Pull-based broadcasting with timing constraints

Abstract

The rapid growth in wireless and mobile computing has resulted in an increase in applications that would benefit from a broadcast system that considers timing constraints. Previously proposed scheduling algorithms for broadcast systems aim to minimize the mean access time, while our goal in this paper is to identify scheduling algorithms for broadcast systems that ensure requests meet their deadlines. We consider a pull-based approach in which users send explicit requests for data to a server. There are several well-known scheduling algorithms for real-time systems in non-mobile environments, and previous work in real-time mobile systems has utilized these scheduling algorithms to minimize missed deadlines. However, we demonstrate that traditional real-time algorithms do not always perform the best in a mobile environment and may perform worse than some non-real-time strategies. In this paper, we present a model of a pull-based real-time broadcast system and an analysis of the system as a Markov decision process (MDP). We also propose an efficient scheduling algorithm, called aggregated critical requests-b (ACR-b), which is designed for timely delivery of data to mobile clients. Our results indicate that ACR-b outperforms existing algorithms when timing constraints and the benefits of broadcasting are considered.

Scheduling on-demand broadcast with timing constraints

“…The values in this state window correspond directly to the values in the state matrix in Eq. (8). The only difference is that since we cannot consider more than K time slots into the future, it is optimal based on the current pending requests.…”

“…(8) in Box I, as a result of a Markov transition from state S(t) to S(t + 1), the first column of S(t) is eliminated since we now consider t +1. In the remaining columns the number of requests with deadlines is equal to the number of existing requests with deadlines plus the newly arrived requests with deadlines, except for the row in the matrix representing the page broadcast.…”

“…The Appendix presents the ACR-ω algorithm in more detail, and for further discussion and examples see [8]. In the next section, we present an algorithm with a simpler approach and a lower complexity.…”

“…This strategy is based directly on the state transition matrix in Eq. (8) and also on the expected β-discount cost defined in Eq. (10) in Section 3.…”

“…In this section we describe the model we used in [9] to study the on-demand broadcast scheduling problem with timing constraints. We begin by modeling the requests with deadlines and the broadcast system to satisfy them using the notation in Table 1.…”