Reliability and Local Delay in Wireless Networks: Does Bandwidth Partitioning Help?

TL;DR

This paper analyzes how bandwidth partitioning affects reliability and delay in wireless networks, revealing a tradeoff where BWP can improve reliability but may increase delay depending on the approach.

Contribution

It provides closed-form expressions for local delay and reliable transmission density under BWP in the ultrareliable regime, highlighting the reliability-delay tradeoff.

Findings

BWP reduces interference but increases SIR requirements or decreases data rate.

In the ultrareliable regime, BWP's impact depends on the SIR approach used.

A fundamental tradeoff exists between reliability and delay when applying BWP.

Abstract

This paper studies the effect of bandwidth partitioning (BWP) on the reliability and delay performance in infrastructureless wireless networks. The reliability performance is characterized by the density of concurrent transmissions that satisfy a certain reliability (outage) constraint and the delay performance by so-called local delay, defined as the average number of time slots required to successfully transmit a packet. We concentrate on the ultrareliable regime where the target outage probability is close to 0. BWP has two conflicting effects: while the interference is reduced as the concurrent transmissions are divided over multiple frequency bands, the signal-to-interference ratio (SIR) requirement is increased due to smaller allocated bandwidth if the data rate is to be kept constant. Instead, if the SIR requirement is to be kept the same, BWP reduces the data rate and in turn increases the local delay. For these two approaches with adaptive and fixed SIR requirements, we derive closed-form expressions of the local delay and the maximum density of reliable transmissions in the ultrareliable regime. Our analysis shows that, in the ultrareliable regime, BWP leads to the reliability-delay tradeoff.