Luan, S. and Gligor, V.D. : "A Fault Tolerant Protocol for Atomic Broadcast", IEEE Trans. on Parallel and Distributed Systems, vol. 1, No. 3, pp. 271 -- 285 (1990).  Home/Search   Document Not in Database   Summary   Related Articles   Check

....Asymmetric protocols ( 7, 9, 21] use a centralized coordinator for ordering the messages. The problems with this method are in the serial bottleneck they create at the coordinating site, and with the costly handling of faults in case the coordinator crashes. Existing symmetric protocols ([28, 6, 17, 24]) require all machines to relay their up todate view on the latest delivered messages in order for a new message to be admitted into the total order. These methods are essentially all ack, i.e. require all machines to send ack for a single message admittance. Unfortunately, all ack protocols ....

....ordering of messages have been designed that circumvent the impossibility result. Synchronous protocols ( 16, 12, 18] circumvent the impossibility result by explicitly assuming synchrony. Probabilistic protocols ( 4, 29, 8, 10, 13, 26] introduce random steps to the protocol. In other protocols ([9, 7, 6, 24]) the system halts and reconfigures when processors fail or join. In Transis, the impossibility result is circumvented in the automatic maintenance of dynamic membership ( 1] The impossibility result stems from the inability to distinguish between slow and faulty machines in the asynchronous ....

[Article contains additional citation context not shown here]

S. W. Luan and V. D. Gligor. A fault-tolerant protocol for atomic broadcast. IEEE Trans. Parallel & Distributed Syst., 1(3):271--285, July 90.

....size contains the data structures necessary to perform the certi cation test (e.g. readset and writeset) Atomic broadcast settings are described in the next section. 6. 2 Atomic Broadcast Implementation The literature on atomic broadcast algorithms is abundant (e.g. 14] 15] 16] 17] [18], 19] 20] and the multitude of di erent models (synchronous, asynchronous, etc. and assumptions about the system renders any fair comparison dicult. However, known atomic broadcast algorithms can be divided into two classes. We say that an Atomic Broadcast algorithm scales well, and belongs ....

S. W. Luan and V. D. Gligor, \A Fault-Tolerant Protocol for Atomic Broadcast," IEEE Transactions on Parallel & Distributed Syst., vol. 1, pp. 271-285, July 1990.

.... systems motivated this work, and are described in [2, 8] Atomic broadcast has been traditionally used to achieve fault tolerance by replicating functionality (e.g. using the state machine approach [18] and previous work on atomic broadcast algorithms are in the context of fault tolerance [10, 6, 11, 16, 15, 14, 5, 9, 1]. In these works, one main concern is to guarantee total order and agreement when the broadcaster fails during the broadcast. To obtain this guarantee, a recipient needs to send messages to other recipients, either to inform them of a previously delivered message, or to inform them of the intent ....

.... do not fail or, equivalently, that if they fail then they always recover and can retrieve all previously broadcast messages from stable storage) Deterministic merge algoriths for atomic broadcasts are special cases of symmetric protocols, in which all recipients execute the same code (e.g. [6, 15, 9]) In contrast, asymmetric protocols are usually based on coordinators or tokens (e.g. 5, 1] Broadcasts with weaker and stronger semantics than atomic broadcast have been proposed, such as reliable broadcast and causal broadcast. 12] provides a good overview. The Bellman equation first ....

S. W. Luan and V. D. Gligor. A fault-tolerant protocol for atomic broadcast. IEEE Transactions on Parallel and Distributed Systems, 1(3):271--285, July 1990.

.... recently published well known protocols algorithms on reliable multicast are surveyed in the next two sections, such as Isis and Horus systems of Birman [1, 2] token list [3] and its implementation [4] sequencer [5, 6] two phase [7, 8] tree [9] logical token ring [1, 10, 11] 3 phase [12], discrete acknowledgments [13, 14] and module [2, 15, 16] etc. We classify the well known algorithms as token passing, discrete acknowledgement and two phase approaches and briefly describe their features and associated problems. 1.1. Token passing approach Among the existing multicast ....

Luan, S. and Gligor, V. D. (1990) A fault-tolerant protocol for atomic broadcast. IEEE Trans. Parallel Distrib. Syst., 1, 271--285.

