Translating Service Level Objectives to lower level policies for multi-tier services

Abstract

Service providers and their customers agree on certain quality of service guarantees through Service Level Agreements (SLA). An SLA contains one or more Service Level Objectives (SLO)s that describe the agreed-upon quality requirements at the service level. Translating these SLOs into lower-level policies that can then be used for design and monitoring purposes is a difficult problem. Usually domain experts are involved in this translation that often necessitates application of domain knowledge to this problem. In this article, we propose an approach that combines performance modeling with regression analysis to solve this problem. We demonstrate that our approach is practical and that it can be applied to different n-tier services. Our experiments show that for a typical 3-tier e-commerce application in a virtualized environment, the SLA can be met while improving CPU utilization by up to 3 times.

On understanding the economics and elasticity challenges of deploying business applications on public cloud infrastructure

“…Some cloud server bundles are restricted to a certain computing resource which vary between CISPs. For example, Terremark vCloud 11 and FlexiScale Server 12 are packaged in terms of CPU and RAM resources only. Therefore, if an application workload has variable CPU and I/O requirements, then adding new cloud server instances will only scale the CPU and RAM resources and additional network resources has to be rented and managed separately.…”

“…It should be indicated that achieving such needs is not an easy or straightforward task at all. Due to its significance, some researchers tried to address similar issues [9,11]. In their SLA Decomposition approach, Chen et al [11] proposed analytical models for capturing the relationships between high level system SLOs and low-level system component goals.…”

“…Due to its significance, some researchers tried to address similar issues [9,11]. In their SLA Decomposition approach, Chen et al [11] proposed analytical models for capturing the relationships between high level system SLOs and low-level system component goals. The low-level goals are then used to efficiently allocate and monitor corresponding computing resources in a private virtualized datacentre.…”

A Reference Architecture for Multi-Level SLA Management

“…It strongly relates to (performance) prediction and SLA translation techniques. Prominent approaches in this area rely on (layered) queueing networks (LQNs) [13], [14] or stochastic Petri nets (SPN) [15]. Chen et al [13] uses LQNs for translating service level objectives (SLOs) into low-level system thresholds.…”

Heavy-traffic revenue maximization in parallel multiclass queues

“…This holds because non-iterative expressions for the stationary performance indexes are lacking for multiclass closed models, thus most formulations require to use gradient-hill methods in combination with some model evaluation technique [10,17,24], e.g., approximate mean-value analysis (MVA) [15,33,43]. These formulations are popular in applications [5,16,31,39] but, as we show in the paper, they often fail to converge to a local optimum at the short timescales of minutes at which online capacity management systems operate today. This implies that practitioners can be forced to work with solutions that may be quite far from the global (or even a local) optimum.…”