A. M. Ricciardi and K. P. Birman. Using Process Groups to ImplementFailure Detection in Asynchronous Environments. In 11th ACM Symposium on Principles of Distributed Computing, pages 341--353, 1991.  Home/Search   Document Not in Database   Summary   Related Articles   Check

.... the same group or capriciously install new views excluding correct and non suspected processors) and yet it should be weak enough to remain solvable [58] These problems have been identified in two papers widely referenced to give rigorous definitions of group membership for asynchronous systems: [59] for the primary component type, and [60] for the partitionable one. Since the work of [58] several other group membership specifications have appeared. De21 spite this intense activity, the distributed system community has yet to agree on a formal definition of the group membership problem, ....

A. Ricciardi and K. P. Birman, "Using process groups to implement failure detection in asynchronous environments," in Proceedings of the Tenth ACM Symposium on Principles of Distributed Computing, pp. 341--352, 1991.

....Section 6. 2 Related Work The membership problem was first defined for synchronous systems in [4] Since then it has also been the subject of intense investigation for asynchronous systems. There are two types of membership services for asynchronous systems: primary partition membership services [21, 17, 19, 18, 6, 15] and partitionable membership services [1, 16, 20, 7, 6] We use the timed asynchronous system model as the foundation of our work [5] This system model assumes that processes have access to a local unsynchronized hardware clock with a bounded drift rate. It uses the following failure model: ....

....model does not define any bound on the failure frequency. It is not only an accurate description of a network of workstations running Unix or Windows NT, it also describes many nearly synchronous systems accurately. Some of these membership services explicitly assume the time free system model [21, 19, 15], some assume the timed system model [16, 6] and for others it is not entirely clear if the underlying model is timed or time free. It was shown in [3] that the membership problem is not solvable in time free asynchronous systems. This impossibility result does not hold for timed systems [6] ....

A. Ricciardi and K. Birman. Using process groups to implement failure detection in asynchronous environments. In Eleventh ACM Symposium on Principles of Distributed Computing, Montreal, Canada, Aug 1991.

....member of a group in a site, only one message is sent there, and then copied to all recipients. Likewise, when a site fails, a single run of a site failure detection algorithm needs to be executed, instead of having many runs executing in parallel as in some other earlier generation groups systems[33, 26, 7]. On the other hand, site group members process the send requests from their local senders, and run the protocols that ensure delivery to the members, with the requested guarantees. Senders are not full right members of a group. They get from the system just the necessary information and support ....

....activity. Site membership concerns site groups, which are a mapping of groups on the site structure, and its aim is to assist intersite communication on behalf of group participants. Both objectives have traditionally been aggregated to the Group Membership Problem in the generality of other works [13, 23, 33]. The main data structures of a ##, as shown in Figure 6, are the Group Membership, and the Activity Criteria. The Group Membership structure holds the membership and views of all groups. The Activity Criteria are defined on a per group basis, when the group is created. They establish the rules ....

Aleta M. Ricciardi and Kenneth P. Birman. Using process groups to implement failure detection in asynchronous environemnts. In Proceedings of the 10th ACM Symposium on Principles of Distributed Computing, pages 341--351, August 1991.

....in order to model network partitions. In the paper we only consider the primary partition membership service. 2.2 Solving Group Membership Many algorithms have been proposed to solve the group membership problem. These algorithms have in common to be complex. This is the case of the protocol in [ 24], but there are two more recent examples. In [17] Lotem et al. describe a membership protocol that requires the introduction of notions such as quorums, sub quorums, ambiguous sessions, last formed sessions, resolution rules, learning rules. In [ 7] and in its recent version [20] an Atomic ....

A.M. Ricciardi and K. P. Birman. Using Process Groups to Implement Failure Detection in Asynchronous Environments. In Proc. of the 10th ACM Symposium on Principles of Distributed Computing, pages 341--352, August 1991.

....denoted here, satisfies only the following two properties: 1. There is a time after which every process that crashes is always suspected by some correct process. 2. There is a time after which some correct process is never suspected by any correct process. A di#erent approach was taken in [RB91] a correct process that is wrongly suspected to have crashed, voluntarily leaves the system. It may later rejoin the system by assuming a new identity. In [CT91] this is denoted 3W . Note that, at any given time t, processes cannot use to determine the identity of a correct process. ....

....easy to see that the root of some # , k i, that appears in that sequence is 1 valent, while the root of # k 1 is 0 valent. By definition, k is a critical index. 2 The rest of the proof remains unchanged. 7. 4 Failure detectors with infinite range of output values The failure detectors in [RB91, CT91] only output lists of processes suspected to have crashed. Since the set of processes is finite, the range of possible output values of these failure detectors is also finite. In this paper our model allows for failure detectors with arbitrary ranges of output values, including the ....