....handling is discussed in Section 7, and Section 8 concludes the paper. 2 Related Work The paper is at the intersection of two domains: 1) Atomic Broadcast algorithms, and (2) optimistic algorithms. The literature on Atomic Broadcast algorithms is abundant (e.g. 1] 3] 4] 5] 7] 10] [13], 15] However, the multitude of di#erent models (synchronous, asynchronous, etc. and as1 Spontaneous total order message reception occurs for example with very high probability when network broadcast or IP multicast are used. 2 sumptions needed to prove the correctness of the algorithms ....

S. W. Luan and V. D. Gligor. A Fault-Tolerant Protocol for Atomic Broadcast. IEEE Trans. Parallel & Distributed Syst., 1(3):271--285, July 90.

....In [48] the role of Atomic Broadcast in distributed systems is compared to that of message passing mechanism in the operating systems of today. Horus [50] is the object oriented extension of Isis. The algorithms in [51, 52] are for Atomic Multicast on a point to point network over a spanning tree. [53] uses a 3 phase commit protocol to order messages. Their algorithm uses voting to avoid blocking. The efficiency is low at low loads but increases with increasing load. Amoeba [23] implements a reliable totally ordered multicast, at the kernel level. A sequencer in the group orders messages for ....

V. Gligor and W. Luan, "A fault-tolerant protocol for atomic broadcast," IEEE Trans. Par. Distr. Syst., vol. 1, no. 3, pp. 271--285, 1990.

....networks are low ( 10 09 ) entities may fail to receive PDUs due to the buffer overrun because the processing speed of the entities is slower than the transmission speed [1] Schneider et al. 17] present a reliable broadcast protocol which uses the one to one communication. Luan and Gligor [12] present a broadcast protocol which provides the total ordering of received messages based on majority consensus decision. Garcia Molina and Spauster [6] characterize message ordering properties in a reliable broadcast protocol using the one to one network. Takizawa [19, 20] presents a cluster ....

Luan, S. W. and Gligor, V. D., "A Fault-Tolerant Protocol for Atomic Broadcast," IEEE Trans. Parallel and Distributed Systems, Vol.1, No.3, 1990, pp.271-285.

....the states of the objects. Here, a group of multiple objects have to be communicated. That is, objects send messages to multiple objects in the group. It is important to support the causally ordered delivery of messages in the distributed applications. Many group communication protocols [1, 3, 4, 6, 10, 11, 13, 16, 18 22, 24, 25, 28 30] have been discussed so far. They have discussed how to support the atomic and ordered delivery of network messages, i.e. packets at the communication level. In order to support the atomic and ordered delivery of messages in the presence of message loss and object faults, O(n 2 ) processing ....

Luan, S. W. and Gligor, V. D., "A Fault-Tolerant Protocol for Atomic Broadcast," IEEE Trans. Parallel and Distributed Systems, Vol.1, No.3, 1990, pp.271--285.

....to stable storage is required when (re)entering the prepared state. 7 Related Work Several other proposals have been made for deploying commit protocols as part of a telecommunications service and we give three examples here. 7. 1 Atomic Broadcast Luan and Gligor s atomic broadcast algorithm [9] has each site (end point nodes) accumulating received broadcast messages in a queue. When an application at one of the sites wants to consume a prefix of these messages, it initiates a three phase commit protocol in which the other sites vote according to whether they have received all messages ....

S. Luan and V.D. Gligor. A fault-tolerant protocol for atomic broadcast. IEEE Transactions on Parallel and Distributed Systems, 1(3):271--285, July 1990.

....considered to be only failure in the highspeed network. In this paper, we would like to discuss how to provide atomic delivery of PDUs among multiple entities and some receipt ordering of them by using high speed broadcast networks. Reliable broadcast communication systems have been discussed in [4, 5, 10, 11, 13, 17, 18, 19, 20, 22, 24, 25, 26, 27]. 22] presents a reliable broadcast protocol which uses one to one communication. 5, 10, 13] discuss centralized protocols which use Ethernet. 11] characterizes message ordering properties in a reliable broadcast protocol using the conventional one to one network. 24, 25, 19, 20, 21] present a ....

Luan, S. W. and Gligor, V. D., "A Fault-Tolerant Protocol for Atomic Broadcast," IEEE Trans. on Parallel and Distributed Systems, Vol.1, No.3, 1990, pp.271-285.

