Fault-aware Scheduling in Grid Environment based on Linear Programming
Mehdi Sarikhani, Bahman Javadi, Askari Parichehre
Islamic Azad University of Qazvin, Iran
Sarikhani.mehdi@Gmail.com
Javadi@aut.ac.ir
apmastermail@gmail.com
POSTER PAPER
ABSTRACT
Grid system provides the sharing, selection and
aggregation on distributed autonomous resources while it
is an error prone environment. So, grid component like
scheduler must provide the user‘s Quality of Service
(QoS) requirements by selecting the appropriate
resources for user’s jobs. In this paper, we have proposed
a fault-aware economy scheduling model based on binary
integer programming (FES-BIP) to allocate resources to
application jobs such that the users’ requirements are met
while we know the distribution of resource failure in the
grid environment. FES-BIP algorithm guarantees optimal
resource selection.
KEYWORDS: Quality of Service, economy model,
fault-aware scheduling, binary integer programming.
1. INTRODUCTION
Grid becomes more and more popular in large-scale
computing, because they enable the sharing of computing
resources that are autonomous resources and distributed
throughout the world. We assume the market-based grid
should consider the economy factors as a QoS user
requirement. In market-based mechanisms, time, cost, and
number of Processing Elements (PE) which a user needs
are the important QoS factors. Consequently, the
scheduling algorithms try to meet QoS factors and to
optimize these factors. Grids are more error prone than
traditional parallel machines as there are potentially
thousands of resources that are heterogeneous and sharing
among various applications [12]. So fault tolerance is an
important characteristic in grid as the dependability of
individual grid resources may not be guaranteed and
resources are autonomous. It becomes increasingly
difficult to guarantee that a resource being used is not
malicious in some way [17]. So providing appropriate
resource selection is vital to use grid systems perfectly.
Also resource selection and scheduling in grid
environment is always a controversial subject especially
when fault tolerant scheduling is crucial due to the lots of
failure. Therefore, the fault tolerant scheduling based on
economic factors is so complicated (NP-hard) and finding
the optimal answer is so difficult. There are several
approaches to deal with scheduling but one of the best
approaches to prove the scheduling result which is
affirmative and optimal is using mathematical model and
especially linear programming model. So by modeling
fault-aware scheduling system which can provide optimal
scheduling results, the scheduler can meet the user‘s QoS
requirement and prepare fault tolerant grid resource.
Albeit linear programming scheduling model has been
used in previous works to solve scheduling problems, all
of them work only on resource and job but obviously such
models cannot prepare a real grid model[9][10]. Because
they have to assume that all resource PEs has same cost
and computational power. They can’t take any decision
about the tasks of jobs, in particular about the length of
tasks and have to take all tasks with the same length. So it
is crucial to put another parameter into model which is the
resource PEs information, having assumed that the
number of PEs in each resource (machine) and cost of
each PE per second is possible. Our proposed scheduling
algorithm by adding such information about PEs of
resource on economy grid meet requirement of the time,
cost, and number of PEs which users need. In addition it
optimizes the cost and ensures reliability.
The rest of the paper is organized is as follows. In section
2, we review some related works. In section 3, we put
forward our linear programming model. In section 4, we
978-1-4244-6829-4/10/$26.00 ©2010 IEEE 656