Cost function based event triggered Model Predictive Controllers application to Big Data Cloud services

Abstract

High rate cluster reconfigurations is a costly issue in Big Data Cloud services. Current control solutions manage to scale the cluster according to the workload, however they do not try to minimize the number of system reconfigurations. Event-based control is known to reduce the number of control updates typically by waiting for the system states to degrade below a given threshold before reacting. However, computer science systems often have exogenous inputs (such as clients connections) with delayed impacts that can enable to anticipate states degradation. In this paper, a novel eventtriggered approach is proposed. This triggering mechanism relies on a Model Predictive Controller and is defined upon the value of the optimal cost function instead of the state or output error. This controller reduces the number of control changes, in the normal operation mode, through constraints in the MPC formulation but also assures a very reactive behavior to changes of exogenous inputs. This novel control approach is evaluated using a model validated on a real Big Data system. The controller efficiently scales the cluster according to specifications, meanwhile reducing its reconfigurations.

Anomaly detection and diagnosis for cloud services: Practical experiments and lessons learned

“…The wide adoption of virtualization in several application domains gave rise to a large body of research work covering a variety of topics, dealing for instance with elastic resource contention issues of VMs or services Kundu et al (2012); Matsunaga and Fortes (2010); Bodík et al (2009); Silvestre et al (2015a), server reconfiguration Cerf et al (2016), resources and energy management Chase et al (2001); Berral et al (2010), and security Bhat et al (2013); Gander et al (2013). Other works also propose frameworks dedicated to the processing of large datasets (or big data) sourced from cloud infrastructures such as Pop (2016).…”

A control theoretical view of cloud elasticity: taxonomy, survey and challenges

“…The paper, however, lacks details about the results obtained. Whereas, Cerf et al [38] focused on the idea of reducing the number of elasticity decisions for big data cloud systems using an MPC controller coupled with an event triggering mechanism. The event triggering mechanism serves as an additional layer to determine whether the MPC decision will be carried out or not.…”

“…In cloud elasticity domain, the majority of existing solutions belong to the category of either regulatory control (e.g., [28][29][30][31][32][33]) or optimisation (e.g., [34][35][36][37][38][39]). The control solutions having the objective of disturbance rejection often assist another control solution (e.g., [40][41][42][43]).…”