....The PO protocol [16] provides the LO service. That is, if an entity broadcasts two PDUs p and q in this order, every destination entity receives q after p. In the TO service, all the destinations receive PDUs in the same order in addition to the sending order. Most reliable broadcast protocols [3, 4, 6, 7, 9, 10, 14, 15] provide the TO service. In the CO service, PDUs received are ordered by the happened before relation [8] If p is sent logically before q, p is delivered to every destination before q. In the LO and TO services, it is discussed in what order each entity can receive PDUs. In the CO service, it ....

Luan, S. W. and Gligor, V. D., "A Fault-Tolerant Protocol for Atomic Broadcast," IEEE Trans. on Parallel and Distributed Systems, Vol.1, No.3, 1990, pp.271-285.

....needed for both ATC and WKC is (N 1) In these protocols, a message loss in the network is handled by timeout based retransmissions of the message. Other protocols to totally order messages or for atomic message delivery are also possible such as acknowledgements piggybacked on other messages [2, 12, 13] 2 . 3.2 Impact of communication semantics on application level mechanisms The communication requirements of an application have to be mapped onto one of O Bcast, A Bcast and W Bcast semantics. Since ordering of messages is the strongest requirement in applications and it matches exactly with ....

S. W. Luan and V. D. Gligor. A Fault-tolerant Protocol for Atomic Broadcast. IEEE Transactions on Parallel and Distributed Systems Vol.1, No.3, pp.271--285, July 1990.

....among multiple objects. Then, messages sent by each object are delivered to the destination objects in the group. This type of group communication is intragroup communication [16, 17, 18, 20] In this paper, we discuss the intra group communication among multiple application objects. The papers [1, 3, 4, 6, 8, 9, 12, 14, 16, 17, 18, 20, 23] have discussed how to support the causally ordered delivery of messages at the network level in the presence of omission faults of the network, i.e. message loss and stop faults of the objects. O(n 2 ) processing overhead and O(n) to O(n 2 ) communication overhead are implied for number n of ....

Luan, S. W. and Gligor, V. D., "A Fault-Tolerant Protocol for Atomic Broadcast," IEEE Trans. Parallel and Distributed Systems, Vol.1, No.3, pp. 271-285, 1990.

....the group communication, messages sent by one process have to be delivered to either all the destinations or none in the group, i.e. atomic delivery . In addition, each process has to receive messages sent by the processes in some order. Group communications have been studied in [3, 4] 7] 9] [11, 12], 13] 16] and [17] 21] 17] presents a reliable broadcast protocol which uses the one to one communication. An important problem in the group communication is which process coordinates the cooperation of multiple processes in the group. Most approaches [4, 7, 17] adopt the centralized ....

Luan, S. W. and Gligor, V. D., "A Fault-Tolerant Protocol for Atomic Broadcast," IEEE Trans. on Parallel and Distributed Systems, Vol.1, No.3, 1990, pp.271-285.

....and they send messages to the objects in the group. Another type of group communication is multicast [3, 4] where each object sends messages to group or groups of objects. In this paper, we would like to discuss the intra group communication among multiple application objects. Many papers [1,3,4,6,8,10,13,15,17 19,21,22,25,26] have discussed how to support the reliable and ordered delivery of network messages, i.e. packets at the network level in the presence of message loss and object faults. O(n 2 ) processing overhead and O(n) to O(n 2 ) communication overhead are implied for number n of objects in the group ....

Luan, S. W. and Gligor, V. D., "A Fault-Tolerant Protocol for Atomic Broadcast," IEEE Trans. Parallel and Distributed Systems, Vol.1, No.3, 1990, pp.271--285.

....objects. Then, messages sent by each object are delivered to the destination objects in the group. This type of group communication is intra group communication [16 18, 20] In this paper, we would like to discuss the intra group communication among multiple application objects. Many papers [1,3, 4,6, 8,9,12,14, 16 18,20,23] have discussed how to support the causally ordered delivery of messages at the network level in the presence of message loss and stop faults of the objects. O(n 2 ) processing overhead and O(n) to O(n 2 ) communication overhead are implied for number n of objects in the group [16] On the ....

Luan, S. W. and Gligor, V. D., "A Fault-Tolerant Protocol for Atomic Broadcast," IEEE Trans. Parallel and Distributed Systems, Vol.1, No.3, 1990, pp.271--285.

