Optimisation of Resource Selection - Capabilities

This capability allows optimization of the execution of an Application / Task / Service Instance with an associated contract, namely the SLA, in order to maximize the probability of SLA satisfaction.

The requests, under the form of an SLA, get processed to select the best host, among all the available ones, to which the elaboration of the requests is assigned.

Obviously, the definition of best for a host depends on the state of the variables (e.g. the metrics of the SLA) in the system: available resources, resources that are needed to satisfy SLA requests and objectives to optimize. Moreover, some of the objectives are directly related to the knowledge of the requests in an SLA (completion time minimization, cost minimization, success probability maximization etc.) while other objectives are related to the state of the system (for example, workload balance).

The capability works according to the following flows:

·Mapping of high level application-specific business objectives (negotiated between client and provider and formalized into a Pre-SLA document) into low-level infrastructural parameters (such as CPU, bandwidth, memory usage) that can be quantitative measured.

·After the mapping of high level objectives into low-level parameters, the mathematical objective function is defined (the function representing the goals of the problem that we want to optimize subject to problems constraints).

·When an SLA requires a resource, the service, considering the state of the system, defines a list of available resources. The list is ordered according to a parameter that defines the quality in matching the given SLA to a given resource.