Path Enumeration Solution for Evaluating Reliability, Based on the Path Algebra Framework

Abstract The demand for connectivity in this fourth industrial revolution, is growing in an exponential rate. As the Fourth-Generation communication systems are replaced by the Fifth and Sixth-Generation ones, a plethora of fundamental performance issues need to constantly be addressed. Virtual network reliability is one of those main challenges for deploying resilient service-centric and user-centric scenarios in Next-Generation Networks. The need for strengthening critical applications dictates the emergence of mechanisms that could inherently produce more reliable resource allocations, either at the cloud or edge level. However, until the adoption of fully artificially intelligent management systems, the enhancement of existing resource evaluation and allocation mechanisms is mandatory during this transitional era. To reach this pressing milestone, we will have to revise the already deployed reliability frameworks, assessing their orchestrating stages. Our research focuses on providing a novel evaluation mechanism that could successfully replace traditional path choosing techniques in reliability prone algorithms, regardless of their orientation. The proposed synthesized metrics enumerate the candidate paths validly and optimally, by adhering to the Path Algebra fundamental properties of monotonicity and isotonicity. Such an implementation can be deployed seamlessly, either on legacy or modern architectures, cementing the innovative qualities of its design. The experimental results of our simulations indicate that the proposed solution favors substantially more reliable paths than its shortest-path counterpart.