SUMMARY REMARKS & TESTBEDS
Modio is developing a new solution for computationally-based NFV autoscaling. Our Value Proposition is formed around our technical innovation to be able to cope with large variations of resource demand without avoiding either over allocation of resources that leads to excessive operational costs to the 5G telco operator or under allocation that leads to clients’ SLA violations of the offered services. The Proof of Concept of IntelligentNFVAutoscalerhas recently been validated at small-scale within Modio’s Open Call SoftFIRE project1. Specifically, we have experimented with state of the art forecasting algorithms (ARIMA, Holt Winters and Recurrent Neural Networks) to predict the upcoming workload and dynamically calculate the parameters of scale-out and scale-in actions. In comparison with static autoscaling approaches used by current NVFOs, such as OpenBaton, we have demonstrated that our predictive models are effective in deriving appropriate scale-out and scale-in actions’ parameters that outperform autoscaling policies with fixed step. The above have been validated via series of experiments within SoftFIRE’s federation.However, although the obtained results have shown that our autoscaling concept has a promising potential for commercial exploitation, additional work is still needed to (a) implement and experiment at a much larger-scale involving both stationary and mobile WebRTC clients, (b) to support the Kubernetes framework that was not supported by the SoftFIRE federation. Fed4FIRE+ can provide us the support to implement and experiment with our additional work via its Tengu, Virtual Wall and w-iLab.t testbeds. This experiment shall allow us to move from a small-scale Proof of Concept demonstration to an enhanced and refined implementation which will bevalidated at large scale. By achieving this goal, we will be in competitive position to actively pursueVC backing to develop a Minimum Viable Product that will aid us to enter the 5G NFV market.