Michel Gokan Khan won the best student paper award at IEEE NetSoft 20202020-08-27
At the IEEE Conference on Network Softwarization 2020, Michel Gokan Khan, doctoral student at the Department to Computer Science, together with Prof. Javid Taheri, Dr. Mohammad Ali Khoshkholghi and Prof. Andreas Kassler from Karlstad University, was awarded the best student paper award for paper entitled "A performance modelling approach for SLA-aware resource recommendation in cloud native network functions”.
Network Function Virtualization (NFV) becomes the primary driver for the evolution of next generation 5G networks that a broad launch of it is expected to take place in 2021. The software core behind modern 5G networks are based on a modern cloud-native and microservice architecture, that decouple components of each network function into multiple independently manageable microservices. Michel’s work mainly focuses on performance optimization of the software core behind modern 5G networks that is based on NFV.
- In this work, we used a machine learning approach to model the impact of each microservice’s resource configuration (i.e., CPU and memory configuration) on the most influential metrics (i.e. serving throughput and latency) of each chain of services in a performance testing environment. Then, considering the targeted Service Level Objectives (SLO), we proposed an algorithm to predict each microservice’s resource capacities in a production environment. The accuracy of our prediction on a Kubernetes-based prototype of a cloud native 5G Home Subscriber Server (HSS) provided by the Ericsson showed 78%-95% accuracy., says Michel.
In 2018, he won the best demo paper award for his NFV-Inspector’s paper, and in this work, he extended the core engine behind the NFV-Inspector.
- I am very happy to receive another award before my licentiate seminar and my forth award in total during my Phd studies in Karlstad university, says Michel.
The patent-pending technology proposed in this paper was a collaboration between Karlstad University and Ericsson and was based on the framework of the NFV-Optimizer project, partially funded by the Knowledge Foundation of Sweden.
A performance modelling approach for SLA-aware resource recommendation in cloud native network functions