A Sensory-Data-Hosting Oriented Scheduling Strategy on Virtual Machine

Sensory data is a main kind of data resources against the background of Internet of Things, it puts forward higher requirements on the underlying IT infrastructure in the aspects of data scale, growth rate and demand of application processing. With the IaaS(Infrastructure as a Service)-layer cloud services whose main character is virtualization, the goal of pooling, managing and flexible using of the IT infrastructure resources can be achieved, meanwhile the IaaS-layer cloud services is beneficial for meeting the need of sensory-data-hosting or sensory-data-application-hosting. But when directly applying those existing solutions to sensory-data-hosting, the phenomenon of load tilt and unbalace may appear, resulting in the less-than ideal condition of emerging hot spots or cold spots. This paper puts forward a sensory-data-hosting oriented scheduling strategy on virtual machine, aim at solving the problem of existing solutions can not schedule dynamically according to the changing needs of sensory-data-hosting, also we have designed and implemented the scheduling system as well as the scheduling algorithm based on the greedy algorithm. Preliminary sensory data-hosting cases have shown that the strategy proposed in this paper can be used to realize business-sensitive scheduling on virtual machine, and the goal of load balancing and optimized energy saving can also be achieved.