Overview

The GridWay Metascheduler enables large-scale, reliable and efficient sharing of computing resources (clusters, computing farms, servers, supercomputers), managed by different LRM (Local Resource Management) systems, such as PBS, SGE, LSF, Condor, within a single organization (enterprise grid) or scattered across several administrative domains (partner or supply-chain grid). GridWay is a Globus project, adhering to Globus philosophy and guidelines for collaborative development and so welcoming code and support contributions from individuals and corporations around the world.

Core Functionalities:

A. Scheduling Capabilities
- A.1. Dynamic Scheduling
- A.2. Opportunistic Migration
- A.3. Performance Slowdown Detection
- A.4. Self-adaptive Applications
- A.5. Checkpointing Support
- A.6. Scheduling Policies
- A.7. Scheduling Reporting and Accounting
- A.8. Job Dependency
- A.9. Single and Multi-user Support

B. Fault Detection & Recovery Capabilities
- B.1. Job Cancellation
- B.2. Remote System Crash or Outage
- B.3. Network Disconnection
- B.4. Client Fault Tolerance
C. User Interface Functionality
- C.1. Broad Application Scope
- C.2. DRM Command Line Interface
- C.3. DRMAA Application Programming Interface
D. Installation & Configuration Issues
- D.1 Modular Architecture
- D.2 Requirements on Core Grid Services
- D.3 Supported Remote Services
- D.4 Supported Client Platforms
- D.5 Decentralized Architecture
- D.6 Security
- D.7 Meta-scheduling Infrastructure Scenarios

State-of-the-art: Scheduling Policies

The new Grid scheduling policy comprises two components:

Job prioritization policies: Pending jobs are prioritize according to four policies: fixed, fair-share, deadline and waiting-time. The job policies are used to sort the jobs of the users of a given scheduling domain (GridWay instance).
Resource prioritization policies: A given job can be executed on those resources that match it requirements. The resource policies are used to sort the matching resource list of each job. The matching resources are prioritize according four policies: fixed, usage, failure and rank.

Support for Job Submission Description Language (JSDL) : standard GridWay now supports the JSDL Open Grid Forum standard. JSDL is a language used to describe the requirements of jobs for submission to Grids (like the current job template used in GridWay). JSDL can be used in an alternative way to write job templates. So GridWay can be easily integrated with current applications/projects using JSDL to describe jobs.

Support for MPI jobs : GridWay provides support for the remote execution of MPI jobs to a single resource. MPI jobs are defined with the new TYPE template attribute (TYPE=mpi), and the NP attribute to set the number of processors required to execute the job.

New functionality in GridWay CLI : Last release includes new functionality in the following GridWay commands:

gwhost: A new option, -f, has been added to show all host parameters in a parameter-value format.
jsdl2gw: New command to translate JSDL requests to Job Template.
gwsubmit: A new option, -p, has been added to specify the fixed priority of the job amd it is also able to read the job templates for stdin. * gwps: New format options.
all: The help information provided with -h has been reviewed.

You can find out more information in the commands help (-h option) or in the command reference guide.

New attributes in the Job Template : Last release includes new attributes in the Job Template:

NAME to specify a name for the job.
DEADLINE to specify the deadline date for the new deadline policy for Job Prioritization.
TYPE and NP to indicate that a job uses MPI and the needed number of processors (NP), respectively.

Test suite : Last release incorporates a test suite to evaluate the quality of the software. The test suite can be used to check the installation of the meta-scheduler.

Support for EGEE VOs : The user is now able to specify the VO in the requirement expression. So, a single GridWay instance is able to provide execution support for users of different VOs and a user is able to manage different VOs from the same GridWay instance.

Further Details

Link to project site : http://www.gridway.org/
Development Status: Mature
License : Apache Software License
Used technologies : C, Globus Toolkit, GridFTP, Java, Python, Ruby
OS : Linux, SunOS/Solaris, MacOS
Link to component download : http://www.gridway.org/cgi-bin/GridWay/download.cgi?stable