Today more and more business processes are using different telecommunication and information technologies to facilitate and improve business efficiency. Each business process is organized as a sequence of components which cooperate together in order to achieve common goal. This goal is to provide high quality services for clients. The main component that is used for building such systems is web-service. Web-service is a component that incapsulates a set of functions that can be invoked remotely over the Internet. The request can be directed through different order of web-services depending on the business process’s workflow. Each web-service on this route has different parameters of quality of service. That is why it is very important to ensure high quality of service for the whole route. One possible way is to change the route by using web-service which has better parameters of quality of service and provides the same functionality. For that purpose special routing algorithms have been created. These algorithms are implemented as a separate web-service which is called broker. Broker dynamically analyses quality of service parameters for all web-services that may be used to process the request and chooses the best one with the highest parameters of quality of service. But to ensure that a specific business process will flow with the highest efficiency and high cost profit, in this paper it is proposed to use the multivariate access method. This method uses the technology of adaptive migration of web-services and creates copy of the web-service that has been overloaded. As a result broker can use two instances of the same web-service to direct requests, and thus it can do dynamic load balancing. Such solution allows to the owner of the web-service to increase quality of service and get higher profit. This paper analyses the technology of web-services and parameters of quality of service of distributed systems for performing complex business processes. Here is proposed a way to increase quality of service for clients and optimize the business process using multivariate access method.