Lundelius, J., Lynch, N.: A new fault-tolerant algorithm for clock synchronization. In: Proceedings of the Third annual ACM Symposium on Principles of Distributed Computing, Vancouver, Canada (1984) 75--88  Home/Search   Document Details and Download   Summary   Related Articles   Check

.... phrases like . algorithm xx requires synchronized clocks with precision . are commonly used in modern textbooks on very different subjects, like communication protocols and data base systems. The clock synchronization algorithms of Halpern Simons Strong Dolev [3] Lundelius Lynch [5], Lamport Melliar Smith [4] Mahaney Schneider [6] Cristian Aghili Strong [1] Srikant Toueg [7] and the many variants derived from them) but also related theoretical work form a well established basis for applications that require synchronized clocks. So why wasting time and money by ....

J. Lundelius, N. Lynch. A New Fault-Tolerant Algorithm for Clock Synchronization, Proc. 3rd ACM Symposium on Principles of Distributed Computing, Vancouver, Canada, August 1984, p. 75--88.

....is then specialized to the class of algorithms using averaging functions, yielding a theory that corresponds to those of Shankar and Miner. As examples of the verification of concrete, published algorithms, the formal verification of an instance of an averaging algorithms (by Welch and Lynch [1]) and of a non averaging algorithm (by Srikant and Toueg [9] is exhibited. 1 Introduction Clock synchronization is one of the central elements of distributed dependable real time systems. Many mechanisms for realizing dependability properties in distributed real time systems rely on the fact ....

....a different algorithm and do not rely on a convergence function. In this paper we report on the formal analysis of a broad class of clock synchronization algorithms. Starting from formalizations of the non averaging algorithm of Srikant and Toueg [9] and the averaging algorithm of Welch and Lynch [1], several similarities between both averaging and non averaging algorithms became apparent. Concepts like interval clocks, rounds and faulty processing nodes are common to both types of algorithms and are needed in either formalization, whereas other concepts like triggering events or convergence ....

[Article contains additional citation context not shown here]

J. Lundelius Welch and N. Lynch. A new fault-tolerant algorithm for clock synchronization. Inf. and Comp., 77(1):1--36, 1988.

....of the primitive of Figure 2, it is easy to see that the proofs of Lemmas 1 10 and Theorem I hold. Also, by Theorem 5, correct processes send O(n ) messages for each resynchronization round. Thus the number of messages sent by correct processes for each resynchronization is similar to that in [10]. In Section 3.3 we showed how the authenticated algorithm could be modified to achieve optimal accuracy. Translating this modified algorithm with our broadcast primitive results in a nonauthenticated algorithm that achieves optimal accuracy. broadcast (round 0) using the primitive in Figure ....

....now describe how a process joins a system of synchronized clocks. This could be used by new processes to enter the system, or by processes that have become unsynchronized (possibly due to failures) to reestablish synchronization with the rest of the system. The algorithms are based on the idea in [10], modified to the context of our algorithms. When a process p wishes to join the system, it sends a message (joining) to the processes already in the system. It then receives messages from these processes and determines the number i of the round being executed. Since p could have started this ....

[Article contains additional citation context not shown here]

LUNDELIUS, J., AND LYNCH, N. A new fault-tolerant algorithm for clock synchronization. In Proceedings of the 3rd Annual ACM Symposium on Principles of Distributed Computing (Vancouver, Canada, Aug.). ACM, New York, 1984, pp. 75-88.

....different rates) The restart problem is to reestablish synchronization after transient faults have afflicted one or more (or all) controllers. The synchronization problem is well understood and many algorithms to solve it have been developed, analyzed, and formally verified [LMS85, Min93, Sha92, WL88] The algorithm employed in TTA belongs to the general class of averaging synchronization algorithms [Sch87] each controller i estimates its skew relative to each controller p by comparing the reading of its local clock at the instant when the message in slot p arrives against s i (p) adjusted ....

J. Lundelius Welch and N. Lynch. A new fault-tolerant algorithm for clock synchronization. Information and Computation, 77(1):1--36, April 1988. 25

....system is explained. 1. Introduction It is frequently necessary for processors to agree on an approximation to some value. In certain applications, processors can use different values provided these values are close to each other and to the value being approximated. Clock synchronization [HSSD84] [LL84] [LM84] MO83] is one application where such Inexact Agreement is acceptable. Other applications arise in industrial control where sensors measure some physical quantity and, under program control, Room 2C 323, AT T Bell Laboratories, Murray Hill, N.J. 07974 Department of Computer Science, ....

....to implement fault tolerant protocols for synchronizing clocks. The approach we take here is based on the interactive convergence algorithm of [LM84] There, processor clocks are It is also possible to replace the Approximate Agreement protocol used in the clock resynchronization protocol of [LL84] with FCA or CCA. The protocols of [HSSD84] and [MO83] however, cannot benefit from using FCA or CCA because they are not based periodically brought together using an Inexact Agreement protocol. FCA has better precision than the Inexact Agreement protocol used in [LM84] so simply replacing ....

