Error-Tolerant Resource
Allocation and Payment Minimization for Cloud System
ABSTRACT:
With virtual machine (VM) technology being
increasingly mature, compute resources in cloud systems can be partitioned in
fine granularity and allocated on demand. We make three contributions in this
paper: 1) we formulate a deadline-driven resource allocation problem based on
the cloud environment facilitated with VM resource isolation technology, and
also propose a novel solution with polynomial time, which could minimize users’
payment in terms of their expected deadlines. 2) By analyzing the upper bound
of task execution length based on the possibly inaccurate workload prediction,
we further propose an error-tolerant method to guarantee task’s completion
within its deadline. 3) We validate its effectiveness over a real
VM-facilitated cluster environment under different levels of competition. In
our experiment, by tuning algorithmic input deadline based on our derived
bound, task execution length can always be limited within its deadline in the
sufficient-supply situation; the mean execution length still keeps 70 percent
as high as user specified deadline under the severe competition. Under the
original-deadline-based solution, about 52.5 percent of tasks are completed within
0.95-1.0 as high as their deadline, which still conforms to the
deadline-guaranteed requirement. Only 20 percent of tasks violate deadlines,
yet most (17.5 percent) are still finished within 1.05 times of deadlines.
EXISTING
SYSTEM:
In literatures, traditional optimization
problems are often subject to the precise prediction of task’s characteristic
(or execution property), which is nontrivial to realize in practice.
Traditional job scheduling is often
formulated as a kind of combinatorial optimization problem (or queue-based multiprocessor
scheduling problem, due to the nonguaranteed performance isolation for multiple
tasks running on the same machines. That is, most of the existing deadline-driven
task scheduling solutions (from single cluster environment confined in LAN to
the Grid computing environment suitable for WAN are also strictly subject to
the queuing model under which a single machine’s multiple resources cannot be
further split to smaller fractions at will. This will eventually cause the raw-grained
resource allocation, relatively low resource utilization and suboptimal task
execution efficiency
DISADVANTAGES
OF EXISTING SYSTEM:
Users may wish to minimize their
payments when guaranteeing their service level such that their tasks can be
finished before deadlines. Such a deadline-guaranteed resource allocation with
minimized payment is rarely studied in literatures. Moreover, inevitable errors
in predicting task workloads will definitely make the problem harder.
PROPOSED
SYSTEM:
We make three contributions in this
paper:
1) We formulate a deadline-driven
resource allocation problem based on the cloud environment facilitated with VM
resource isolation technology, and also propose a novel solution with
polynomial time, which could minimize users’ payment in terms of their expected
deadlines.
2) By analyzing the upper bound of task
execution length based on the possibly inaccurate workload prediction, we
further propose an error-tolerant method to guarantee task’s completion within
its deadline.
3) We validate its effectiveness over a
real VM-facilitated cluster environment under different levels of competition.
ADVANTAGES
OF PROPOSED SYSTEM:
All the theoretical conclusions are
confirmed with our experiments. Specifically, in the situation with relatively
sufficient resources, the worst case tasks under the stricter deadline-based
allocation only take as about 0.75 times as their deadlines to complete, as
compared to the 1.2 times of the deadlines under the original user-predefined
deadline based allocation. We also observe that in the competitive environment,
the latter algorithm performs much more stable than the former instead, which
means that the latter tolerates the resource competition better. We also confirm
the effectiveness of our solution via the distribution of the number of tasks
with respect to execution times and user payments: in the competitive
situation, majority of tasks can be guaranteed to be completed within
deadlines.
SYSTEM ARCHITECTURE:
MODULES:
1.
Task Assignment
2.
Physical Node
3.
Virtual Machine
4.
Cloud server
5.
Allocated and Available resource
MODULES DESCRIPTION:
1. Task Assignment:
In cloud systems, the cloud proxy (a.k.a., server)
continually receives and responds to user requests (or tasks) with customized
requirements (or virtual machines). All tasks will be handled based on their
priorities (like Google task scheduler) or in terms of First-Come-First-Serve
(FCFS) policy when the tasks are of the same priorities. Each task’s execution
may involve multidimensional resources, such as CPU and disk I/O. A data mining
task, for example, usually needs to load a large set of data from disk before
or in the middle of its computation. Eventually, such a task may store its
computation results onto the local disk or a public server through network.
2.
Physical
Nodes:
A
physical node is typically a server or a virtual machine, but it can be any
active device attached to a network that is capable of sending, receiving, and
forwarding information over a communications channel. In other words, a
physical node is any active device attached to a network that can run a
chef-client and also allow that chef-client to communicate with a server.
3.
Virtual
Machines:
In our cloud model, any task will be executed on one
or more virtual machines with user-reserved resources and the payment is
calculated based on the customized resource (a.k.a., pay-by-reserve policy).
Adopting such a pricing policy is driven by three reasons. First, the
efficiencies of many applications usually rely on multiple resources but it is
nontrivial to precisely evaluate the exact amount of their consumption
separately on individual resources. Second, quite a few users prefer to
reserving resources for tolerating usage burst and guaranteeing their service
levels. Lastly, the alternative pricing policy, pay-as-you-consume, is rather
simple because its payment is always fixed regardless of the resource
allocation.
4.
Cloud
server:
In this module, we create a local Cloud
and provide priced abundant storage services. The users can upload their data
in the cloud. We develop this module, where the cloud storage can be made
secure. However, the cloud is not fully trusted by users since the CSPs are
very likely to be outside of the cloud users’ trusted domain. Similar to we
assume that the cloud server is honest but curious. That is, the cloud server
will not maliciously delete or modify user data due to the protection of data
auditing schemes, but will try to learn the content of the stored data and the
identities of cloud users.
5. Allocated and Available resource:
In this module, proved optimal for
minimizing the payment cost within user-defined deadline for his/her task, the
deadline still may not be guaranteed due to two factors, either bounded
available resources or inaccurate workload vector information about the task.
We propose effective technique, which provides a necessary and sufficient
condition of guaranteeing the task’s deadline given accurate prediction and
relatively sufficient re-sources.
SYSTEM CONFIGURATION:-
HARDWARE CONFIGURATION:-
ü Processor - Pentium –IV
ü Speed - 1.1
Ghz
ü RAM - 256
MB(min)
ü Hard
Disk - 20
GB
ü Key
Board - Standard
Windows Keyboard
ü Mouse - Two
or Three Button Mouse
ü Monitor - SVGA
SOFTWARE CONFIGURATION:-
ü Operating System :
Windows XP
ü Programming Language : JAVA
ü Java Version : JDK 1.6 & above.
REFERENCE:
Sheng Di, Member, IEEE, and Cho-Li Wang,
Member, IEEE-“Error-Tolerant Resource Allocation and Payment Minimization for
Cloud System” IEEE TRANSACTIONS ON
PARALLEL AND DISTRIBUTED SYSTEMS, VOL. 24, NO. 6, JUNE 2013.
SEE THE PROJECT OUTPUT HERE