Aleta Ricciardi and Kenneth P. Birman. Using process groups to implement failure detection in asynchronous environments. In Proceedings of the Tenth ACM Symposium on Principles of Distributed Computing, pages 341--351. ACM Press, August 1991.

.... of the inherent difficulty of formally specifying membership services, it has been shown in [Anceaume et al. 1995] that the two most referenced formal specifications of membership services, both for the primary component and for the partitionable type, are unsatisfactory: the primary component one [Ricciardi Birman 1991, Ricciardi Birman 1993] has flaws in the formalism and allows undesirable executions, while the partitionable one [Dolev et al. 1994, Dolev et al. 1995] can be satisfied by useless protocols. 17 18 19 A precise definition of partition in time free asynchronous systems can be found in ....

A. Ricciardi and K. P. Birman, "Using Process Groups to Implement Failure Detection in Asynchronous Environments", in loth ACM Syrup. on Principles of Distributed Computing, (Montreal, Canada), pp.341-51, 1991.

....mg is then discarded by all correct replicas in the group. Handling Partitions As has been discussed in section 3.1, GCP supports a paritionable membership model, in order to achieve non blocking membership protocols in asynchronous systems. However, we follow the Primary Partition approach [20] in the design of the RSP layer. In case of system partitioning, the RSP layers of different replicas deliver partitioned views of the group. The replicas in at most one of the partitions continue providing the service. For example, this may be the majority partition (if any) or the partition ....

A. Ricciardi and K. Birman, "Using Process Groups to Implement Failure Detection in Asynchronous Environments," presented at 10th Annual ACM Symposium on Principles of Distributed Computing, 1991.

....services may use the core mechanisms, but need to be careful not to create infinitely recursive dependencies. Other than the core services, they can make use of local operating system facilities such as disk storage. The first external service, the Failure Detector (FD) performs fault analysis [8]. FD is a distributed service that uses information passed up by MUTS, in addition to information made available by the host operating system, to determine the sites that have become unreachable. As its output, it provides a stream of failure notifications, which is multicast to all relevant ....

A. Ricciardi and K. Birman, "Using Process Groups to Implement Failure Detection in Asynchronous Environments," ACM Symposium on Principles of Distributed Computing, pp. 341-353, Montreal, Quebec, Canada, August 19-21, 1991. - 5 - -

....network [2] Different servers may dedicate different amounts of storage to Tangler, but each publicly certifies its capacity. There is a general consensus on the public keys and capacities of available servers. The list of servers is maintained using a standard group membership algorithm (e.g. [16]) Convincing other nodes of a server s corruption is straight forward. Many forms of corruption result in two contradictory messages digitally signed by the same server a succinct proof of the server s misbehavior. The other main threat is that a bad server will refuse to answer requests ....

A. M. Ricciardi and K. P. Birman. Using process groups to implement failure detection in asynchronous environments. In Proceedings of the 10th Annual ACM Symposium on Principles of Distributed Computing, pages 341--353, August 1991.

....ours. Device sharing clusters, including Tandem Non Stop Computers, VAXclusters [85, 145] and Microsoft Cluster Service [157] use a six round membership protocol to coordinate resource failover during configuration changes. Atomic broadcast protocols also use membership protocols similar to ours [4, 49, 131]. The theoretical study of membership protocols has begun only recently. Chandra and Toueg proved the impossibility of membership agreement in asynchronous networks [31] Cristian and Schmuck defined the timed asynchronous distributed system model, in which messaging delay is bounded in the ....

A. M. Riccardi and K. P. Birman. Using process group to implement failure detection in asynchronous environments. In 10th ACM Symp. on Princ. of Distr. Computing (PODC), pages 341--352, Montreal, Quebec, August 1991. 3.5.1 147

....delays, but if the bounds used are not correct then we are in a situation similar to the one where the impossibility result of [7] holds. In order to solve this problem we use an approach similar to the one used by some research groups to solve the asynchronous problem a failure detector [13, 2]. However, there is a major difference between these two scenarios we want to reach agreement in a known bounded time. For that, we need to have (real) synchronous channels to implement the failure detector. Our claim is that it is reasonable to make that assumption on most real time ....

....levels of synchronism in [5] Based on these analysis and studies, we have to come out with a conclusion about what type of QoSs can be provided or not in a given system, and whether we need additional mechanisms in order to provide some QoSs. This idea is similar to what was done in Isis [13] to overcome the impossibility result of [7] As the system is supposed to run in asynchronous environments and be able to provide consensus among group members, they had to find a mechanism to solve the problem in a practical case. That is the function of the failure detector [13] With the ....

