[graap-wg] GSA-RG session at GGF 16
Philipp Wieder
ph.wieder at fz-juelich.de
Thu Feb 9 15:57:02 CST 2006
Dear All,
the GSA-RG would like to invite you to our session at GGF 16.
Date: Tuesday, February 14, 2006
Time: 3:30 pm - 5:00 pm
Location: Olympia B
We are especially looking forward to see three presentations on
state-of-the-art scheduling systems. These presentations will examine
generic Grid scheduling requirements and will therefore also be of
interest to groups like GRAAP or OGSA-RSS. Please find the respective
abstracts below.
I hope to see you in Athens.
Best regards, Philipp.
ABSTRACTS:
----------
1. "Workflow Scheduling in the ASKALON Grid Environment" by Marek
Wieczorek, University of Innsbruck
The ASKALON Grid environment is a full Grid environment designed to compose
and execute scentific workflow applications. The ASKALON scheduler is a
dynamic scheduling service using different optimization techniques to
maximize the user's profit in accessing the Grid. Pluggable architecture
of the scheduler enables the user to apply different workflow scheduling
algorithms, to apply advance reservations, and to consider different
optimization criteria, among them execution time and economic cost.
Acting as a part of a full Grid environment, the scheduler interacts
with other components of the ASKALON, namely with the resource
management, the enactment engine, the performance prediction and the
monitoring services, trying to make scheduling decisions based on the
most reliable and up-to-date information. Our goals are to make the
service SLA-aware and to extend it in the directions recommended by the
GGF community. Besides the functionality of ASKALON scheduler we will
also reflect general requirements for a Grid scheduler.
2. "The VIOLA Meta-scheduling Service" by Wolfgang Ziegler, Fraunhofer SCAI
Distributed applications or workflows usually require different
specialised resources like compute nodes, visualisation devices, storage
devices, or network connectivity with a defined QoS at the same time or
within a given period of time to be executed successfully. Orchestrating
such resources on a local level within one organisation is only a minor
organisational problem. Orchestration of resources on a Grid level
requires a service that is able to solve the same problems in an
environment that probably stretches across several administrative
domains. Additional conditions have to be taken into account, like site
autonomy and different site policies.
In this talk we first describe the environment where co-allocation of
resources is of vital importance, the requirements for the
MetaScheduling service that provides the required co-allocation means,
and related work. In the next step we characterise the functionality of
the MetaScheduling service, followed by the description of the current
implementation. As a use case we show the integration of the scheduling
system into the UNICORE Grid middleware and finally we present a new
project based on this work that started end of 2005.
3. "GridWay Scheduling Architecture" by Ignacio Martín Llorente,
Universidad Complutense de Madrid
GridWay, on top of Globus services, enables large-scale, secure and
reliable sharing of computing resources (clusters, computing farms,
servers, supercomputers...), managed by different resource management
systems (PBS, SGE, LSF, Condor...), within a single organization
(enterprise grid) or scattered across several administrative domains
(partner or supply-chain grid). GridWay is an open source
meta-scheduling technology that performs job execution management and
resource brokering, allowing unattended, reliable, and efficient
execution of jobs, array jobs, or complex jobs on heterogeneous and
dynamic Grids. GridWay performs all the job scheduling and submission
steps transparently to the end user and adapts job execution to changing
Grid conditions by providing fault recovery mechanisms, dynamic
scheduling, migration on-request and opportunistic migration. The
presentation describes the scheduling requirements for partner grid
computing, the functionality provided by GridWay to meet those
requirements, the GridWay scheduling architecture and finally its
requirements on core Grid services for the implementation of such
functionality.
More information about the graap-wg
mailing list