A Quest For Optimizing The Data Processing Decision For Cloud-Fog Hybrid Environments

In the Cloud, computing loads found a new virtualized and dynamically scaling home. Cloud data centers offered a safe haven from risky capital investments and high dependence on internal resources for enterprises and all of that was achieved by going back to the central computing concept. The pros are ideal for tried and tested scenarios of predictable or at least bounded enterprise loads. However, because of the pay-as-you-go business concept that is dominant in the Cloud, scenarios where a big percentage of requests turn up with low-value data (as in incomplete, meaningless or out-of-sync) can be financially detrimental to the Cloud tenant. The Internet of Things offerings, for example, expand the horizons of data sources for a certain Cloud based service but it also jumps to a new dimension in terms of the data filtering and preprocessing required of the same service to get to the core value-returning requests. This contributed to the emergence of Fog (edge) computing where some of the processing is done on the edge of the network with the aim of distributing the load and minimizing the network congestion caused by low-value data. The availability of Fog nodes poses a challenge to Cloud service designers regarding how to optimize the process. The decision as to where to perform each step of the data management can make the difference for Cloud providers in mitigating both the risk of pushing high loads to the Cloud servers and network and the risk of almost localizing the whole process and losing the benefits from Cloud services. To tackle this challenge, we consider the question of optimizing the decision process for a data-intensive highly distributed Cloud service. A novel optimization model is presented exploring the factors in effect with the objective of maximizing the Cloud provider value and initial experimental results are presented. Shown results offer some insight into the contradicting factors in play (cost, network capacity, node and Cloud capacity). This work builds towards a comprehensive technique that helps Cloud providers decide their data processing strategy in any mixed Cloud-fog Cloud service platform.