RUDA, Miroslav, Jiří DENEMARK and Luděk MATYSKA. Scheduling Virtual Grids: the Magrathea System. In Second International Workshop on Virtualization Technology in Distributed Computing. USA: ACM digital library, 2007, p. 1-7. ISBN 978-1-59593-897-8.
Other formats:   BibTeX LaTeX RIS
Basic information
Original name Scheduling Virtual Grids: the Magrathea System
Name in Czech Plánování virtuálních gridů: systém Magrathea
Authors RUDA, Miroslav (203 Czech Republic, guarantor, belonging to the institution), Jiří DENEMARK (203 Czech Republic, belonging to the institution) and Luděk MATYSKA (203 Czech Republic, belonging to the institution).
Edition USA, Second International Workshop on Virtualization Technology in Distributed Computing, p. 1-7, 7 pp. 2007.
Publisher ACM digital library
Other information
Original language English
Type of outcome Proceedings paper
Field of Study 10201 Computer sciences, information science, bioinformatics
Country of publisher Czech Republic
Confidentiality degree is not subject to a state or trade secret
RIV identification code RIV/00216224:14330/07:00023038
Organization unit Faculty of Informatics
ISBN 978-1-59593-897-8
Keywords in English cluster virtualization; grid computing; preemption; heterogeneous environment
Tags cluster virtualization, grid computing, heterogeneous environment, preemption
Tags International impact, Reviewed
Changed by Changed by: Mgr. Miroslav Ruda, učo 1284. Changed: 11/4/2012 15:02.
Abstract
Contemporary production grids do not usually offer the flexibility users are looking for. While different user communities have often contradictory requirements on the operating system, libraries, and applications, the production Grids provide only one rigid environment. This rigidness can be overcome by virtualization, when every user community or even individual user can be provided with its own instance of a virtual Grid, running optimized and tailored operating system and services. The promise of higher flexibility of virtual Grids is compensated by the increase in scheduling complexity. In this paper, we present the Magrathea system that extends the Grid resource management systems with support for virtual environment. After discussing the design requirements, we introduce the Magrathea architecture that consists of three components: the master and slave processes running on virtualized resources and the cache process to provide the information about virtual machine state to the scheduler. Two virtual machines sharing one physical resource and used exclusively, preemption of a lower priority job running in a virtual machine, support for more than two concurrently domains and support for ``frozen'' services that are repeatedly invoked and suspended are the use scenarios discussed in the second part of the paper. We demonstrate how they are supported by the Magrathea system and what modifications to the Grid resource management system are necessary. The Magrathea is currently in the pre-production use on the Czech national Grid environment MetaCenter.
Abstract (in Czech)
Článek popisuje systém Magrathea, který umožňuje plánovat virtuální stroje na pracovních uzlech clusteru. Na příkladech použití demonstruje sdílení jednoho uzlu mezi dvěma různými gridy, preempci jednoho virtuálního stroje druhým privilegovaným a podporu služebních domén, které jsou v době nečinosti suspendované.
Links
MSM0021622419, plan (intention)Name: Vysoce paralelní a distribuované výpočetní systémy
Investor: Ministry of Education, Youth and Sports of the CR, Highly Parallel and Distributed Computing Systems
PrintDisplayed: 27/7/2024 13:31