The construction of a common objective function for analytical infrastructures

A decision maxim for efficient task realization within analytical network infrastructures

“…As public clouds in particular incorporate large computing infrastructures, the aspect of energy-efficient computing has become a significant theme (Luo et al (2012)). While each research study focuses on a certain problem, in accordance with Grum et al (2017), each focuses on a highly specific optimization problem and from a general, systemwide perspective, only local optima can be identified with each.…”

“…Previously presented architectural infrastructure levels were identified by Grozev & Buyya (2014), and combined with task realization strategies on the basis of heuristics in analytic infrastructures by . An integration of available concepts was researched by Grum et al (2017), with the aim of identifying global optima in analytical task processing over scenarios and environmental changes. A quantified benchmark of different task realization strategies, such as real-world live tests or simulations, remains to be developed.…”

“…13, an optimization may consider several influencing factors. Based on the design-oriented approach of Grum et al (2017), the following refers to a set of factors validated in three ways by experts providing a high level of expertise in analytical infrastructures.…”

“…• The transfer cost tr t i,j,k refers to the costs incurred because of the transfer of job part j from CPS i to CPS k at time step t. As one assumes these to develop over time because of current bottlenecks, they are time dependent, and we initially assume these to be constant. The assessment of efficient task processing is based on the empirically validated objectives of Grum et al Grum et al (2017), resulting in the use of Eq. 26.…”