F. Cristian and C. Fetzer, "The timed asynchronous system model," in Proceedings of the 28th Annual International Symposium on Fault-Tolerant Computing, Munich, Germany, June 1998, pp. 140--149.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....in the system exchange messages over reliable and FIFO communication links. For each link connecting two server processes, it does exist a known time constant # such that if processes p i and p j are not crashed, then a message sent from p i to p j at time t will be received by p j within t # [7]. This is a reasonable assumption if server processes are connected over a local area network (see Section 4) Concerning failures, we assume the DNS does not crash and other entities follow a crash failure behavior. In particular, we assume that if an entity (i.e. server process, failure ....

....physical address is returned. For simplicity and without lost of generality, we will no longer consider the distributed nature of the DNS database in the following. Bounded drifts with respect to real time are allowed in practical models. e.g. the Cristian Fetzer Timed Asynchronous Model [7]. Failure Detection. To pursue the aim of implementing highly available servers, a server process needs to detect the failure of another one. In the past it has been argued that this detection should be decoupled by its client processes and implemented as a middleware service [5, 11, 15] A ....

[Article contains additional citation context not shown here]

Cristian F., Fetzer C, "The Timed Asynchronous System Model", In Proc. of the 28th Annual International Symposium on Fault-Tolerant Computing, Munich, Germany, 1998.

....believe to be a reasonable assumption for a generic framework. In short, NAVTECH has a model of partial or uncertain synchrony, that can vary anywhere in the space between synchronous and asynchronous. The specific models that best reflect NAVTECH s synchrony assumptions are the timed asynchronous[12] and the quasi synchronous[46] models, or the recent timely computing base model[41] 3.1 Topological Model NAVTECH was designed for large scale applications. Topology issues and communication system characteristics have a fundamental impact on the achievable scale of computations, 4 as systems ....

F. Cristian and C. Fetzer. The timed-asynchronous system model. Technical Report CS97-517, UCSD, January 1997.

....the cost of short term inconsistency can be very high in mobile computation scenarios. Mobile hosts interact over wide spaces, and inconsistency can propagate indefinitely and cause irreparable damage in mission critical applications. Similar strict consistency has been considered by Cristian [12] for the primary group membership problem in timed asynchronous systems, but has not been investigated for the partitionable group membership problem. The goal of our group membership service is not only to create a consistent view of group membership among participants, but also to help users ....

F. Cristian and C. Fetzer. The timed asynchronous system model. Technical Report CS97519, UCSD, January 1997.

....the Protocol for Asymmetric Duplex Redundancy (PADRE) of [5] starting by highlighting some of its system assumptions. 4.1 Timeliness assumptions PADRE has been conceived assuming a system that satisfies several timeliness properties. These properties are those of the Timed Asynchronous model [4]: 1) the drift between process clocks is bounded and the bound known, 2) messages sent over the network have performance omission failure semantics, 3) there is a known bound for message handling by processes: processes either handle received messages within this interval or halt. Based on ....

....by PADRE since for all configuration changes we have v = 1. Indeed, in this case we have only one process in v i v i 1 : violation of view synchrony can only occur with a least two processes in v i 1 . The same view delivery property is ensured in PADRE by fail aware datagrams [4]. PADRE computes an upper 21 A knows that B is in quarantine, i.e. that B will never install Figure 12. Unit A delays the installation of v 2k 1 to avoid view divergence bound on the real transmission delay of each message, and classifies messages as fast and slow. A fast message is ....

[Article contains additional citation context not shown here]

F. Cristian and C. Fetzer. The Timed Asynchronous System Model. Technical Report CSE97-519, Department of Computer Science, UCSD, 1997.

