K. Birman and T. Joseph. Exploiting Virtual Synchrony in Distributed Systems. In Proc. of the 11th ACM Symp. on Operating Systems Principles, pages 123--138, December 1987.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....group communication, random graphs. 1INTRODUCTION AND BACKGROUND L ARGE SCALE reliable group communication. Reliable group communication protocols are essential for distributed systems and applications such as publish subscribe systems [8] distributed databases [6] consistency management [13] and distributed failure detection [26] The growth of the Internet has influenced the scale and the reliability requirements of distributed systems. Traditional solutions applicable in small scale settings often do not scale well to very large system sizes. Network layer multicast protocols ....

....negative acknowledgments to repair packet losses. However, IP multicast is not currently deployed in the Internet. Consequently, application level multicast [17] has recently received increasing attention. Centralized or partially centralized approaches, proven efficient in local area networks [3] 18] do not scale well to large groups. For instance, log based reliable multicast (LBRM) 16] uses loggers to provide stable storage and to handle retransmission of missing messages; however, the amount of information to be stored grows with the number of nodes and loggers could be ....

[Article contains additional citation context not shown here]

K.P. Birman and T.A. Joseph, "Exploiting Virtual Synchrony in Distributed Systems," Proc. ACM Symp. Operating Systems Principles (SOSP), Dec. 1987.

....Every message destined to a multicast group is reliably disseminated among its current members. Traditionally, GCSs were employed mostly in replicated object and database applications [KD96, Ami95, KA98] These applications utilize GCS services with strong semantics (e.g. virtual synchrony [BJ87a] for supporting consistency and fault tolerance. Since the implementation of such services involves agreement protocols, the system performance degrades significantly as the group size and the message transmission volume increase. These performance problems become even worse in wide area network ....

K. P. Birman and T. A. Joseph. Exploiting virtual synchrony in distributed systems. In Proceedings of the 11th Annual ACM Symposium on Operating Systems Principles, pages 123--138, November 1987.

....RELATED WORK Our work deals with solving the Consensus problem [33] one of the most fundamental problems in distributed computing. Consensus is the building block for replication paradigms such as state machine replication [30, 45] group membership (see [41, 11] for survey) virtual synchrony [5], atomic broadcast [10] total ordering of messages [29, 19] etc. Consensus is known to be unsolvable in most realistic models such as asynchronous message passing systems [18] and asynchronous shared memory with read write registers [36, 27, 16] if even a single process can fail by crashing. ....

i K. Birman and T. Joseph. Exploiting virtual synchrony in distributed systems. In Proceedings of the 11th Annual Symposium on Operating Systems Principles, pages 123--138, November 1987.

....network. 1 Introduction Reliable multicast has been recognized as a key feature in many distributed systems, as it allows to reliably disseminate the same message to a large number of recipients. Consequently, reliable multicast is supported by many middlewares such as group communication (Isis [5], Horus [24] Transis [8] Ensemble [1] Relacs [4] Phoenix [17] and Totem [19] to name a few) protocols like RMTP [20] and SRM [9] and in the near future standards like CORBA [2] Reliable multicast typically involves storing copies of each message either by several dedicated servers, or as ....

....this is also an important aspect of hypercube based protocols (and tree based protocols) since in practical distributed systems, the underlying network topology is rarely known, and might change as the system and or network evolve. 1. 1 Related Work Many group communication toolkits, e.g. ISIS [5], Horus [24] and Ensemble [12] employ a fully distributed protocol, along the lines of the FullDist protocol we present in Section 2. As we discuss later in Section 3, this protocol is not very scalable. Guo et al. investigated the scalability of a variety of stability detection protocols in ....

K. Birman and T. Joseph. Exploiting Virtual Synchrony in Distributed Systems. In Proc. of the 11th ACM Symp. on Operating Systems Principles, pages 123--138, December 1987.

....Application processes can communicate with the members of a group by addressing messages to the group. Most GCSs strive to present di#erent members of the same group with mutually consistent perceptions of the communication done in the group. This perception is known as Virtual Synchrony semantics [12]. Traditionally, GCSs were designed for deployment in local area networks (LANs) E#cient GCSs that operate in wide area networks (WANs) is still an open area of research. Designing such GCSs is challenging because in WANs communication is more expensive and connectivity is less stable than in ....

....other sees her as alive and well. Middleware systems that hide from the application some of the underlying inconsistencies and instead present them with a more consistent picture of the distributed execution facilitate development of distributed applications. Group communication services, such as [3, 5, 44, 12], are examples of such middleware systems. They are particularly useful for building applications that require reliable multi point to multi point communication among a group (or groups) of processes. Examples of such applications are data replication (for example, 29, 4, 22, 33, 24] ....

