Big data over wireless pipe-Path: Queuing model for multi-Directional radio links

Abstract

Big data needs to be delivered to remote devices via wireless links in many applications such as North Pole monitoring, UAV communications, smart city crowd-sourcing, etc. Single routing path may not be able to meet the ultra-high-throughout requirements of wireless big data delivery. A pipe-like path consisting of closely coupled links (forming a bundle) is more suitable to big data transmissions. The queuing model for such a coupled routing pipe is important to network congestion diagnosis and avoidance. Particularly, the queuing analysis between neighboring hops can help to analyze the pipe traffic accumulation dynamics. Our pioneering work presented here first develops a queuing model for a single node of the pipe. We design a full Markov-Chain-based queuing model for bulk arrival, bulk service queues with multiple vacations. We then extend such a single node queuing model to the pipe-based queuing scenarios, and deduce the closed-form equations for the typical queue statistics such as the service delay. The numerical results can serve as the guidelines for big data transmission protocols. Comprehensive, accurate simulations have been conducted to verify the correctness of our proposed pipe queuing models.