R. van Renesse and K. Birman. Fault-tolerant programming using process groups. In F. M. T. Brazier and D. Johansen, editors, Distributed Open Systems, pages 96--112. IEEE Computer Society Press, Los Alamitos, California, 1994.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....maintains one or a set of trading operations. If it fails, these trading operations cannot proceed, but other (trading and database) operations are not affected. 2. 4 Fault Tolerant Models for Servers Various models have been proposed in building fault tolerant servers based on server replication [4, 9, 10]. In [14] and [16] we have proposed four types of fault tolerant servers, two of them are described below. A service providing server (Figure 3(a) has a buddy server running (or in a cold stand by state, i.e. ready to be executed) somewhere in the network, but no communication between the server ....

R. van Renesse and K. Birman. Fault-tolerant programming using process groups. In F. M. T. Brazier and D. Johansen, editors, Distributed Open Systems, pages 96--112. IEEE Computer Society Press, Los Alamitos, California, 1994.

....Concurrent C [2] and the fault tolerant version of the SR language [3] The second approach provides a faulttolerant computing toolkit or a model that can be used together with general programming languages and standard operating systems. Typical examples of this approach are the ISIS toolkit [4], the ARGUS system [5] the location based replication paradigm [6] the fulfillment transactions approach [7] and the RPC transaction management system [8] This paper follows the second approach. Remote Procedure Call (RPC) is perhaps the most popular model used in today s distributed software ....

....simplifies the understanding and reasoning about programs. There have been some efforts to combine twoofthe three techniques ; RPC, replication, and transaction ; together to achieve reliable computing. However, none of the existing systems proposals are completely satisfactory. The ISIS toolkit [4] is a distributed programming environment, including a synchronous RPC system, based on virtually synchronous process groups and group communication. ISIS combines RPC and replication techniques to achieve the goal of developing reliable programs. A special process group, called a fault tolerant ....

R. van Renesse and K. Birman. Fault-tolerant programming using process groups. In F. M. T. Brazier and D. Johansen, editors, Distributed Open Systems, pages 96--112. IEEE Computer SocietyPress,Los Alamitos, California, 1994.

....to CORBA, these models address issues related to partial failures, process replication, reliable multicast, asynchronous communication, and ordering of events. Implementation of robust distributed systems on conventional hardware is enabled by toolkits that follow these models, e.g. by Horus [25], Isis [3] Transis [1] and Consul [15] Unfortunately, the programming interface provided by these toolkits is proprietary and rather low level; it mainly consists of a rich set of C procedures presenting access to light weight processes, unstructured messages, process groups, message passing ....

van Renesse, R., and Birman, K. P. FaultTolerant Programming using Process Groups. In Distributed Open Systems, F. Brazier and D. Johansen, Eds. IEEE Computer Society Press, 1994.

....in use at several sites. However, ISIS is not particularly flexible: it presents users with a sharp set of tools suitable for one particular kind of fault tolerant system, but it is not particularly useful if other kinds of system organization are preferred. In the most recent formulation of ISIS [16, 17], the sender of a message must itself be a member of a group in order to send a message to that group. This means that an ISIS process group cannot replace a single existing server process directly. If a client wishes to make a request of a server group to which it does not belong, it must send ....

Renesse, R. and Birman, K. Fault-tolerant Programming using Process Groups . To appear, some IEEE publication or other. . November 1992.

....messages on non volatile storage and deliver them to their recipients as soon as they become available and register with the GTS again. 1. 2 Related Work Examples of state of the art toolkits offering process groups and reliable, order preserving multicast are Consul [12] Electra [10] Horus [14], Isis [4] and Transis [1] The default policy of these toolkits is to remove failed members from their groups. A recovered process will have to re join the system, and to request the current application state from a group member. If all members of a group fail, processes which depend on the ....

van Renesse, R., and Birman, K. P. Fault-Tolerant Programming using Process Groups. In Distributed Open Systems, F. Brazier and D. Johansen, Eds. IEEE Computer Society Press, 1994.

....operation, encrypted communication, reconfiguration, and in that applications may transmit messages without waiting until they have been delivered. Examples of state of the art toolkits offering process groups and reliable, order preserving multicast are Consul [14] Electra [12] Horus [16], Isis [4] and Transis [1] These toolkits primarily aim to support applications running within one LAN. If a process has been unresponsive for a certain (usually short) period of time, their default behavior is to regard the process as faulty and to exclude it from the system. If a process or a ....

van Renesse, R., and Birman, K. P. FaultTolerant Programming using Process Groups. In Distributed Open Systems, F. Brazier and D. Johansen, Eds. IEEE Computer Society Press, 1994.

....flexibility it is not easy to achieve. Different approaches for group oriented synchronous CSCW use well established groupcommunication technologies as group communication sub systems. However, state of the art toolkits offering process group and reliable order preserving multicast like [6] [7], 8] and [9] aim, primarily, the support for fault tolerance in a LAN environment. Moreover, these systems offer too low level primitives and abstractions for collaborative purposes. This originates a considerable gap between the abstractions needed to describe collaborative processes and the ....

Van Renesse, R., Birman, K. "Fault Tolerant Programming Using Process Groups", in Distributed Open Systems, F. Brazier and D. Johansen, Eds IEEE Computer Society Press (1994).

....also offers security features, and is able to deal with network partitioning. Horus system provides the following properties, which are group and point to point communication layered communication semantics, support for multiple transport protocols, threads, security model and failure consistency [10,12,13,14,15]. Transis is a transport level group communication service that simplifies the development of fault tolerant distributed applications. The system distinguishes itself in allowing partitionable operation, in which multiple network components that are temporarily disconnected from each other operate ....

Van Renesse R., Birman K.P., Fault-tolerant Programming Using Process Groups, IEEE Distributed Open Systems, December 1993.

....as soon as they become available and register with GTS again. Addressing is by Uniform Resource Locators similar to the World Wide Web. 1. 2 Related Work Examples of state of the art toolkits offering process groups and reliable, order preserving multicast are Consul [12] Electra [10] Horus [14], Isis [4] and Transis [1] These toolkits primarily aim to support applications running within one LAN. If a process has been unresponsive for a certain (usually short) period of time, their default behavior is to regard the process as faulty and to exclude it from the system. If a process or a ....

van Renesse, R., and Birman, K. P. FaultTolerant Programming using Process Groups. In Distributed Open Systems, F. Brazier and D. Johansen, Eds. IEEE Computer Society Press, 1994.

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