Its general goal is to build a private Cloud environment that can provide end

Its general goal is to build a private cloud

This preview shows page 160 - 162 out of 386 pages.

China. Its general goal is to build a private Cloud environment that can provide end-users with a high-level programming interface, and users can utilize com- puting resources they need without considering where these computing resources come from (i.e. the layer of program interface is independent of the layer of computing resources). YML [ 14–16 ] is a large-scale workflow programming framework, developed by PRiSM laboratories at the university of Versailles and Laboratoire d’Informatique Fondamentale de Lille (LIFL, Grand Large Team, INRIA Futurs) at the University of Science and Technology of Lille. The aim of YML is to pro- vide users with an easy-to-use method to run parallel applications on different distributed computing platforms. The framework can be divided into three parts: end-users interface, YML frontend, and YML backend. End-users interface is used to provide an easy-to-use and intuitive way to submit applications, and appli- cations, can be developed using a workflow-based language, YvetteML. YML frontend is the main part of YML, which includes compiler, scheduler, data repository, abstract component, and implementation component. The role of this
Image of page 160
147 9 A Reference Architecture Based on Workflow for Building Scientific Private Clouds part is to parse parallel programs, into executable tasks and schedule these tasks to appropriate computing resources. YML backend is the layer to connect differ- ent Grid and Desktop Grid middleware through different special interfaces, and users can develop these interfaces very easily. The YML is a component-based framework in which components can interact with each other through well- defined interfaces and researchers can add/modify one or several interfaces for other middleware to YML very easily. Paper [ 18 ] presents a method of resource management in Clouds through a grid middleware. Here, we will extend YML to build scientific private Clouds for non-big enterprises and research institutes. We call this project “YML-PC.” Three steps are needed to make this project a reality. The first step is to inte- grate volunteer computing resources into dedicated computing resources through YML and make them work in co-ordination. Volunteer computing resources can be a supplement to dedicated computing resources and a volun- teer computing resources-based platform has the ability to expand computing resource pools dynamically by nature. If dedicated computing resources are not enough for users, volunteer computing resources can be utilized to implement their tasks. But users do not know whether their tasks are run on dedicated computing resources or volunteer computing resources, and they need not know. The key issue of this step is how to allocate tasks to different kinds of computing resources more reasonably and make those computing resources work with high efficiency. The second step is to develop an interface for Hadoop and integrate it into YML. Then, some evaluations will be made on cluster environment + Hadoop. The third step is to try to build a hybrid Clouds
Image of page 161
Image of page 162

You've reached the end of your free preview.

Want to read all 386 pages?

  • Spring '16
  • Mr Gebre

  • Left Quote Icon

    Student Picture

  • Left Quote Icon

    Student Picture

  • Left Quote Icon

    Student Picture

Stuck? We have tutors online 24/7 who can help you get unstuck.
A+ icon
Ask Expert Tutors You can ask You can ask You can ask (will expire )
Answers in as fast as 15 minutes