....partial synchronous models. Research in distributed systems has traditionally been divided between two canonical models: fully synchronous and fully asynchronous. Partial synchronous models try to give the best of both worlds, allowing timeliness specifications but accepting that they can fail [22, 6]. The TTCB provides a set of time services whose main objective is precisely to detect the failure of timeliness specifications. This is only possible because the TTCB is timely, i.e. the TTCB is a real time (synchronous) component. Many of the time related ideas and services of the TTCB were ....

F. Cristian and C. Fetzer. The timed asynchronous system model. In Proc. the 28th IEEE International Symposium on Fault-Tolerant Computing, pages 140--149, 1998.

....from soft to hard real time. These are challenging problems in largescale systems with uncertain synchrony, especially where wireless communication is employed. We intend to use and build on previous results on partial synchrony systems, such as the timed asynchronous and quasi synchronous models [11,37,40]. 2.4. Scalability Scalability represents a crucial transparency property concerning the ability to accommodate growth in a large scale distributed system. Thus, connecting more participants to the system dynamically, including adding entire additional networks, or providing new services, should ....

F. Cristian and C. Fetzer. The Timed Asynchronous System Model. Proceedings of the 28th Annual International Symposium on Fault-Tolerant Computing. pp.140-149. Munich, Germany. June 1998.

....properties determining the possibility or impossibility of solving various distributed system problems like consensus, atomic broadcast or leader election. The timed asynchronous model adds some synchronism to the system by assuming the existence of local hardware clocks with bounded drift rate [7]. This allows processes to measure the passage of time and use timeouts. In timed systems it is possible to construct a special failure detector, a fail aware failure detector [8] which can be implemented if some additional progress assumptions are made. In [10] Fetzer proposes an approach to ....

Flaviu Cristian and Christof Fetzer. The timed asynchronous system model. In Digest of Papers, The 28th Annual International Symposium on Fault-Tolerant Computing, pages 140--149, Munich, Germany, June 1998. IEEE Computer Society Press.

.... commit protocol when the system is stable and consensus has to be reached on a sequence of values (a three phase protocol is needed only for the #rst consensus problem, while all the subsequent ones can be solved with a two phase exchange of messages) Cristian s timed asynchronous model [4] is similar to the distributed setting considered in this paper. It assumes, however, a bounded clock drift even when the system is unstable. Our model is weaker in the sense that makes no assumption on clock drift when the system is unstable. The Clock GTA provides a formal way of modelling the ....

F. Cristian, C. Fetzer, The timed asynchronous system model, Dept. of Computer Science, UCSD, La Jolla, CA. Tech. Rep. CSE97-519.

....the Protocol for Asymmetric Duplex Redundancy (PADRE) of [5] starting by highlighting some of its system assumptions. 4.1 Timeliness assumptions PADRE has been conceived assuming a system that satisfies several timeliness properties. These properties are those of the Timed Asynchronous model [4]: 1) the drift between process clocks is bounded and the bound known, 2) messages sent over the network have performance omission failure semantics, 3) there is a known bound for message handling by processes: processes either handle received messages within this interval or halt. Based on ....

....PADRE since for all configuration changes we have jv i v i 1 j = 1. Indeed, in this case we have only one process in v i v i 1 : violation of view synchrony can only occur with a least two processes in v i v i 1 . The same view delivery property is ensured in PADRE by fail aware datagrams [4]. PADRE computes an upper v 2k = fA i ; B j g A suspects B A knows that B is in quarantine, i.e. that B will never install view fB j g Figure 12. Unit A delays the installation of v 2k 1 to avoid view divergence bound on the real transmission delay of each message, and classifies ....

[Article contains additional citation context not shown here]

F. Cristian and C. Fetzer. The Timed Asynchronous System Model. Technical Report CSE97-519, Department of Computer Science, UCSD, 1997.

....that determine the possibility or impossibility of solving various distributed system problems like consensus, atomic broadcast or leader election. The timed asynchronous model adds some synchronism to the system by assuming the existence of local hardware clocks with bounded drift rate [7]. This allows processes to measure the passage of time and use timeouts. In timed systems it is possible to construct a special failure detector, a fail aware failure detector [8] which can be implemented if some additional progress assumptions are made. In [10] Fetzer proposes an ap proach to ....