[Article contains additional citation context not shown here]

K. Birman and T. Joseph, Exploiting virtual synchrony in distributed systems, in 11th ACM SIGOPS Symposium on Operating Systems Principles (SOSP), ACM, Nov 1987, pp. 123-- 138.

....Israel 1 Introduction Reliable multicast has been recognized as a key feature in many distributed systems, as it allows reliable dissemination of the same message to a large number of recipients. Consequently, reliable multicast is supported by many middlewares such as group communication (Isis [5], Horus [35] Transis [10] Ensemble [1] Relacs [3] Phoenix [23] and Totem [26] to name a few) protocols like RMTP [28] and SRM [12] and in the near future standards like CORBA [27] Reliable multicast typically involves storing copies of each message either by several dedicated servers, or ....

....based protocols (and tree based protocols) since in practical distributed systems, the underlying network topology is rarely known, and might change as either the system or the network changes or both of them evolve. 1. 1 Related Work Many group communication toolkits, for example, ISIS [5], Horus [35] and Ensemble [16] employ a fully distributed protocol, along the lines of the FullDist protocol we present in Section 2. As we discuss later in Section 3, this protocol is not very scalable. Guo et al. investigated the scalability of a variety of stability detection protocols in ....

K. Birman and T. Joseph. Exploiting Virtual Synchrony in Distributed Systems. In Proc. of the 11th ACM Symp. on Operating Systems Principles, pages 123--138, December 1987.

....asymmetric [4,13,19] with the exception of [16] and they all require that less than one third of the participants are malicious. Group membership is usually changed in a synchronous fashion by a separate agreement protocol that interrupts normal message delivery (the Virtual Synchrony paradigm [2, 15]) We present here a new methodology for designing symmetric protocols for use in malicious settings. Characteristic to our approach is the use of only one type of message. It is broadcast to all other participants and primarily carries application data. It also carries auxiliary information that ....

Kenneth P. Birman and Thomas Joseph. Exploiting Virtual Synchrony in distributed systems. In Proceedings of the 11th ACM Symposium on Operating Systems Principles, pages 123--138, Austin, Texas, November 1987.

....t from which a set of connected servers does not experience any crashes recoveries or network partitions merges, the servers will switch to the operational (RUN) state. In order to guarantee these properties we rely on the group communication system to follow the Virtual Synchrony properties [6, 14] in partitionable systems and to provide Agreed message delivery. The Virtual Synchrony property specifies that any two servers that advance together from one membership to the next one, will deliver an identical set of messages in the first membership. The Agreed delivery property guarantees ....

K. P. Birman and T. A. Joseph. Exploiting virtual synchrony in distributed systems. In Proceedings of the ACM Symposium on OS Principles, pages 123--138, 1987.

....processes can communicate with the members of a group by addressing messages to the group. Most GCSs strive to present di#erent members of the same group with mutually consistent perceptions of the communication done in the group. This perception is known as virtual synchrony (VS) semantics [12]. Traditionally, GCSs were designed for deployment in local area networks (LANs) E#cient GCSs that operate in wide area networks (WANs) is still an open area of research. Designing such GCSs is challenging because in WANs communication is more expensive and connectivity is less stable than in ....

....disconnected, while the other sees her as alive and well. Middleware systems that hide from the application some of the underlying inconsistencies and instead present them with a more consistent picture of the distributed execution facilitate development of distributed applications. GCSs, such as [3, 5, 44, 12], are examples of such middleware systems. They are particularly useful for building applications that require reliable multipoint to multipoint communication among a group (or groups) of processes. Examples of such applications are data replication (for example, 29, 4, 22, 33, 24] highly ....

[Article contains additional citation context not shown here]

K. Birman and T. Joseph, Exploiting virtual synchrony in distributed systems, in Proceedings of the 11th ACM SIGOPS Symposium on Operating Systems Principles (SOSP), 1987, pp. 123--138.

....multicast system is Overcast [1] Like Bayeux, Overcast requires that joining nodes coordinate with a central root node. A significant amount of work has also gone into overlay networks and application level multicast systems not designed to scale, such as Resilient Overlay Networks (RONs) [16], End System Multicast [3] and ISIS Horus style Virtual Synchrony [17] but which provide other benefits. Of course, all the work for constructing multicast distribution trees builds upon the techniques originally developed for IP Multicast [18] 19] VII. CONCLUSION We have explored some of ....