....parameters Communication parameters Control data size 1 KB Atomic Broadcasts per second 1; 180; 800=n Communication overhead 12000 ins Figure 6: Simulation model parameters 6. 2 Atomic Broadcast Implementation The literature on atomic broadcast algorithms is abundant (e.g. 7] 9] 8] 11] [23], 33] 5] and the multitude of different models (synchronous, asynchronous, etc. and assumptions about the system renders any fair comparison difficult. However, known atomic broadcast algorithms can be divided into two classes, according to scalability issues. An atomic broadcast algorithm ....

S. W. Luan and V. D. Gligor. A Fault-Tolerant Protocol for Atomic Broadcast. IEEE Trans. Parallel & Distributed Syst., 1(3):271--285, July 1990.

....space limitations, all proofs have been omitted. They can be found in [12] 2 Related Work The paper is at the intersection of two issues: 1) Atomic Broadcast algorithms, and (2) optimistic algorithms. The literature on Atomic Broadcast algorithms is abundant (e.g. 1] 3] 4] 5] 6] 8] [11], 14] However, the multitude of different models (synchronous, asynchronous, etc. and assumptions needed to prove the correctness of the algorithms renders any fair comparison difficult. We base our solution on the Atomic Broadcast algorithm as presented in [4] because it provides a ....

S. W. Luan and V. D. Gligor. A Fault-Tolerant Protocol for Atomic Broadcast. IEEE Trans. Parallel & Distributed Syst., 1(3):271--285, July 90.

....delivery of a multicast message to either all or none of the destinations. In addition, many protocols guarantee that multicast messages adressed to a group are received in the same order by all the members of the group. A number of reliable multicast protocols have been proposed [1] 2] 3] [4], 5] 6] Each of these protocols ensures reliable transfer of messages with different types of message ordering. On the other hand, an unreliable IP multicast service [7] is now supported on a wide part of the internet the MBONE or Multicast backBONE. However, most the reliable multicast ....

S-W. Luan and V. D. Gligor, "A Fault-Tolerant Protocol for Atomic Broadcast," Proc. IEEE 7th Reliable Distributed Systems Symposium, pp. 112-126, 1988.

....Order. The sequence of delivered messages is identical for all group members. 3. Termination. Once a message has been successfully issued by a sender, eventually all group members will receive the message. Early distributed agreement protocols are based on variations of two or three phase commit [LG90] central control [BSS91] or token ring [BG93, AMMS 93] In addition to reliable and atomic multicast primitives, a membership facility provides virtually synchronous group of processes : all working processes receive the same set of messages between configuration changes. The communication ....

S. Luan and V. Gligor. A fault tolerant protocol for atomic broadcast. IEEE Transactions on Parallel and Distributed Systems, 1(3), 1990.

....mutual exclusion to shared resources, dissemination of multimedia information, announcing resource availability, announcing terminal conditions, initiating emergency procedures, etc. There is an extensive literature that deal with broadcast related issues in static networks, see for example [4, 6, 13, 16, 19, 17]. In this paper, we consider the problem of broadcast in cellular structured wireless networks that support user mobility. In such networks, broadcast proceeds in two phases. During the first phase, the broadcast message is sent from the originating or source station to a collection of base ....

S. Luan and V. Gligor, "A fault tolerant protocol for atomic broadcast," IEEE Transactions on Parallel and Distributed Systems, vol. 1, pp. 271 -- 285, July (1990).

....the asynchronous model which constitutes the most general model. In the context of asynchronous systems, the usual classification of total order algorithms distinguishes between symmetric and asymmetric algorithms. In a symmetric algorithm, all processes execute the same code [Lam78, CM84, BJ87, LG90, DKM93, CT93, Anc93] in an asymmetric algorithm, one process plays a special role, e.g. defines the ordering of messages [KTHB89, BSS91] The asymmetric algorithms are interesting as they are more efficient (they require less phases and or fewer messages than the symmetric algorithms) The ....

S. W. Luan and V. D. Gligor. A Fault-Tolerant Protocol for Atomic Broadcast. IEEE Trans. Parallel & Distributed Syst., 1(3):271--285, July 90.

....be delivered right away. Meanwhile, the communication in the system can continue as usual. At some later point, when enough information has been gathered from other machines, the message can be delivered. It is perhaps easiest to understand this through the simple idea of an all ack protocol (see [22, 4, 15, 20, 24]) ffl Wait for a message from each machine in the system. ffl Deliver the set of messages that do not causally follow any other, in lexicographical order. This kind of agreed service has a post transmission delay, caused by the first step. The length of the delay is set by the slowest ....