Flaviu Cristian and Christof Fetzer. The timed asynchronous system model. In Digest of Papers, The 28th Annual International Symposium on Fault-Tolerant Computing, pages 140-- 149, Munich, Germany, June 1998. IEEE Computer Society Press.

....as the bound for message delay. In [12] the bound is not known or the point in time when it starts to hold is not known. More recently there has been another work that has some similarities to our work from the point of view of goals [15] That work is based on a model called Timed Asynchronous [10]. That system model assumes that processes have access to local clocks that although not synchronized have a bounded drift rate. The fault model assumes that processes may crash or experience performance failures and messages can have omissions or performance failures. On top of this model a ....

F. Cristian and C. Fetzer. The timed asynchronous system model. Technical Report CSE97-519, Dept. of Computer Science, UCSD, La Jolla, CA 9

.... ensures time domain correctness of applications in environments of uncertain synchronism[14] The timely computing base model addresses a broader spectrum of problems than those solved by previous timed partially synchronous models, such as the quasi synchronous[12] and the timedasynchronous models[4]. All these works share the same observation: synchronism or asynchronism are not homogeneous properties of systems. That is, they 5 vary with time, and they vary with the part of the system being considered. In the timely computing base model, it is assumed that systems have an architecture such ....

F. Cristian and C. Fetzer. The timed asynchronous system model. In Proceedings of the 28th Annual International Symposium on Fault-Tolerant Computing, pages 140-149, Munich, Germany, June 1998. IEEE Computer Society Press.

.... Toueg have given a failure detector which, should the system be synchronous for a long enough period, would allow to terminate consensus[1] Cristian Fetzer have devised the timed asynchronous model, where the system has just enough synchronism to make decisions such as fail safe shutdown [2]. We have devised the quasi synchronous model where parts of the system have enough synchronism to perform real time actions with a certain probability[3] 3. The Timely Computing Base The architecture of a system with a Timely Computing Base is suggested by Figure 1. Whilst there is a ....

Flaviu Cristian and Christof Fetzer. The timed asynchronous system model. In Proceedings of the 28th Annual International Symposium on Fault-Tolerant Computing, pages 140-- 149, Munich, Germany, June 1998. IEEE Computer Society Press.

.... ensures time domain correctness of applications in environments of uncertain synchronism[14] The timely computing base model addresses a broader spectrum of problems than those solved by previous timed partially synchronous models, such as the quasi synchronous[12] and the timedasynchronous models[4]. All these works share the same observation: synchronism or asynchronism are not homogeneous properties of systems.Thatis,they 5 vary with time, and they vary with the part of the system being considered. In the timely computing base model, it is assumed that systems have an architecture such ....

F. Cristian and C. Fetzer. The timed asynchronous system model. In Proceedings of the 28th Annual International Symposium on Fault-Tolerant Computing, pages 140--149, Munich, Germany, June 1998. IEEE Computer Society Press.

.... in environments of uncertain synchronism [Verissimo et al. 2000] The timely computing base model addresses a broader spectrum of problems than those solved by previous timed partially synchronous models, such as the quasi synchronous [Verssimo Almeida 1995] and the timed asynchronous models [Cristian Fetzer 1998]. All these works share the same observation: synchronism or asynchronism are not homogeneous properties of systems. That is, they vary with time, and they vary with the part of the system being considered. However, each model has treated these asymmetries in its own way: some relied on the ....

F. Cristian and C. Fetzer, "The Timed Asynchronous System Model", in 28th IEEE Int. Symp. on Fault-Tolerant Computing (FTCS-28), (Munich, Germany), pp.140-9, IEEE Computer Society Press, 1998.

