A control theory approach for managing cloud computing resources: A proof-of-concept on memory partitioning

Abstract:

Autonomic cloud services need to adjust the number and partitioning of resources depending on observed workload changes. The goal is to maximize system performance while minimizing costs and meeting service level objectives (SLOs). This problem is well suited for the use of control engineering principles, where the system can be designed as a closed-loop controller. In this position paper, we discuss why control engineering approaches are suitable in this context and illustrate our argument using a motivating problem: dynamic memory partitioning for cloud caches. Basic control system approaches are presented as a starting research point in this direction.

Año de publicación:

2017

Keywords:

• control engineering
• memory partitioning
• CLOUD COMPUTING
• Resource Management

Fuente:

google

scopus