Abstract:
In complex computing environment the management of services, middleware and their
coordination is needed. Cloud orchestration is the term used in this concept to manage the
resources and coordinate with the complex services and software. Cloud computing is at infancy
stage and gradually organizations are moving toward it. In cloud, a software cost is minimized
and management of cloud is much easy as compared to the traditional services. Despite of giving
different facilities some issues must be addressed to gain the confidence for this emerging field.
The important issue that is centralized approach which causes the bottleneck performance and
other issue is when failure arises. In that case the solution which comes in mind is to replace the
centralized approach with peer to peer in which servers are replaced by the nodes which may act
as client and can be a server also. In this research work, middleware has been proposed to
overcome the issue of scalability and efficiency especially in peer to peer network. This includes
the division of processing among all the peers. The focus of this research is to introduce a
middleware that intelligently measures the workload of each peer and assign the work
accordingly, to better utilize the processing power of each peer without using a centralized
control node based approach. In peer to peer tasks are divided among different peers and each
peer is giving its resources in term of processing. At the top the cloud user specify the task to be
executed. At middle the load of each peer is to be checked to specifically assign the task to the
peer having fewer loads. The task is assigned and executed at that peer and results are sent to the
users who have requested the service. The proposed architecture is called SEPMid. The
middleware is designed to do this task automatically. User only requests the service and gets the result the whole task is done automatically. The proposed solution will be a generic solution.
