Modeling and Analysis of Converged Network-Cloud Services

TL;DR

This paper develops an analytical model to evaluate tradeoffs in cloud network design, balancing resource centralization benefits against communication costs, with a case study in the oil and gas industry.

Contribution

It introduces a novel analytical framework for assessing the tradeoffs in converged network-cloud environments considering application needs and costs.

Findings

Centralization offers multiplexing benefits but increases communication costs.

The model helps optimize resource placement based on specific tradeoffs.

Case study demonstrates practical application in the oil and gas sector.

Abstract

Networks connecting distributed cloud services through multiple data centers are called cloud networks. These types of networks play a crucial role in cloud computing and a holistic performance evaluation is essential before planning a converged network-cloud environment. We analyze a specific case where some resources can be centralized in one datacenter or distributed among multiple data centers. The economy of scale in centralizing resources in a sin- gle pool of resources can be overcome by an increase in communication costs. We propose an analytical model to evaluate tradeoffs in terms of application requirements, usage patterns, number of resources and communication costs. We numerically evaluate the proposed model in a case study inspired by the oil and gas industry, indicating how to cope with the tradeoff between statisti- cal multiplexing advantages of centralization and the corresponding increase in communication infrastructure costs.