....Asymmetric protocols ( 6, 7, 17] use a centralized coordinator for ordering the messages. The problems with this method are in the serial bottleneck it creates at the coordinating site, and with the costly handling of faults in case the coordinator crashes. Most existing symmetric protocols ([22, 4, 15, 20, 24]) require all machines to relay their up to date view on the latest delivered messages in order for a new message to be added to the total order. These methods are essentially all ack, i.e. require all machines to send ack for a single message delivery. Unfortunately, all ack protocols typically ....

[Article contains additional citation context not shown here]

S. W. Luan and V. D. Gligor. A Fault-Tolerant Protocol for Atomic Broadcast. IEEE Trans. Parallel & Distributed Syst., 1(3):271--285, July 90.

....Asymmetric protocols ( 9, 11, 23] use a centralized coordinator for ordering the messages. The problems with this method are in the serial bottleneck they create at the coordinating site, and with the costly handling of faults in case the coordinator crashes. Most existing symmetric protocols ([29, 7, 21, 26, 31]) require all machines to relay their upto date view on the latest delivered messages in order for a new message to be admitted into the total order. These methods are essentially all ack, i.e. require all machines to send ack for a single message admittance. Unfortunately, all ack protocols ....

....m as the number of machines that p receives messages from, before it can deliver m. These messages may be concurrent to m, or follow m. The efficiency of an agreed multicast protocol may be measured in terms of the worst case and average case index of synchrony. Most existing symmetric protocols ([29, 7, 21, 26]) have index of synchrony n, the number of machines. It is easy to see that genuinely distributed protocols for total ordering of messages, such as ToTo, must bear an index of synchrony greater then n 2 (think of two messages received and acknowledged by two disjoint sets of n 2 processors) ....

S. W. Luan and V. D. Gligor. A fault-tolerant protocol for atomic broadcast. IEEE Trans. Parallel & Distributed Syst., 1(3):271--285, July 90.

....overflows due to the shortage of the processing time for messages. ii) is caused by the loop back mechanism of the operating system by which the messages sent from a node is stored to its message buffer directly. For the above problems, we can use the techniques proposed in (Melliar et al. 1990, Luan and Gligor 1990) so that the assumptions hold on Ethernet. However, for efficiency, here we adopt to place a proxy server which broadcasts messages forwarded from all nodes. Each node does not broadcast its messages directly but asks the server to broadcast its messages. By placing the server at an independent ....

Luan, S. and Gligor, V.D. : "A Fault Tolerant Protocol for Atomic Broadcast", IEEE Trans. on Parallel and Distributed Systems, vol. 1, No. 3, pp. 271 -- 285 (1990).

....the incremental changes to this graph. By maintaining the antecedence graph this protocol does not need to send acknowledgements to confirm that the message is stored at r members. On the other hand, the application must potentially be rolled back when a processor fails. The protocol described in [Luan and Gligor 1990] is one of the protocols that require additional hardware. In the LG protocol each member must be equipped with a disk. Using these disks the pro tocol can provide fault tolerant ordered broadcasting, even if the network partitions. It uses a majority consensus decision to agree on a unique ....

Luan, S. W. and Gligor, V. D., "A Fault-Tolerant Protocol for Atomic Broadcast," IEEE Trans. Parallel and Distributed Systems, Vol. 1, No. 3, pp. 271-285, July 1990.

....and global ordering 2 of messages and require a multicast transport scheme that delivers messages to group members from a single source in sequence and without loss. Group membership management has been considered in these schemes to varying degrees of detail, from a complete absence of it [9] to a fairly detailed discussion of it [3] The capabilities of even the most general of these mechanisms, however, is restricted by the environment in which the corresponding multicast protocol operates. For instance, the protocol given in [7] assumes a broadcast environment and the protocol ....

S-W. Luan and V. D. Gligor "A Fault-Tolerant Protocol for Atomic Broadcast," Proc. IEEE 7th Reliable Distributed Systems Symposium, pp 112-126, 1988.

