[Paper Review] On the Capacity Region of ALOHA Protocol for the Internet of Things

Moslem Noori, Samira Rahimian+1 more|arXiv (Cornell University)|May 27, 2016

IoT Networks and Protocols11 references3 citations

TL;DR

This paper proposes a framework to characterize the capacity region of irregular repetition slotted ALOHA (IRSA) in heterogeneous Internet of Things (IoT) networks with multiple user classes. By introducing the concept of a dual homogeneous network, the authors show that the capacity region of a multi-class network is fully determined by the maximum achievable throughput of its dual network, enabling precise control of class-specific throughputs through user activation and optimal IRSA parameter selection.

ABSTRACT

Accommodating the needs of a large number of diverse users in the Internet of Things (IoT), notably managing how the users access the common channel, has posed unique challenges to the network designers. In this paper, we study a heterogeneous IoT network consisting of multiple classes of users who may have different service requirements. For this network, we consider the application of irregular repetition slotted ALOHA (IRSA) that is shown to offer large throughput for single-class networks. Then, we focus on finding the network performance boundaries by studying the set of feasible throughput values for each class, called the capacity region. To this end, we first introduce the concept of dual network of a multi-class network meaning a homogeneous network with the same number of users. We then prove that finding the capacity region of the assumed multi-class network boils down to finding the maximum achievable throughput of its dual network. Using this finding, we then discuss how any given point of the capacity region can be achieved. Further, a delay performance study is conducted to evaluate the average and maximum packet transmission delay experienced by the users of each class.

Motivation & Objective

• To address the challenge of managing diverse service requirements in heterogeneous IoT networks with sporadic, uncoordinated traffic.
• To define and characterize the capacity region—the set of feasible throughputs—for multiple user classes sharing a common channel via IRSA.
• To establish a theoretical link between the capacity region of a heterogeneous multi-class network and a homogeneous dual network with the same total number of users.
• To enable the design of user activation and IRSA parameter strategies that achieve any desired point within the capacity region.
• To analyze the impact of user activation and IRSA configuration on average and maximum packet transmission delay.

Proposed method

• Introduce the concept of a dual homogeneous network: a single-class network with the same total number of users as the original multi-class network.
• Prove that the capacity region of the multi-class network is determined by the maximum achievable throughput of its dual network.
• Use the user degree distribution $\Lambda_i(x)$ per class to model transmission behavior, with each user randomly selecting $l$ slots to transmit $l$ replicas.
• Apply successive interference cancellation (SIC) at the base station to resolve packets from singleton and newly revealed singleton slots after interference cancellation.
• Define traffic load $G_i = L_i/M$ for each class $i$, and express the throughput $T_i(\boldsymbol{G}) = S_i(\boldsymbol{G})/M$ as a function of the traffic load vector $\boldsymbol{G}$.
• Derive analytical expressions for the slot degree distribution $\Psi(x)$ and use them to compute the achievable throughput under SIC, leveraging results from single-class IRSA analysis.

Experimental results

Research questions

• RQ1 What is the set of all feasible throughput combinations that can be achieved by different user classes in a heterogeneous IoT network using IRSA?
• RQ2 How can the capacity region of a multi-class IRSA network be characterized in terms of system parameters and user behavior?
• RQ3 Can the capacity region of a heterogeneous multi-class network be derived from a simpler homogeneous dual network?
• RQ4 What user activation and transmission strategy are required to achieve any given point within the capacity region?
• RQ5 How do user activation patterns and IRSA parameters affect the average and maximum packet transmission delay?

Key findings

• The capacity region of a multi-class IRSA network is fully determined by the maximum achievable throughput of its dual homogeneous network, which simplifies the analysis of heterogeneous systems.
• Any point within the capacity region can be achieved by selecting the appropriate number of active users from each class and applying the optimal IRSA scheme derived from the dual network.
• The maximum achievable throughput of the dual network directly sets the upper bound on the sum of throughputs across all classes in the multi-class system.
• The average and maximum packet transmission delays are significantly influenced by the user activation strategy and the choice of repetition distribution in IRSA.
• Numerical results confirm that the proposed framework enables precise control of class-specific throughputs while maintaining low delay, especially when the dual network's optimal parameters are used.
• The framework enables the design of QoS-aware access strategies where high-priority classes can be assigned higher throughputs by adjusting their activation levels and repetition patterns.