[Article contains additional citation context not shown here]

Lundelius, J. and N. Lynch. A new faulttolerant algorithm for clock synchronization. Symposium on Principles of Distributed Computing, August 1984, ACM, 75-88.

....us to identify the different imple mentation choices made by each protocol in solving three subproblems it defines. This permits the e.g. Babaoglu Dnunmond 87] Cristian et al. 86] Halpern eta . 84] Kopetz Ochsenreiter 87] Lamport Melliar Smith 84] Lampoft Melliar Smith 85] Lundelius Lynch 84] Mahaney Schneider 85] and [Srikanth Toueg 85] 1. various Byzantine clock synchronization protocols 2 to be compared and the contributions of each to be isolated. Previously, clock synchronization algorithms were viewed in terms of three distinct classes: those based on convergence, ....

....delays due to queuing in sites doing relaying can be measured and recorded in the message and therefore can be accounted for. A variation on the clock reading scheme just given, used in the clock synchronization protocols of [Babaoglu Drummond 87] Cristian et al. 86] Halpem et al. 84] Lundelius Lynch 84] and [Srikanth Toueg 85] reduces the number of messages by half but can increase clock reading error. 8 Instead of requesting the time, each processor q periodically broadcasts its virtual clock value (includ ing superscript 0. Upon receipt of such a message, the receiverp updates x[q] as ....

[Article contains additional citation context not shown here]

Lundelius, J. and N. Lynch. A new fault-tolerant algorithm for clock synchronization. Proc. of the Third ACM SIGACT-SIGOPS Symposium on Principles of Distributed Computing, Vancouver, Canada, August 1984, 75-88.

.... considered some plausible modifications to 1;4 (S) but the solutions that he found operate on point, rather than interval, sensor readings, and therefore do not make full use of the available information (an example is the fault tolerant midpoint used in some clock synchronization algorithms [WL88]) An interval based fusion method that does satisfy the Lipschitz Condition has recently been introduced by Schmid and Schossmaier [SS01] Their Fault Tolerant Interval (FTI) fusion function F n (S) is defined as follows. Definition 3 (FTI fusion interval) Let l = max f 1 fS(j) lo g be the f ....

J. Lundelius Welch and N. Lynch. A new fault-tolerant algorithm for clock synchronization. Information and Computation, 77(1):1--36, April 1988. 19

....clocks to drift apart so periodically (several hundred times a second) they must be resynchronized. What makes this difficult is that some of the clocks may be faulty. The clock synchronization algorithm used in TTA is a modification of the WelchLynch (also known as Lundelius Lynch) algorithm [78], which itself can be understood as a particular case of the abstract algorithm described by Schneider [66] Schneider s abstract algorithm operates as follows: periodically, the nodes decide that it is time to resynchronize their clocks, each node determines the skews between its own clock and ....

J. Lundelius Welch and N. Lynch. A new fault-tolerant algorithm for clock synchronization. Information and Computation, 77(1):1--36, April 1988. 3

....waves of echo or accept events may be needed to make this fault tolerant. Schneider [Sch87] gives a general description that applies to all averaging clock synchronization algorithms; these algorithms differ only in their choice of fault tolerant average. The Welch Lynch algorithm [WL88] is a popular choice that is characterized by use of the fault tolerant midpoint as its averaging function. We assume n clocks and the maximum number of simultaneous faults to be tolerated is t (3t n) the fault tolerant midpoint is the average of the t 1 st and n t th clock reading, when ....

J. Lundelius Welch and N. Lynch. A new fault-tolerant algorithm for clock synchronization. Information and Computation, 77(1):1--36, April 1988.

....average that is largely insensitive to a certain number of readings from faulty clocks. Schneider [18] gives a general description that applies to all averaging clock synchronization algorithms; these algorithms differ only in their choice of fault tolerant average. The Welch Lynch algorithm [25] is a popular choice that is characterized by use of the fault tolerant midpoint as its averaging function. Event based algorithms rely on nodes being able to sense events directly on the interconnect: each node broadcasts a ready event when it is time to synchronize and sets its clock when it ....

....providing clock synchronization, and message sequencing and transmission functions. The interconnect bus is duplicated and each controller drives both of them through partially independent bus guardians. TTA uses an averaging clock synchronization algorithm based on that of Lundelius and Lynch [25]. This algorithm is implemented in the controllers, but requires too many resources to be replicated in the bus guardians. The guardians, which have independent clocks, therefore rely on their controllers for a start of frame signal. This compromises their independence somewhat (they also share ....

J. Lundelius Welch and N. Lynch. A new fault-tolerant algorithm for clock synchronization. Information and Computation, 77(1):1--36, April 1988.

