Monitoring Platform Evolution towards Serverless Computing for 5G and Beyond Systems

• DOI: 10.36227/techrxiv.19095788
• Publication Date: 2022-02-07T07:11:26Z
• Authors: Matteo Pergolesi; Ramon Perez; Priscilla Benedetti; Jaime Garcia-Reinoso; Aitor Zabala; Pablo Serrano; Mauro Femminella; Gianluca Reali; Kris Steenhaut; Albert Banchs
• Keywords: Microservices

Fifth generation (5G) and beyond systems require flexible efficient monitoring platforms to guarantee optimal key performance indicators (KPIs) in various scenarios. Their applicability Edge computing environments requires lightweight solutions. This work evaluates different candidate technologies implement a platform for 5G these environments. For data plane technologies, we evaluate virtualization including bare metal servers, virtual machines, orchestrated containers. We show that containers not only offer superior flexibility deployment agility, but also allow obtaining better throughput latency. In addition, explore the suitability of Function-as-a-Service (FaaS) serverless paradigm deploying functions used manage platform. is motivated by event oriented nature those functions, designed set up infrastructure newly created services. When FaaS warm start mode used, gives users perception resources are always available. cold running application’s modules automatically destroyed when application use. Our analysis compares both them with standard microservices. The experimental results produces significant latency increase, along potential instabilities. this reason, its usage recommended despite savings resources. Conversely, executing configuration tasks infrastructure, it can provide similar execution times microservice-based deployment. approach significantly simplifies code logic comparison microservices, reducing lines less than 38%, thus development time. Thus, represents best technology implements such management functions.