Analiza mechanizmów tolerowania awarii w architekturach przetwarzania rozproszonego.

• Autorzy: Karwaczyński P.

Warianty tytułu

EN

Analysis of fault tolerance mechanisms in distributed processing architectures.

• Języki publikacji: PL

Abstrakty

PL

Wraz z postępującym rozwojem społeczeństw informacyjnych rośnie znaczenie systemów informatycznych, pozwalających na uzyskanie dostępu do informacji, jej przetwarzanie oraz wymianę. Ponieważ źródła informacji najczęściej są geograficznie rozproszone, stąd jednym z najbardziej eksploatowanych współcześnie obszarów informatyki jest przetwarzanie rozproszone. Przetwarzanie rozproszone w stosunku do przetwarzania scentralizowanego jest nieporównywalnie bardziej złożone i zawodne. Aby uprościć tworzenie systemów rozproszonych z powodzeniem rozwija się i wykorzystuje oprogramowanie warstwy pośredniczącej (middleware) oraz wszelkiego rodzaju architektury i technologie wspierające. Tematem niniejszego artykułu jest analiza występujących w wybranych rozwiązaniach tego typu mechanizmów dotyczących podniesienia poziomu niezawodności.

EN

As a result of the ongoing growth of information societies, the role of computer systems allowing access to information as well as its processing and dissemination becomes increasingly significant. Due to the geographic distribution of diverse sources of information, one the most exploited domains in present computer science is distributed processing. Distributed processing, in comparison to cetralized processing, is far more complex and unreliable. In order to simplify the development of distributed systems, the middleware software and supporting architectures and technologies are successfully developed and utilized. The article deals with the analysis of fault tolerance mechanisms present in selected distributed processing architectures. The most typical means to provide fault tolerance is a hardware/software replication. It may allow system users to continue realization of their tasks in spite of failures of nodes, network infrastructure or software, providing that the appropriate redundancy is implemented. Advanced distributed systems are expected to behave in a self-controlled way, concealing failures and recovering from them when possible. This article details support of CORBA, Jini, Web Services and dotNET Remoting towards coping with failures. CORBA (Common Object Request Broker Architecture) is a technology consistent with the RM-OMA model. Its core element is ORB (Object Request Broker), responsible for all the communication performed between the client and the server. Systems using CORBA can be fault tolerant as they can take advantage of the implementations based on the OMG formal specification ,,Fault Tolerant CORBA". The specification introduces entity redundancy, fault detection and recovery mechanisms. The advanced implementation of this specification is Eternal, a component-based middleware framework. The term Web Services applies to the technology developed by W3C, which goal is to integrate the applications using the XML-based standards. Such integration should be possible due to the set of software components, middlewares and, above all, common protocols. Web Services specifications do not directly address fault tolerance. Nevertheless, preliminary, third-party solutions aiming at fault tolerant web services have been provided. Jini is completely Java-based technology. It consists of the programming interfaces and the network protocols, which allow development of distributed systems based on the paradigm of the federation of services. In this context a service is any software or hardware entity connected to the computer network, which can perform any usable function. The natural redundancy of services cooperating in their federation together with the lease indirectly improve the level of fault tolerance of a system. The dotNet Remoting is a technology developed by Microsoft, as a part of the dotNet platform. It is the successor of the DCOM technology, which allows the distributed objects to cooperate. The dotNet Remoting improves the DCOM making it more flexible and removes some of the drawbacks of the predecessor as well. Nonetheless, despite the lease mechanism, it does not support fault tolerance.

Słowa kluczowe

PL

przetwarzanie rozproszone; niezawodność; tolerowanie awarii

EN

distributed processing; reliability; fault tolerance

• Wydawca: Komitet Budowy Maszyn PAN
• Czasopismo: Zagadnienia Eksploatacji Maszyn
• Rocznik: 2005
• Tom: Vol. 40, nr 3
• Strony: 61--71
• Opis fizyczny: Bibliogr. 15 poz.

Twórcy

autor

Karwaczyński P.

• Politechnika Wrocławska, Instytut Informatyki Stosowanej, Wybrzeże Wyspiańskiego 27, 50-370 Wrocław, tel. [phone], fax: [phone],

Bibliografia

• [1] Coulouris G., Dollimore J., Kindberg T.: Systemy rozproszone. WNT, Warszawa 1998.
• [2] Dialani, V., Miles, S., Moreau, L., De Roure, D. and Luck, M.: Transparent fault tolerance for web services based architectures. Proceedings of 8th International Europar Conference (EURO-PAR'02), s. 889-898, Paderborn (Germany), 2002.
• [3] Gruber J., Jóźwiak I.J.: Niezawodność przemysłowych instalacji sieci intranetowych. Zagadnienia eksploatacji maszyn 2001, vol. 36, s. 53-68.
• [4] Jóźwiak I.J.: Reliability exploration of microcomputer systems using the Weibull distribution. Microelectron. Reliab. 1992, vol. 32, s. 337-340.
• [5] Jóźwiak I.J.: The influence of error occurrence on branched computer network reliability. Microelectron. Reliab. 1992, vol. 32, s. 487^192.
• [6] Jóźwiak I.J.: The influence of the number of users on the reliability of a branched computer network. Microelectron. Reliab. 1992, vol. 32, s. 483-486
• [7] Jóźwiak I.J.: Use of concomitant variables for reliability exploration of microcomputer systems. Microelectron. Reliab. 1992, vol. 32, s. 341-344.
• [8] McDonough T.: Object Level Fault Tolerance for CORBA-based Distributed Computing. 1999, http://dantanner.tripod.com/FT.htm.
• [9] Monson-Haefel R.: J2EE Web Services. Addison Wesley, Boston 2003.
• [10] Narasimhan P., Moser L.E., Melliar-Smith P.M.: Eternal - a component-based framework for transparent fault-tolerant CORBA. Software - Practice and Experience 2002, vol. 32, s. 771-788.
• [11] Nikander P.: Fault Tolerance in Decentralized and Loosely Coupled Systems. Proceedings of Ericsson Conference on Software Engineering (ECSE'2000), Stockholm (Sweden), 2000.
• [12] OMG: The Object Management Architecture Guide, Third Edition. June 1995. http://cgi.omg.org/docs/ab/97-05-05.pdf.
• [13] OMG: The Common Object Request Broker Architecture: Core Specification. March 2004, http://www.omg.org/cgi-bin/apps/doc7formal/04-03-12.pdf.
• [14] Srinivasan P.: An Introduction to Microsoft. Net Remoting Framework. July 2001, http://msdn.microsoft.com/library/default.asp?url=/hbrary/enus/dndotnet/html/hawkremoting.asp.
• [15] Sun Microsystems: Jini Architecture Specification. December 2001, http://www.sun.com/jini/specs/jinil_l.pdf.