Task Allocation for Energy Optimization in Fog Computing Networks With Latency Constraints

Fog networks offer computing resources of varying capacities at different distances from end users. A Node (FN) closer to the network edge may have less powerful compared cloud, but processing computational tasks in FN limits long-distance transmission. How should be distributed between fog and cloud nodes? We formulate a universal non-convex Mixed-Integer Nonlinear Programming (MINLP) problem minimizing task transmission- processing-related energy with delay constraints answer this question. It is transformed Successive Convex Approximation (SCA) decomposed using primal dual decomposition techniques. Two practical algorithms called Energy-EFFicient Resource Allocation (EEFFRA) Low-Complexity (LC)-EEFFRA are proposed their effectiveness tested for various traffic scenarios. Using EEFFRA/LC-EEFFRA can significantly decrease number requests unmet requirements when baseline solutions (from 48% 24% 10 MB requests). Utilizing Dynamic Voltage Frequency Scaling (DVFS) minimizes consumption (by one-third) while satisfying requirements.