....modeling decisions and a stimulus for the development of algorithm verification techniques. Similar results should be possible for real time systems. Some examples of complexity results that have already been obtained for real time systems are the many results on clock synchronization, including [8, 11, 17, 20, 32] (see [31] for a survey) In this paper, we embark on a study of complexity results for real time systems. We begin this study by considering timing based variations of certain problems that have previously been studied in asynchronous concurrent systems. In particular, we study a variant of the ....

....that its worst case response time is at most n. m (d 2) plus some terms involving c 2 and l) A corresponding lower bound can also be proved. A formal treatment of these results requires several changes to our model, and we prefer not to present it here. The clock synchronization algorithm of [20] yields synchronization points that can be used by a distributed allocation algorithm whose response time is at most n .m (n 1 )d. Since the lower bound of [20] implies that this clock synchronization algorithm is optimal, it does not appear that a naive use of clock synchronization produces ....

[Article contains additional citation context not shown here]

LUNDELIUS, J., AND LYNCH, N. {1988), A new fault-tolerant algorithm for clock synchronization, Inform. and Cornput. 77, 1-36.

....round r until it has received a round r Gamma 1 message from every nonfaulty process. The partial synchronization works by using a combination of two criteria for advancing to further phases, one based on elapsed local time and the other based on messages received. A similar technique is used in [WL88] to initiate new rounds of clock resynchronization. In particular, our criteria for ending round 1 is essentially the same as the criteria used in [WL88] for ending every round; our criteria for subsequent rounds is different. 4.1 The simulation algorithm The algorithm simulates a synchronous ....

....for advancing to further phases, one based on elapsed local time and the other based on messages received. A similar technique is used in [WL88] to initiate new rounds of clock resynchronization. In particular, our criteria for ending round 1 is essentially the same as the criteria used in [WL88] for ending every round; our criteria for subsequent rounds is different. 4.1 The simulation algorithm The algorithm simulates a synchronous algorithm by ensuring that each nonfaulty process receives all round r messages of the synchronous algorithm from all other nonfaulty processes before ....

J. L. Welch and N. Lynch. A new fault-tolerant algorithm for clock synchronization. Information and Computation, Vol. 77, No. 1 (April 1988), pp. 1--36. 67

No context found.

Lundelius, J., Lynch, N.: A new fault-tolerant algorithm for clock synchronization. In: Proceedings of the Third annual ACM Symposium on Principles of Distributed Computing, Vancouver, Canada (1984) 75--88

No context found.

Lundelius, J., and Lynch, N., A new fault-tolerant algorithm for clock synchronization. In Proc. of the 3d Annual ACM Symposium on Principles of Distributed Computing. 1984.

No context found.

J. Lundelius and N. Lynch. A new fault-tolerant algorithm for clock synchronization. In Proceedings of the 3rd Annual ACM Symposium on Principles of Distributed Computing, pages 75--88, Vancouver, Canada, August 1984.

No context found.

J. L. Welch, and N. Lynch, " A New Fault-Tolerant Algorithm for Clock Synchronization ", Information and Computation 77, 1-36, 1988.

No context found.

J. Lundelius and N. Lynch. A new fault-tolerant algorithm for clock synchronization. In ACM Symp. on Principles of Distributed Computing, pages 75--88, 1984.

No context found.

J. L. Welch, and N. Lynch, A New Fault-Tolerant Algorithm for Clock Synchronization , Information and Computation 77, 1-36, 1988.

No context found.

Lundelius, J., and N. Lynch, "A New Fault-Tolerant Algorithm for Clock Synchronization:, Proc. Third Annual ACM Symposium on Principles of Distributed Computing, pgs. 75-88, August 1984.

No context found.

J. Welch and N. Lynch. A new fault-tolerant algorithm for clock synchronization. Information and Computation, 77(1):1--36, 1988.

No context found.

J. L. Welch and N. Lynch. A new fault-tolerant algorithm for clock synchronization. Information and Computation, 77(1), pages 1--36, 1988.

No context found.

J. L. Welch and N. Lynch, `A new fault tolerant algorithm for clock synchronization', Information and Computation, 77, (1), 1--36 (1988).

No context found.

J. Lundelius and N. Lyunch, "A new fault-tolerant algorithm for clock synchronization," Information and Computation, vol. 77, no. 1, pp. 1--36, 1988.

No context found.

Lundelius, J. and N. Lynch. A new fault-tolerant algorithm for clock synchronization. Proc. of the Third ACM SIGACT-SIGOPS Symposium on Principles of Distributed Computing, Vancouver, Canada, August 1984, 75-88.

No context found.

J. Lundelius-Welch and N. Lynch. A new fault-tolerant algorithm for clock synchronization. Inform. and Comput., 77(1):1--36, April 1988.

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