....model for Group Membership in networks subject to partitions and message loss. I will return to this open problem in the Proposal section of this report. Another model for networks subject to partitions is called the timed asynchronous model. It was first mentioned in [CS95] and updated in [CF97] The model is based on experiments with distributed systems consisting of networked workstations in a laboratory setting. Messages can be lost or experience unbounded communication delay but most are delivered within a known time bound. Processes may crash or experience unbounded delays between ....

Flaviu Cristian and Christof Fetzer. The timed asynchronous system model. Technical report, Department of Computer Science, University of California, San Diego, 1997.

.... able to terminate[13] Cristian Fetzer have devised the timed asynchronous model, where the system alternates between synchronous and asynchronous behavior, and where parts of the system have just enough synchronism to make decisions such as detection of timing failures or fail safe shutdown [15]. We have devised the quasi synchronous model where parts of the system have enough synchronism to perform real time actions with a certain probability[36] These works share a same observation: synchronism or asynchronism are not homogeneous properties of systems. That is, they vary with time, ....

Flaviu Cristian and Christof Fetzer. The timed asynchronous system model. In Proceedings of the 28th Annual International Symposium on Fault-Tolerant Computing, pages 140--149, Munich, Germany, June 1998. IEEE Computer Society Press.

....the device to normal operation, and remove the device from its list of failed devices. Synchrony model. We have assumed a synchronous system model to simplify the analysis of correctness. We are investigating whether the system is correct under a more relaxed model, such as timed asynchrony [9]. The only difference between our current assumptions and this model is it permits performance failures in the network, resulting in late messages. Failure suspectors. The layout control protocol we have presented uses leases as failure suspectors, but leases only work in systems with limited ....

F. Cristian and C. Fetzer. The timed asynchronous system model. Technical Report CSE97--519, Computer Science Dept., Univ. of California at San Diego, 1997.

....which can communicate with a majority of time servers in a timely manner and which is timely (i.e. it does not suffer performance failures) has to set its synchronization indicator to true. We show how fail aware clock synchronization can efficiently be implemented in timed asynchronous systems [3]. These systems are characterized by a precise notion of what it means that a process or message to be timely, unbounded message transmission and process scheduling delays, and each process has access to a local hardware clock with a drift rate within some given bounds. The proposed protocol ....

F. Cristian and C. Fetzer. The timed asynchronous system model. Technical Report CS97-519, UCSD, Jan 1997.

....better results, essentially for three reasons: i) they allow timeliness speci cations; ii) they admit failure of those speci cations; iii) they provide timing failure detection. Previous timed partially synchronous models, such as the quasi synchronous [99] and the timed asynchronous models [29], share the same observation: synchronism or asynchronism are not homogeneous properties of systems. That is, they vary with time, and they vary with the part of the system being considered. However, each model has treated these asymmetries in its own way: some relied on the evolution of ....

F. Christian and C. Fetzer. The timed asynchronous system model. In Proceedings of the 28th IEEE International Symposium on Fault-Tolerant Computing (FTCS-28), pages 140-149, Munich, Germany, 1998. IEEE Computer Society Press.

No context found.

F. Cristian and C. Fetzer, "The timed asynchronous system model," in Proceedings of the 28th Annual International Symposium on Fault-Tolerant Computing, Munich, Germany, June 1998, pp. 140--149.

No context found.

F. Cristian and C. Fetzer. The timed asynchronous system model. Technical Report CS97-519, University of California, San Diego, January 1997.

No context found.

F. Cristian and C. Fetzer. The timed asynchronous system model. Technical Report CS-97519, University of california - San Diego, January 1997.

No context found.

6 Christian, F., Fetzer, C.: The timed asynchronous system model. In: Proceedings of the 28th IEEE International Symposium on Fault-Tolerant Computing. (1998) 140-149

No context found.

F. Cristian and C. Fetzer. The Timed Asynchronous System Model. Technical Report CSE97-519, Department of Computer Science, UCSD, 1997.

First 50 documents

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