....data are updated consistently in the presence of faults. Fault tolerance also requires that faulty processors, and any concomitant erroneous computations, are recognized and that the system is reconfigured to eliminate those faults. Several advanced designs of fault tolerant networks of computers [2, 7, 10, 11] employ the idea of placing a partial order and a total order on messages (typically broadcast within a process group) This approach to building fault tolerant networks of computers greatly simplifies the design of the application but depends heavily on efficient fault tolerant protocols to keep ....

Luan, S. W. and Gligor, V. D., "A fault tolerant protocol for atomic broadcast," IEEE Transactions on Parallel and Distributed Systems 1, 3 (July 1990), 271-285.

....primarily to display information from protocols that use positive and or negative acknowledgments to achieve atomic broadcasts and orders on messages [ 4, 6, 12 ] With the addition of retransmission request messages, the Psync protocol [ 16 ] can be displayed. The protocol by Erramilli and Singh [ 11 ] would require the addition of commit message types. More extensive modifications to the graphical interface would be required to display protocols, such as some of the ISIS protocols [ 1 ] which use rounds and timestamps to establish the ordering of a message instead of acknowledgments. The ....

S. W. Luan and V. D. Gligor, "A fault-tolerant protocol for atomic broadcast," Proceedings of the 7th Symposium on Reliable Distributed Systems, Columbus, OH (October 1988), 112-126.

....the same messages, even in the face of malicious multicast initiators. Atomic multicast adds the property that honest members deliver these messages in the same order. While reliable and atomic multicast protocols tolerant of only benign failures have been the focus of much systems research (e.g. [5, 16, 17, 13, 10, 3, 1]) relatively little has been done to extend these results to the more stringent requirements of security. In this paper we present new reliable and atomic group multicast protocols tolerant of malicious processes. We have implemented these protocols as part of Rampart, a toolkit for building ....

S. W. Luan and V. D. Gligor. A fault-tolerant protocol for atomic broadcast. IEEE Transactions on Parallel and Distributed Systems, 1(3):271--285, July 1990.

....21 32210.91, as part of the ESPRIT Basic Research Project BROADCAST (number 6360) and by PPR IF under contract number 5003 34344. classification of total order algorithms distinguishes between symmetric and asymmetric algorithms. In a symmetric algorithm, all processes execute the same code [13, 6, 3, 14, 8, 5, 1]; in an asymmetric algorithm, one process plays a special role, e.g. defines the ordering of messages [11, 4] The asymmetric algorithms are interesting as they are more efficient (they require less phases and or fewer messages than the symmetric algorithms) The classification ....

S. W. Luan and V. D. Gligor. A Fault-Tolerant Protocol for Atomic Broadcast. IEEE Trans. Parallel & Distributed Syst., 1(3):271--285, July 90.

....is a difficult problem (that has been shown to be as difficult as the consensus problem, see Sect. 3.2) and (2) the TO multicast is precisely the adequate mechanism to lock replicated objects. TO multicast based on DTM Implementations for TO multicasts have been described in numerous papers [3, 4, 5, 9, 17, 21]. Liveness is however not ensured for any of these algorithms when assuming only the properties of the failure detector 3S. This is not true for the total order broadcast algorithm given in [6] This algorithm implements however TO broadcast, and not TO multicast. TO multicast can be implemented ....

S. Luan and V. Gligor. A fault-tolerant protocol for atomic broadcast. IEEE Transactions on Parallel and Distributed Systems, 1(3) pages 271-285. July 1990.

....protocols are given in [46] 2.1. TCP 11 Some protocols for reaching consensus in asynchronous systems are asymptotic in the sense that the probability of reaching consensus asymptotically approaches one as time increases. An example of an asymptotic atomic broadcast protocol can be found in [38]. Much of the difficulty in reaching consensus in an asynchronous system is caused by the difficulty of distinguishing between a failed processor and a slow processor. Many protocols use timeouts to detect failure; these failure detectors are termed unreliable because they may sometimes ....

S. W. Luan and V. D. Gligor. A fault-tolerant protocol for atomic broadcast. In Proceedings of the 7th Symposium on Reliable Distributed Systems, pages 112--126, Columbus, Ohio, October 1988.

....also a semantic ordering, delivering messages early by giving a priority to read operations. Failure notifications are causally ordered, but the membership protocol does not guarantee that all group members view the same sequence of configuration changes (virtual synchrony) Luan and Gligor [58] derived an atomic broadcast from the three phase commit protocol. In the first phase, a process initiates a voting round. When enough acknowledgments have been collected, the initiator enters a notification phase to specify which messages need to be ordered. If enough acknowledgments are also ....

