Verissimo, P., Rodrigues, L., Casimiro, A.: CesiumSpray: A Precise and Accurate Global Time Service for Large-Scale Systems. Journal of Real-Time Systems 12 (1997) 243--294  Home/Search   Document Not in Database   Summary   Related Articles   Check

....broadcasts. Moreover, it assumes the presence of synchronized global clocks, such as GPS, at many points in the network, and so the main focus is reducing the variance along the paths to these time oracles. There are several proposals for synchronizing clocks within a single broadcast domain [18, 17, 13]. These all exploit the special properties of broadcast media and achieve high precision. However, they cannot synchronize nodes that do not lie within the same broadcast domain. Since our focus here is on global clock synchronization, we don t discuss these more local approaches further. Two ....

VERISSIMO, P., RODRIGUES, L., AND CASIMIRO, A. Cesiumspray: a precise and accurate global time service for large-scale systems. Technical Report NAV-TR-97-0001, Universidade de Lisboa, 1997.

....various types of failures, self configuration, security in the face of deliberate sabotage, and ubiquitous deployment. NTP allows construction of a hierarchy of time servers, multiply rooted at canonical sources of external time. Perhaps most closely related to our work is the CesiumSpray system [28], the foundations of which were described by Verssimo and Rodrigues in [27] Their system, like ours, makes use of the inherent properties of broadcast media to achieve superior synchronization. In a certain basic scenarios, RBS and CesiumSpray are essentially equivalent specifically, if the ....

....may be unknown, and highly variable from message to message, the nature of a broadcast dictates that for a particular message, these quantities are the same for all receivers. This observation was also made by Verssimo and Rodrigues in [27] and later became central to their CesiumSpray system [28]. The fundamental property of Broadcast Reference Synchronization is that a broadcast message is only used to synchronize a set of receivers with one another, in contrast with traditional protocols that synchronize the sender of a message with its receiver. Doing so removes the Send Time and ....

P. Verissimo, L. Rodrigues, and A. Casimiro. Cesiumspray: a precise and accurate global time service for large-scale systems. Technical Report NAV-TR-97-0001, Universidade de Lisboa, 1997.

....[12] One example that provides a more system independent interface is the Spearmints system [13] Furthermore, most hardware solutions are designed 12 to work in LAN type environments. A notable exception is the CesiumSpray system that uses GPS receivers to synchronize machines, even across WANs [14]. Systems also exist that use both hardware and software to synchronize globally timed measurements [15] Using some hardware can help to reduce the intrusion and improve accuracy of some software algorithms. Many software algorithms are limited by the variance in message transmission times. ....

P. Verissimo, L. Rodrigues, and A. Casimiro, "CesiumSpray: A precise and accurate global time service for large-scale systems," Real-Time Systems, vol. 12, pp. 243-294, May 1997.

....They provide a known framework in dependable system design vis vis accidental failures. Generally, they consist of allocating different assumptions to different subsets or components of the system, and have been used in a number of systems and protocols [Meyer Pradhan 1987, Powell et al. 1988, Verissimo et al. 1997] 5.1.2 Composite fault model With hybrid assumptions some parts of the system would be justifiably assumed to exhibit fail controlled behaviour, whilst the remainder of the system would still be allowed an arbitrary behaviour. This would be advantageous in modular and distributed system ....

P. Verissimo, L. Rodrigues and A. Casimiro, "Cesiumspray: a Precise and Accurate Global Time Service for Large-Scale Systems", Journal of Real-Time Systems, 12 (3), pp.243-94, 1997. 84

....with the user requirements of the vast majority of today s on line applications. Hybrid assumptions combining both kinds of failure assumptions would be desirable. They are a known framework in dependable system design vis vis accidental failures [Meyer Pradhan 1987] Powell et al. 1988] [Verissimo et al. 1997]. Generally, they consist of allocating different assumptions to different subsets or components of the system, and have been used in a number of systems and protocols. Hybrid models allow stronger assumptions to be made about parts of the system that can justifiably be assumed to exhibit ....

P. Verissimo, L. Rodrigues and A. Casimiro, "Cesiumspray: a Precise and Accurate Global Time Service for Large-Scale Systems", Journal of Real-Time Systems, 12 (3), pp.243-94, 1997.

....of the system. 3.2.6 Clock Synchronization Almost all services provided byAtomic Group rely on the presence of synchronized clocks. The clocks can be either dedicated hardware clocks (radio controlled or satellite controlled clocks) or internal hardware clocks that are synchronized by software [54, 79]. Logical (Lamport) clocks suce for correct operation of the Atomic Group protocols, but the latency of the protocol will be signi cantly worse than the latency of a protocol based on physical clocks. Physical clocks cannot be perfectly synchronized if the communication medium that connects the ....

P. Verissimo, L. Rodrigues and A. Casimiro, \Cesiumspray: A precise and accurate global time service for large-scale systems," Real-Time Systems, vol. 12, no. 3 (May 1997), pp. 243-294.

