SluggishQCloud: Towards the Sluggish Queueing assisted Cloud System for Efficient Resource Management

Abstract

Translate Article

The provision of computing resources such as com- puting power, storage, networking, and IT infrastructure dynamically and virtually by service providers over the Internet on a pay-per-use basis is known as cloud computing. Customers can send a lease to the cloud service provider via the internet to request cloud-based resources. Leases are resource provisioning abstractions that include all necessary resources, the provisioning time, and the customer’s budget, among other things, in termsof the quantity, capacity, and duration of nodes. The service provider then accepts all customer leases, generates virtual machines for them, and schedules them with the assistance of a lease manager. Lease managers utilize their algorithms to handle leases, which include scheduling leases. OpenNebula uses Haizea, an open-source VM-based leasing management framework, as a resource manager. It acknowledges the user’s lease request and subsequently arranges the lease according to the physical machines’ resource availability. It enables users to submit requests as leases. In haizea, preemptible best-effort leases can be supplanted by leases with a higher priority, such as those with an advanced reservation or an immediate lease. This article models each VM as a service provider and extracts its different performance metrics, including average number of request leasesin the system/ queue, the mean waiting time of the leases in the system/queue.