Luan and Gligor. A fault-tolerant protocol for atomic broadcast. IEEE Transactions on Parallel and Distributed Systems, 1(3):271--285, 1990.

.... Total order multicast in asynchronous systems Many total order multicast algorithms for the asynchronous system model have been proposed in the literature [19] These algorithms can be classified as being either symmetric or asymmetric: in a symmetric algorithm all processes perform the same code [6, 25, 10, 4], whereas in an asymmetric algorithm one process plays a special role, i.e. defines the ordering of messages [23, 7] Asymmetric algorithms require less phases and are thus more efficient, but are subject to the contamination problem [16, 35] Token based algorithms [11, 3] can be classified ....

S. W. Luan and V. D. Gligor. A Fault-Tolerant Protocol for Atomic Broadcast. IEEE Trans. Parallel & Distributed Syst., 1(3):271--285, July 90.

No context found.

Luan, S. and Gligor, V.D. : "A Fault Tolerant Protocol for Atomic Broadcast", IEEE Trans. on Parallel and Distributed Systems, vol. 1, No. 3, pp. 271 -- 285 (1990).

No context found.

Luan, S. and Gligor, V.D. : A Fault Tolerant Protocol for Atomic Broadcast, IEEE Trans. on Parallel and Distributed Systems, Vol. 1, No. 3, pp. 271 -- 285 (1990).

No context found.

Luan, S. and Gligor, V.D. : A Fault Tolerant Protocol for Atomic Broadcast, IEEE Trans. on Parallel and Distributed Systems, Vol. 1, No. 3, pp. 271 -- 285 (1990).

No context found.

LUAN, S.-W. AND GLIGOR, V. D. 1990. A fault-tolerant protocol for atomic broadcast. IEEE Trans. Parall. Distrib. Syst. 1, 3 (July), 271--285.

No context found.

S. Luan and V. D. Gligor, `A fault-tolerant protocol for atomic broadcast', IEEE Trans. on Parallel and Distributed Syst., 1(3), 271--285 (July 1990).

No context found.

S. W. Luan and V. D. Gligor. A Fault-Tolerant Protocol for Atomic Broadcast. IEEE Trans. Parallel & Distributed Syst., 1#3#:271#285, July 90.

No context found.

Luan, S. W. and Gligor, V. D., "A Fault-Tolerant Protocol for Atomic Broadcast," IEEE Trans. on Parallel and Distributed Systems , Vol.1, No.3, 1990, pp.271-285.

No context found.

Luan, S. W. and Gligor, V. D., "A Fault-Tolerant Protocol for Atomic Broadcast," IEEE Trans. Parallel and Distributed Systems, Vol.1, No.3, 1990, pp.271--285.

No context found.

) Luan, S. W. and Gligor, V. D.: A FaultTolerant Protocol for Atomic Broadcast, IEEE Trans.Parallel and Distributed Systems, Vol.1, No. 3, pp. 271--285 (1990)

No context found.

Luan, S. W. and Gligor, V. D., "A Fault-Tolerant Protocol for Atomic Broadcast," IEEE Trans. on Parallel and Distributed Systems, Vol.1, No.3, 1990, pp.271-285.

No context found.

Luan, S. W. and Gligor, V. D., "A Fault-Tolerant Protocol for Atomic Broadcast," IEEE Trans. on Parallel and Distributed Systems, Vol.1, No.3, 1990, pp.271--285.

No context found.

Luan, S. W. and Gligor, V. D., "A Fault-Tolerant Protocol for Atomic Broadcast," IEEE Trans. on Parallel and Distributed Systems, Vol.1, No.3, 1990, pp.271-285.

No context found.

Luan, S. W. and Gligor, V. D., "A Fault-Tolerant Protocol for Atomic Broadcast," IEEE Trans. on Parallel and Distributed Systems, Vol.1, No.3, 1990, pp.271285.

No context found.

Luan, S. W. and Gligor, V. D., "A fault-tolerant protocol for atomic broadcast," IEEE Transactions on Parallel and Distributed Systems 1, 3 (July 1990), 271-285.

No context found.

S. W. Luan and V. D. Gligor. "A Fault Tolerant Protocol For Atomic Broadcast." IEEE Transactions on Parallel and Distributed Systems. 1, 3 (July 1990). pp. 271-285.

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