K. Birman and T. Joseph, "Exploiting virtual synchrony in distributed systems," in Proc of the ACM SOSP, November 1987, pp. 123--138.

....2 (Liveness) If there is a time t from which a set of connected servers does not experience any network events, the servers will switch to the operational (RUN) state. In order to guarantee these properties we rely on the group communication system to follow the Virtual Synchrony properties [3, 9] in partitionable systems and to provide Agreed message delivery. The Virtual Synchrony property specifies that any two servers that advance together from one membership to the next one, will deliver an identical set of messages in the first membership. The Agreed delivery property guarantees that ....

K. P. Birman and T. A. Joseph. Exploiting virtual synchrony in distributed systems. In Proceedings of the ACM Symposium on OS Principles, pages 123--138, Austin, TX, 1987.

....state of the backups. One way to enforce linearizability is to let the primary use a view synchronous multicast primitive to send the updates. The implementation of this primitive passes through a protocol executed by the replicas in order to 30 agree on the set of messages exchanged in a view [18]. 2.4. A desirable property for software replication Clients and server replicas implementing a replication technique should enjoy the following architectural property. Client server asynchrony. Clients (respectively server replicas) run over a distributed system without 35 any timing ....

Birman K, Joseph T. Exploiting virtual synchrony in distributed systems. Proceedings of the 11th ACM Symposium on 15 Operating Systems Principles, December 1987; 123--138.

....we describe the interface of each of these modules. All functions are non blocking and their results are returned as events. The membership of both participant and site groups at a given instant is called a view. The concept was defined in the context of the view synchrony group semantics [4, 3]. 66 Informally, view synchrony provides membership information to participants (or sites) in the form of views and guarantees that all participants (or sites) that install two consecutive views deliver the same set of messages between these views. However, the architecture of MAFTIA middleware ....

....p BA deliverable. A message M in Pending Bu#er p (M header.status = pending) could only be delivered when the outcome of one of the following rounds the MSD indicates that M was finally BR deliver by all correct processes. The MSD of the BAM protocol is based on the virtual synchrony model [3]. The views are ordered in an acyclic sequence. The MSD protocol is executed in each process, in asynchronous rounds, and in a sequential manner each round is used to detect a set of stable messages (BA deliverable = #M: M header.status = stable ) In the Figure 5.6 is presented the ....

K. Birman and T. Joseph. Exploiting virtual synchrony in distributed systems. In Proceedings of the 11th Symposium on Operating System Principles, pages 123--138, November 1987.

....semantics include: FIFO, Reliable FIFO, Causal, Reliable Causal and Totally Ordered [23] However, most reliable multicast semantics can at least guarantee the property of virtual synchrony. Virtual Synchrony semantics specify how group communication is synchronized with group view delivery [24]. The essence of the virtual synchrony principle is that it guarantees that membership changes within a process group are observed in the same order by all the group members that remain connected. Virtual synchrony guarantees that every two processes that observe the same two consecutive group ....

K. Birman and T. Joseph, "Exploiting Virtual Synchrony in Distributed Systems.", In 11th ACM SIGOPS Symposium on Operating Systems Principles, November 1987.

....that the following four attributes 1 characterize a primary backup system and note that several purported primarybackup protocols (e.g. 1,2,5,8,3] satisfy this characterization. There are, however, purported primary backup protocols that do not satisfy one or more of these attributes (e.g. [4]) pb1: There exists a state predicate PRMY on the state of any server such that PRMY is true on the state of at most one server. pb2: At each client c i there exists a server identity PRMY i . At any time, c i may make a request by sending that request only to PRMY i . Thus, a client interacts ....

Kenneth P. Birman and Thomas A. Joseph. Exploiting Virtual Synchrony in Distributed Systems. In Eleventh ACM Symposium on Operating System Principles, pages 123--138, November 1987.

....time. Finally, in this paper we have attempted to give a characterization of primary backup that is broad enough to include most synchronous protocols that are considered to be instances of the approach. There are protocols, however, that are incomparable to the class of protocols we analyze [5, 16, 18] since they were developed for an asynchronous setting. Such protocols cannot be cast in terms of implementing a (k, A) bofo server for finite values of k and A. We are currently studying possible characterizations for a primary backup protocol in an asynchronous system and hope to extend our ....

Kenneth P. Birman and Thomas A. Joseph. Exploiting virtual synchrony in distributed systems. In Eleventh A CM Symposium on Operating System Principles, pages 123-138, November 1987.

