Workload Patterns for Quality-Driven Dynamic Cloud Service Configuration and Auto-Scaling

Abstract

Abstract-Cloud service providers negotiate SLAs for customer services they offer based on the reliability of performance and availability of their lower-level platform infrastructure. While availability management is more mature, performance management is less reliable. In order to support an iterative approach that supports the initial static infrastructure configuration as well as dynamic reconfiguration and auto-scaling, an accurate and efficient solution is required. We propose a prediction-based technique that combines a pattern matching approach with a traditional collaborative filtering solution to meet the accuracy and efficiency requirements. Service workload patterns abstract common infrastructure workloads from monitoring logs and act as a part of a first-stage high-performant configuration mechanism before more complex traditional methods are considered. This enhances current reactive rule-based scalability approaches and basic prediction techniques based on for example exponential smoothing.

Cloud architecture continuity: Change models and change rules for sustainable cloud software architectures

“…Our research contribution builds upon and extends earlier work into a unifying framework for sustainable cloud architectures that integrates evolution and adaptation under various types of uncertainties. In particular, our contribution are the following: a set of change models and change rules that enable specification of architecture transformations for evolution and adaptation of cloud architectures, an application of the framework to a concrete case on document management, and an evaluation of the effectiveness of the framework based on a retrospective evaluation of different cloud applications we have studied over the past years. …”

“…For the prediction, we need accuracy of the prediction but also good performance to use a potentially time‐consuming technique autonomously in a dynamic cloud environment . In this case, our solution combining traditional collaborative filtering with a workload pattern approach improves the accuracy of standard collaborative filtering, as past observations are gathered in the form of service workload patterns.…”

“…Control‐theoretic foundations are being explored here to manage uncertainty concerns whereby different models from statistics machine learning and fuzzy control are used . Our own work includes the prediction and the use of fuzzy sets . Fuzzy sets are a natural and effective solution and provides the opportunity to ease the specification of rules using intuitive linguistic concepts.…”

“…36,37 Our own work includes the prediction and the use of fuzzy sets. 10,11 Fuzzy sets are a natural and effective solution and provides the opportunity to ease the specification of rules using intuitive linguistic concepts. A performance-oriented adaptation of software systems is proposed in 1 study.…”

Software Architecture for the Cloud – A Roadmap Towards Control-Theoretic, Model-Based Cloud Architecture

“…identify stable quality utilisation patterns. We can map infrastructure workload patterns for CPU, storage and network utilisation at the infrastructure tier to service-level performance patterns, thus linking models (here pattern-based) across tiers [12]. We can implement a prediction technique for the same workload and performance prediction context, based on simple and double exponential smoothing to smoothen outliers and to anticipate trends.…”