VioLinn: Proximity-aware Edge Placementwith Dynamic and Elastic Resource Provisioning

Deciding where to handle services and tasks, as well provisioning an adequate amount of computing resources for this handling, is a main challenge edge systems. Moreover, latency-sensitive constrain the type location devices that can provide needed resources. When available are scarce there possibility some resource allocation requests denied. In work, we propose VioLinn system tackle joint problems task placement, service device provisioning. Dealing with achieved through proximity-aware algorithms ensure tasks handled close end-user. concept spare introduced sudden load variations in time space without having continuously overprovision. Several selection proposed different cost/performance tradeoffs. Evaluations performed both Kubernetes-based testbed using simulations show benefit handling localized spikes higher quality (QoS) lower usage. The study shows it possible achieve increase QoS tradeoffs against cost performance.