A Two-Stage Auction Mechanism for Cloud Resource Allocation

TL;DR

This paper proposes a two-stage auction mechanism for cloud resource allocation in cloud of clouds networks, combining an options-based auction for customer interactions and market analysis for CCN manager negotiations.

Contribution

It introduces a novel two-stage framework with an options-based auction for resource allocation and analyzes market behaviors between CCN managers and cloud providers.

Findings

The options-based sequential auction ensures truthfulness and addresses entrance timing issues.

Analytical foundations and performance metrics for OBSAs are established.

Market analysis for CCN managers' bidding behavior is provided.

Abstract

With the recent growth in the size of cloud computing business, handling the interactions between customers and cloud providers has become more challenging. Auction theory has been proposed to model these interactions due to its simplicity and a good match with real-world scenarios. In this paper, we consider cloud of clouds networks (CCNs) with different types of servers along with customers with heterogeneous demands. For each CCN, a CCN manager is designated to handle the cloud resources. A comprehensive framework is introduced in which the process of resource gathering and allocation is addressed via two stages, where the first stage models the interactions between customers and CCN managers, and the second stage examines the interactions between CCN managers and private cloud providers (CPs). For the first stage, an options-based sequential auction (OBSA) is adapted to the examined market, which is capable of providing truthfulness as the dominant strategy and resolving the entrance time problem. An analytical foundation for OBSAs is presented and multiple performance metrics are derived. For the second stage, two parallel markets are assumed: flat-price and auction-based market. A theoretical framework for market analysis is provided and the bidding behavior of CCN managers is described.