamidessi s impossibility result. It is worthwhile to ....

....divergence is not harmful anymore, as long as there is no danger for live locks. Another method, known from the distributed implementation of concurrent languages, is exploiting a total order among the threads in the system by, for example, always choosing the lock of the smaller thread first [Ber80, BS83, Kna93], when required to make a choice. Then, the above symmetric cyclic wait situation is immediately prevented since both receivers choose the same thread, i.e. lock, as the first to interrogate. Note also that under a total order assumption symmetric networks according to [Pal97] do not exist. In ....

[Article contains additional citation context not shown here]

A. Bernstein. Output Guards and Nondeterminism in "Communicating Sequential Processes". ACM Transactions on Programming Languages and Systems, 2(2):234--238, Apr. 1980.

.... mixed choice for channel based synchronous communication, as exemplified with Concurrent ML [Rep91] and Facile [TLK96] Despite Palamidessi s impossibility result, there also exist algorithms for the distributed implementation of such languages, and in particular of mixed choice, e.g. by Bernstein [Ber80], Buckley and Silberschatz [BS83] and Knabe [Kna93] which all have been proven although rather informally to be correct or, at least, to be deadlock free. The question arises how these practically satisfactory implementations relate to Palamidessi s impossibility result. It is worthwhile ....

....divergence is not harmful anymore, as long as there is no danger for live locks. Another method, known from the distributed implementation of concurrent languages, is exploiting a total order among the threads in the system by, for example, always choosing the lock of the smaller thread first [Ber80, BS83, Kna93], when required to make a choice. Then, the above symmetric cyclic wait situation is immediately prevented since both receivers choose the same thread, i.e. lock, as the first to interrogate. Note also that under a total order assumption symmetric networks according to [Pal97] do not exist. In ....

[Article contains additional citation context not shown here]

A. Bernstein. Output Guards and Nondeterminism in "Communicating Sequential Processes". ACM Transactions on Programming Languages and Systems, 2(2):234--238, Apr. 1980.

....the programmer manipulates can be just the internal name of this thread. In an actual implementation it may be useful to have just one thread of control manage multiple channels. Our algorithm also relies on the existence of a global ordering for processes, an idea which goes back to Bernstein [4]. Each process has a unique id, and any two ids may be compared against one another. This is not an unreasonable assumption, for ids are necessary to route messages unambiguously. We now present the structure of the protocol. We show first the actions associated with the channels, and then turn ....

A. Bernstein. Output guards and nondeterminism in "communicating sequential processes". ACM Transactions on Programming Languages and Systems, 2(2):234--238, Apr. 1980.

....cases where the construct is easy to implement if only certain processes use it. Inefficient implementations of the general construct include those that use global information (e.g. a central coordinator) 18] require an unbounded amount of time [10] or use an unbounded amount of communication [2, 19]. An overview of these techniques can be found in [4] Buckley and Silberschatz [4] were the first to propose a protocol that avoids all three of the inefficiencies above. Bagrodia [1] later proposed a protocol that uses fewer messages. Unfortunately, both protocols are prone to deadlock for ....

A. J. Bernstein. Output guards and nondeterminism in "Communicating sequential processes". ACM Trans. Programming Languages and Systems, 2(2):234-- 238, Apr. 1980.

....cases where the construct is easy to implement if only certain processes use it. Inefficient implementations of the general construct include those that use global information (e.g. a central coordinator) 19] require an unbounded amount of time [9] or use an unbounded amount of communication [2, 20]. An overview of these techniques can be found in [4] Buckley and Silberschatz [4] were the first to propose a protocol that avoids all three of the inefficiencies above. Bagrodia [1] later proposed a protocol that uses Negotiating Non Deterministic Choice 3 fewer messages. Unfortunately, both ....

Arthur J. Bernstein. Output guards and nondeterminism in "Communicating sequential processes". ACM Transactions on Programming Languages and Systems, 2(2):234--238, April 1980.

No context found.

A. Bernstein. Output Guards and Nondeterminism in "Communicating Sequential Processes". ACM Transactions on Programming Languages and Systems, 2(2):234--238, Apr. 1980.

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