IJSRSET1623128 | Received : 25 May 2016 | Accepted : 30 May 2016 | May-June 2016 [(2)3: 417-423]
© 2016 IJSRSET | Volume 2 | Issue 3 | Print ISSN : 2395-1990 | Online ISSN : 2394-4099
Themed Section: Engineering and Technology
417
Cost of Dynamic Migration of Content and Load Balancing
Distribution Services into Hybrid Clouds
P. D. Walunj, Prof. S. K. Sonkar
Department of computer engineering Amrutvahini College of Engineering, Sangamner, Maharashtra, India
ABSTRACT
Cloud computing is a technology which is most recently used for resources sharing for end users. Cloud providers
provide services to end user like software, Application and Platform as services to complete their job. One of the
objectives of cloud computing is load balancing. In this paper we used hybrid cloud for taking advantages of both
cloud i.e. private cloud and public cloud. Content distribution services where cloud provider switch contents on
different servers to improve load balancing. The contents are migrated or distributed dynamically on different
servers. We used the Lyapunov optimization theory to minimize operational cost of the application with quality of
services guarantees. In Lyapunov theory operational cost is reduces using threshold and rating parameter. If rating
is above the threshold values then cost is increases same for if rating value is below then the cost is decreases.
Threshold values is one kind of limit is define as constant. We proposed dynamic control algorithm is used to design
for dynamic load balancing on servers and also provide higher security.
Keywords: Content-Distribution; Dynamic Resource Allocation; Load Balancing; Dynamic Migration; Lyapunov
Optimization theory.
I. INTRODUCTION
Cloud computing is recently and widely used technology
where cloud provider provides resources to user to
complete their job. There are types of resources end user
want to complete their need for the application they
might be computational resources or networking
resources. When resources are provide to end user from
the cloud provider the resources are in order and in time.
Resources are servers to end user in number of order and
allocated time slice to each resource. Paper [7] used
selective algorithm for dynamic resource allocation to
end user to complete their need of application .selective
algorithm is based on min max algorithm which is
minimize overall task on machine. [11]user send a
request to need a resources but in some cases the
respected resource is not available to allocated in this
condition request is in longer delay .Cloud provider is
for utilizing and allocating resources within the limit of
cloud environment. We used hybrid cloud is a
combination of both public and private cloud. CDN is
like private cloud and DCN is like a public cloud. One
data center is a place is in each region which is
geographically distributed.
Load balancing is one of the main objectives of the
cloud computing to manage the spaces on the servers.
We used dynamic control algorithm to manage load
balancing. In this algorithm we used concept hot spot
and cold spot. Hot spot technique if user uploads data on
respected server and if server full then migrated content
on different servers and leave the respected server.
When user is requesting to download file or data the user
request is migrated on different server and dispatch from
the data center.
We used Lyapunov optimization theory [11] [12] to
reduce operational cost. There is measure parameter are
used to calculate updated cost is depend upon rating and
time. Rating is parameter where numbers of user
download a data or particular file with in time; cost is
updated on this parameter. Threshold value is predefine
if rating is above than threshold value then cost is
increase with constant values. Same as if value is above
threshold but long duration rating value is low, then cost