....other sees her as alive and well. Middleware systems that hide from the application some of the underlying inconsistencies and instead present them with a more consistent picture of the distributed execution facilitate development of distributed applications. Group communication services, such as [6, 8, 93, 19], are examples of such middleware 15 systems. They are particularly useful for building applications that require reliable multipoint to multi point communication among a group (or groups) of processes. Examples of such applications are data replication (for example, 57, 7, 39, 63, 42] and [30] ....

....services; in particular, they specify how message deliveries are synchronized with view deliveries. This synchronization is done in a way that simulates a benign world in which message delivery is reliable within each view. Many variants of Virtual Synchrony have been suggested (for example, [77, 43, 27, 19, 82, 12]) In addition to other properties, nearly all of them include a key property, called Virtually Synchronous Delivery, which guarantees that processes that receive the same pair of views from the GCS receive the same sets of messages in between receiving the views. Henceforth, when we refer to ....

K. Birman and T. Joseph. Exploiting virtual synchrony in distributed systems. In 11th ACM SIGOPS Symposium on Operating Systems Principles (SOSP), pages 123--138. ACM, Nov 1987.

.... on predecessors, in which a set of messages is agreed upon and delivered at all processes before delivering the change notification; virtual synchrony, in which membership change notifications are delivered to the application at the same point in the application message stream at all processes 2o [Birman Joseph 1987, Schiper Ricciardi 1993] extended virtual synchrony, in which all messages sent by the application prior to receipt of the change notification are guaranteed to be delivered before the change notification itself [Amir et al. 1993, Moser et al. 1994] external synchrony, in which, when a ....

K. P. Birman and T. A. Joseph, "Exploiting Virtual Synchrony in Distributed Systems", ACM Operating Systems Review, 21 (5), pp.123-38, 1987.

....tools provided by these products, together with Astrolabe, could form an interesting marriage. 37 8.3 Event Notification Event Notification or Publish Subscribe services allow applications to subscribe to certain classes of events of interest, and systems to post events. Examples are ISIS [6], TIB RendezVous tm [19] Gryphon [2] and Siena [10] Note that although Astrolabe processes events, it does not route them to subscribers, but processes them to determine some aggregate state snapshot of a distributed system. However, a Publish Subscribe system has been built on top of ....

K. P. Birman and T. A. Joseph. Exploiting virtual synchrony in distributed systems. In Proc. of the Eleventh ACM Symp. on Operating Systems Principles, pages 123--138, Austin, TX, November 1987.

No context found.

K. Birman and T. Joseph. Exploiting Virtual Synchrony in Distributed Systems. In Proc. of the 11th ACM Symp. on Operating Systems Principles, pages 123--138, December 1987.

No context found.

K. P. Birman and T. A. Joseph. Exploiting virtual synchrony in distributed systems. In Proceedings of the ACM Symposium on OS Principles, pages 123--138, Austin, TX, 1987.

No context found.

Birman, K. P. and Joseph, T. A. 1987a. Exploiting virtual synchrony in distributed systems. In Proceedings of the 11th ACM Symposium on Operating Systems Principles (11th SOSP87), Operating Systems Review. 123--138.

No context found.

K. P. Birman and T. A. Joseph. Exploiting virtual synchrony in distributed systems. In Proceedings of the 11th ACM Symposium on Operating Systems Principles (11th SOSP87), Operating Systems Review, pages 123--138, Austin, Texas, Nov. 1987.

No context found.

K. P. Birman and T. A. Joseph. Exploiting virtual synchrony in distributed systems. In Proceedings of the ACM Symposium on OS Principles, pages 123--138, Austin, TX, 1987.

No context found.

Birman, K. P. and Joseph, T. A. Exploiting Virtual Synchrony in Distributed Systems, in Eleventh Symposium on Operating Systems Principles, (Austin, Texas, 1987).

No context found.

K. Birman and T.A. Joseph. Exploiting virtual synchrony in distributed systems. In Proc. 11th ACM Symp. on Operating System Principles, Austin, Texas, Nov. 1987. Also available as technical report TR 87-811 from the Dept. of Computer Science, Cornell Univ., Feb. 1987.

No context found.

K. P. Birman and T. Joseph. Exploiting Virtual Synchrony in Distributed Systems. In 11th Ann. Symp. Operating Systems Principles, pages 123--138, Nov 87.

No context found.

Ken Birman and Thomas Joseph. Exploiting virtual synchrony in distributed systems. In 11th ACM Symp. on Operating System Principles, pages 123-- 138. ACM SIGOPS, 1987.

No context found.

K. P. Birman and T. Joseph, "Exploiting virtual synchrony in distributed systems," in 11 Annual Symposium on Operating Systems Principles, pp. 123--138, November 1987.

