Escalonamento justo em infraestruturas de nuvem com múltiplas classes de serviço

Abstract

Cloud computing providers offer multiple service classes to deal with workload heterogeneity. Classes are distinguished by their expected Quality of Service (QoS), which is defined in terms of Service Level Objectives (SLO). A priority-based scheduling policy is commonly used to guarantee that requests submitted to the different service classes achieve the desired QoS. However, the QoS delivered during resource contention periods may be unfair to certain users. In this paper, we present a SLO-driven scheduling policy which takes the SLOs and actual QoS delivered for each request into account when making decisions. We used simulation experiments fed with traces from a production system to compare the SLO-driven policy with a priority-based one. In general, the SLO-driven policy delivered a better service than the priority-based one.

Exploiting SLAs through Application of Economic Analysis on Datacenters’ Autonomic Management

“…Consolidation heuristics has been extensively adopted to reduce energy consumption by datacenters. However, only Silva et al (2019) does it from the perspective of SLA amortization. Keeping SLO compliance rates far above the minimum agreed is a waste of resources.…”

“…In their approach, any VM arrival/exit event triggers a set of algorithms that use VM profit as a priority level to decide which VMs should be migrated or suspended to rearrange the datacenter. Silva et al (2019) also use a priority strategy, but taking the SLOs and availability delivered for each VM into account when making decisions. Although Silva's approach is similar to ours, the key difference is the strategy we use to push the experienced SLO to the contracted one -they suspend VMs while we migrate VMs to hosts with overbooking.…”

“…In the taxonomy proposed by Bittencourt et al (2018), this work's scheduler can be classified as having the objective of minimizing datacenter costs in the provider perspective. However, such a taxonomy can be augmented with a "SLA exploitation" category, in which this work and [Silva et al 2019] would fit.…”

Escalonamento orientado por qualidade de serviço em infraestruturas de computação na nuvem

“…Essas avaliac ¸ões foram executadas em etapas incrementais. A nova política de escalonamento foi apresentada no artigo "Escalonamento justo em infraestruturas de nuvem com múltiplas classes de servic ¸o", publicado no Simpósio Brasileiro de Redes de Computado-res e Sistemas Distribuídos (SBRC) 2019 [da Silva et al 2019]. Neste artigo comparamos as disponibilidades entregues pelo escalonamento orientado por QoS com as obtidas com escalonamento baseado em prioridade.…”