Transparent Clouds: An Enhancement to Abstraction

TL;DR

This paper proposes mechanisms to enhance transparency and tweakability in cloud systems by leveraging pathways, enabling better resource management and energy efficiency even with abstraction and heterogeneity.

Contribution

It introduces pathway-based mechanisms to improve transparency and tweakability in cloud infrastructures, addressing challenges posed by abstraction and heterogeneity.

Findings

Transparency can be improved without changing core components.

Energy consumption can be reduced through dynamic frequency scaling.

Layered system decomposition aids in implementing these mechanisms.

Abstract

With the introduction of various hardware/software technologies such as Cloud Technologies or Virtualization technologies, there has been a great potential to reuse ICT artifacts thanks to Abstraction and also Exchangeability features achieved via these technologies. These technologies also provide various advantages with respect to sustainability including resource consumption reduction (in the use phase only or in the whole life cycle). However, there is an additional but untapped potential associated with the anonymization of resources introduced by both abstraction and exchangeability features. By realizing on this potential, we can improve cloud solutions and reduce their by-product opacity, which usually prevents leveraging on the specialized but tweakable (i.e., nonessential modifications without changing the main function) features of components that are captured in the component models. This is especially a challenge in the case heterogeneous/disaggregated infrastructure where developing models to cover everything is practically impossible. In this work, by leveraging on the concept of pathways, we develop a few mechanisms that enable transparency and therefore tweakability of features even in the presence of abstraction and heterogeneity. In particular, the layered-stack approach to system decomposition is considered because of its role in both software defined networking (SDN) and Network Function Virtualization (NFV) system decompositions. For a concrete example, the case of dynamic frequency scaling of processors is considered and it is shown that the associated consumption could be considerably reduced without requiring additional changes to the middle components.