No context found.

K. Birman, T. Joseph. Exploiting Virtual Synchrony in Distributed System, Proceedings of the Eleventh Symposium on Operating Systems Principles, pp 123-138. Austin, November 1987. 9

No context found.

K. Birman, T. Joseph. Exploiting Virtual Synchrony in Distributed System, Proceedings of the Eleventh Symposium on Operating Systems Principles, pp 123-138. Austin, November 1987.

No context found.

K. P. Birman and T. A. Joseph. Exploiting virtual synchrony in distributed systems. In Proceedings of the 11th ACM Symposium on Operating Systems Principles, pages 123-- 138. ACM SIGOPS, November 1987.

No context found.

Birman KP, Joseph TA. Exploiting virtual synchrony in distributed systems. Proceedings of the 11th Symposium on Operating Systems Principles, Austin, TX, November 1997. Association for Computing Machinery: New York, NY, 1997; 123--138.

No context found.

K. Birman and T. Joseph. Exploiting Virtual Synchrony in Distributed Systems. In Proc. of the 11th ACM Symp. on Operating Systems Principles, pages 123--138, December 1987.

No context found.

Kennneth P. Birman and Thomas A. Joseph. Exploiting virtual synchrony in distributed systems. In Proceedings of the 11th ACM Symposium on Operating Systems Principles (11th SOSP87), Operating Systems Review, pages 123--138, Austin, Texas, November 1987.

No context found.

K. Birman and T. Joseph. Exploiting virtual synchrony in distributed systems. In Proceedings of the 11th Annual Symposium on Operating Systems Principles, pages 123--138, November 1987.

No context found.

Kenneth P. Birman and Thomas A. Joseph. Exploiting virtual synchrony in distributed systems. Proc 11th Symposium on Operating Systems Principles (SOSP), pages 123--138, November 1987.

No context found.

8 K. Birman and T. Joseph. Exploiting Virtual Synchrony in Distributed Systems. In 11th Ann. Symp. Operating Systems Principles, pages 123--138. ACM, Nov 87.

No context found.

K. P. Birman and T. Joseph. Exploiting Virtual Synchrony in Distributed Systems. In 11th Ann. Symp. Operating Systems Principles, pages 123-138, Nov 87.

No context found.

Ken Birman and Thomas Joseph. Exploiting virtual synchrony in distributed systems. In Proceedings of the 11th ACM Symposium on Operating Systems Principles, December 1987.

No context found.

Kenneth P. Birman, and Thomas A. Joseph, "Exploiting Virtual Synchrony in Distributed Systems", Department of Computer Science, Cornell University, 1987.

No context found.

K. Birman, T. Joseph. Exploiting virtual synchrony in distributed systems. In SOSP, 1987.

No context found.

K. P. Birman and T. A. Joseph. Exploiting virtual synchrony in distributed systems. In Proceedings of the 11th ACM Symposium on Operating Systems Principles (11th SOSP87), Operating Systems Review, pages 123--138, Austin, Texas, Nov. 1987.

No context found.

K. P. Birman and T. A. Joseph. Exploiting Virtual Synchrony in Distributed Systems. In Proc. of 11th ACM Symp. on Operating Systems Principles, pages 123--138, Austin, TX, Nov. 1987.

No context found.

Birman, K.P., Joseph, T.A.: Exploiting virtual synchrony in distributed systems. In: Proceedings of the 11 ACM Symposium on OS Principles, Austin, TX, USA, ACM SIGOPS, ACM (1987) 123--138

No context found.

K. Birman and T. Joseph. Exploiting Virtual Synchrony in Distributed Systems. In 11th Ann. Symp. Operating Systems Principles, pages 123-138. ACM, Nov 87.

No context found.

K. P. Birman and T. Joseph, "Exploiting virtual synchrony in distributed systems," in Proc. of 11th ACM Symposium on Operating Systems Principles, 1987, pp. 123--138.

No context found.

# K.P. Birman and T.A. Joseph, "Exploiting Virtual Synchrony in Distributed Systems," Proc. 11th ACM Symp. Operating Systems Principles, pp. 123--138, Austin, Texas, Nov. 1987.

No context found.

Kenneth P. Birman and Thomas A. Joseph. Exploiting virtual synchrony in distributed systems. In Proc 11th Symposium on Operating Systems Principles #SOSP#, pages 123#138, November 1987.

No context found.

Binnan, K.P., and T.A. Joseph. "Exploiting Virtual Synchrony in Distributed Systems." Operating System Review Op. Sys. Review Vol. 21 No. 5, pp 123-138.

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