....of the system. 3.2.6 Clock Synchronization Almost all services provided by Atomic Group rely on the presence of synchronized clocks. The clocks can be either dedicated hardware clocks (radio controlled or satellite controlled clocks) or internal hardware clocks that are synchronized by software [54, 79]. Logical (Lamport) clocks suce for correct operation of the Atomic Group protocols, but the latency of the protocol will be signi cantly worse than the latency of a protocol based on physical clocks. Physical clocks cannot be perfectly synchronized if the communication medium that connects the ....

P. Verissimo, L. Rodrigues and A. Casimiro, \Cesiumspray: A precise and accurate global time service for large-scale systems," Real-Time Systems, vol. 12, no. 3 (May 1997), pp. 243-294.

....either by hardware or by software. Hardware clock synchronization [SR87, KO87] achieves very tight synchronization through the use of special synchronization hardware at each processor, and uses a separate network solely for clock signals. In contrast, software clock synchronization algorithms [CAS86, HSSD84, ST87, LL88, LMS85, VCR97, PB95, GZ89, MS85, FC95a, FC97, OS94, Arv94, Cri89] use standard communication networks and send synchronization messages to get the clocks synchronized. They do not require specific hardware, but do not provide synchronization as tight as hardware algorithms. Software clock synchronization algorithms decompose themselves in deterministic, ....

.... with algorithms aimed at maintaining logical clocks that reflect properties on the execution of distributed applications, such as causal relationships [Lam78] In addition, for space considerations, we do not detail issues specific to clock synchronization in large scale distributed systems (e.g. [SR87, VCR97]) We classify clock synchronization algorithms according to their internal structure (symmetry of the algorithm) and to three orthogonal and independent basic building blocks we Irisa A Taxonomy of Clock Synchronization Algorithms 5 have identified. The first building block, called the ....

[Article contains additional citation context not shown here]

P. Verissimo, A. Casimiro, and L. Rodrigues. Cesiumspray: a precise and accurate global time service for large scale systems. Journal of Real-Time Systems, 12:243--294, 1997. PI n1103

....may be achieved either by hardware or by software. Hardware clock synchronization [SR87, KO87] achieves very tight synchronization, through the use of special synchronization hardware at each processor, and uses a separate network solely for clock signals. Software clock synchronization algorithms [CAS86, HSSD84, ST87, LL88, LMS85, VCR97, PB95, GZ89, MS85, FC95a, FC97, MO83, OS94, Arv94, Cri89] use standard communication networks and send synchronization messages to get the clocks synchronized. They do not require speci c hardware, but do not provide synchronization as tight as hardware algorithms. Software clock synchronization algorithms decompose themselves in deterministic, ....

.... not concerned in this paper with algorithms aimed at maintaining logical clocks that reAEect properties on the execution of distributed applications, such as causal relationships [Lam78] In addition, we do not detail issues speci c to clock synchronization in large scale distributed systems (e.g. [SR87, VCR97]) Clock synchronization has been extensively studied for the last twenty years. Thorough surveys can be found in [Sch86, RSB90] and [SWL90] In [RSB90] software and hardware clock synchronization algorithms are classi ed with regard to the clock correction scheme used. In contrast, the ....

[Article contains additional citation context not shown here]

P. Verissimo, A. Casimiro, and L. Rodrigues. Cesiumspray: a precise and accurate global time service for large scale systems. Journal of Real-Time Systems, 12:243294, 1997.

....clock synchronization is not required for logical clocks. With the advent of the Global Positioning System (GPS) we now have a tool that allows the use of accurately synchronized physical clocks in networks that span wide areas. By using GPS receivers or GPS based clock synchronization protocols [15, 16], a very precise time source can be made available to the processors. By these means, physical time is provided with an accuracy that is orders The message delivery latency is defined to be the sum of the channel latency (transmission delay plus propagation delay plus intermediate switch delay) ....

P. Verissimo, L. Rodrigues and A. Casimiro, "Cesiumspray: A precise and accurate global time service for large-scale systems," Real-Time Systems, vol. 12, no. 3 (May 1997), pp. 243-294.

....generating a message at any time, with whatever syntax and semantics (form and meaning) and sending it to anywhere in the system. Hybrid assumptions combining both kinds of failure assumptions would be desirable. They are a known framework in dependable system design vis a vis accidental failures[6, 8, 13]. Generally, they consist of allocating di#erent assumptions to di#erent subsets or components of the system, and have been used in a number of systems and protocols. Hybrid models allow stronger assumptions to be made about parts of the system that can justifiably be assumed to exhibit ....

P. Verssimo, L. Rodrigues, and A. Casimiro. Cesiumspray: a precise and accurate global time service for large-scale systems. Journal of Real-Time Systems, 12(3):243--294, May 1997.

No context found.

Verissimo, P., Rodrigues, L., Casimiro, A.: CesiumSpray: A Precise and Accurate Global Time Service for Large-Scale Systems. Journal of Real-Time Systems 12 (1997) 243--294

No context found.

P. Verissimo, L. Rodrigues, and A. Casimiro. Cesiumspray: a precise and accurate global time service for large-scale systems. Technical Report NAV-TR-97-0001, Universidade de Lisboa, 1997. 10

No context found.

P. Verissimo and L. Rodrigues. Cesiumspray: a precise and accurate global time service for large-scale systems. Real-Time Systems, 3(12):243--294, 1997.

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