[Article contains additional citation context not shown here]

Aleta M. Ricciardi and Kenneth P. Birman. Using process groups to implement failure detection in asynchronous environemnts. Technical Report TR 91-1188, Cornell University, Department of Computer Science, 1991. 22

....delays, but if the bounds used are not correct then we are in a situation similar to the one where the impossibility result of [6] holds. In order to solve this problem we use an approach similar to the one used by some research groups to solve the asynchronous problem a failure detector [12, 1]. However, there is a major difference between these two scenarios we want to reach agreement in a known bounded time. For that, we need to have (real) synchronous channels to implement the failure detector. Our claim is that it is reasonable to make that assumption on most real time ....

Aleta M. Ricciardi and Kenneth P. Birman. Using process groups to implement failure detection in asynchronous environemnts. Technical Report TR 911188, Cornell University, Department of Computer Science, 1991.

....some extensions in order to be able to provide early delivery to all participants of the communication group. The idea of using a failure detection service together with the communication protocols, that we address in the present paper 2 , can also be related to work done by the Isis group [10, 3] to solve the problem of consensus in an asynchronous system [7] The major difference between the two scenarios is that we want to reach agreement in a known bounded time. This puts some special requirements on the failure detection service it must be implemented using a synchronous channel. ....

A. M. Ricciardi and K. P. Birman. Using process groups to implement failure detection in asynchronous environemnts. Technical Report TR 91-1188, Cornell University, Department of Computer Science, 1991.

....point. In our paper we introduced extensions in order to be able to provide earlydelivery to all participants of the communication group. The use of a failure detection service together with the communication protocols, that we address in [3] can also be related to work done by the Isis Group [24, 8] to solve the problem of consensus in an asynchronous system [17] The major difference between the two scenarios is that we want to reach agreement in a known bounded time and also detect timing failures. This puts some special requirements on the failure detection service it must be ....

A. M. Ricciardi and K. P. Birman. Using process groups to implement failure detection in asynchronous environemnts. Technical Report TR 91-1188, Cornell University, Department of Computer Science, 1991.

....one another; in this way, the usual membership rituals of group based communication systems (or worse, the hard coded component to component knowledge required for RPC or messaging systems) are handled almost transparently. While Jini makes no attempt to achieve strong membership semantics (e.g. [2, 21]) because its Lookup Service (described below) exists only for objects to locate one another, the discovery, join, and detach mechanisms are parametrized so that the aggregate membership can approximate, to as great a degree as desired, the actual state of the system s connectivity. In other ....

A. Ricciardi and K. Birman. Using Process Groups to Implement Failure Detection in Asynchronous Environments. In Tenth PODC, pages 341-351. ACM, 1991.

No context found.

A. M. Ricciardi and K. P. Birman. Using Process Groups to ImplementFailure Detection in Asynchronous Environments. In 11th ACM Symposium on Principles of Distributed Computing, pages 341--353, 1991.

No context found.

41 A. M. Ricciardi and K. P. Birman. Using Process Groups to Implement Failure Detection in Asynchronous Environments. In Proc. of the 10th ACM Symposium on Principles of Distributed Computing, pages 341--352, August 1991.

No context found.

A. Ricciardi and K.P. Birman. Using process groups to implement failure detection in asynchronous environments. In Proceedings of the 10 Principle of Distributed Computing (PODC), pages 341-352, 1991.

No context found.

A. Ricciardi and K. Birman. Using process groups to implement failure detection in asynchronous environments. In Proc. annual ACM Symposium on Principles of Distributed Computing, Aug. 1991.

No context found.

A. M. Ricciardi and K. P. Birman. Using Process Groups to Implement Failure Detection in Asynchronous Environments. In Proc. of the 10th ACM Symposium on Principles of Distributed Computing, pages 341--352, August 1991.

No context found.

A. Ricciardi and K. Birman, "Using Process Groups to Implement Failure Detection in Asynchronous Environments", Proc. ACM Symp. on Principles of Distributed Computing, Aug. 1991, pp. 341--352.

No context found.

A. M. Ricciardi and K. P. Birman. Using Process Groups to Implement Failure Detection in Asynchronous Environments. In ACM Symp. on Prin. of Distributed Computing (PODC), pages 341--352, August 1991.

No context found.

A. M. Ricciardi, K. P. Birman, "Using Process Groups to Implement Failure Detection in Asynchronous Environments", Annula ACM Symposium of Principles of Distributed Computing, Montreal, Aug. 1991.

No context found.

A. Ricciardi. Using process groups to implement failure detection in asynchronous environments. In Proceedings of the Tenth Annual ACM Symposium on Principles of Distributed Computing, Montreal, P.Q., 19--21 August 1991, pages 341--353.

No context found.

A. Ricciardi, K. Birman, "Using Process Groups to Implement Failure Detection in Asynchronous Environments", in Proceedings of the loth ACM Symposium on Principles of Distributed Computing, pp. 341-352, August 1991.

First 50 documents  Next 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