Optimization algorithms for placement of virtual machines in the cloud computing model SaaS.
The major advantage for users of cloud infrastructure is the ability to refuse from maintaining of expensive computing equipment. Due to the use of virtualization technology the cloud can satisfy the different requirements of a large number of users with a single set of physical hardware.
Cloud architecture has two the most significant parts. One that is visible to the user, includes the application and is used by computer to access the cloud. The second part is the cloud itself, and is composed mainly of servers and storage devices.
Cloud software consists of applications that are available through the Internet, and are used as a service, at a time when they are needed. Applications are built on the base of a cloud architecture and are able to deploy resources dynamically in the necessary time and release them after use.
The broker is an intermediary between users and cloud resources in the data center. Broker is responsible for the allocation of resources between users of cloud, distributing the virtual machines for applications, according to the agreement on the level of service (Service Level Agreement (SLA)). SLA is a formal agreement between the client and supplier of services, containing a description of services, the rights and obligations of the parties and, most importantly, a consistent level of quality of service.
Algorithm for the distribution of tasks could benefit both the user and the owner of the cloud. On the one hand algorithm must satisfy the constraints QoS (Eng. Quality of Service - QoS) for users, on the other hand, it can distribute tasks among virtual machines so as to reduce the load, increase throughput and reduce power consumption.
The project is implemented under the Basic Research Program of the Presidium of RAS №14 "Problems of creation of scientific information and computing environment based on GRID technologies, cloud computing and advanced telecommunication systems".