Jin-Ghoo Choi

Cell-Throughput Analysis of the Proportional Fair Scheduler in the Single-Cell Environment

The fairness concept has been widely studied in the area of data networks. The most well-known fairness criterion (max-min fairness) gives priority to the minimum-rate session. Kelly questioned its appropriateness in his works on the bandwidth sharing among the end-to-end flows and proposed another fairness criterion preferring short-distance flows to enhance the overall throughput, which is called the proportional fairness (PF). A simple scheduler achieving this objective was introduced in wireless access networks and revealed that it can achieve a good compromise between cell throughput and user fairness. Although it has received much attention for some time, research on its performance mainly depended on computer simulations. In this paper, we analyze the PF scheduler to obtain the cell throughput, which is a primary-performance metric, and extend the result to analyze the capacity of multiple-input-multiple-output systems. We evaluate the effect of various parameters on the throughput of the PF scheduler through the numerical analysis

Realistic cell-oriented adaptive admission control for QoS support in wireless multimedia networks

An important quality-of-service (QoS) issue in wireless multimedia networks is how to control handoff drops. We propose admission-control algorithms that adaptively control the admission threshold in each cell in order to keep the handoff-dropping probability below a predefined level. The admission threshold is dynamically adjusted based on handoff-dropping events. We first present a simple admission-control scheme that brings out an important performance evaluation criterion - intercell fairness - and serves as a reference point. We then investigate the intercell unfairness problem and develop two enhanced schemes to overcome this problem. The performance of these protocols is benchmarked and compared with other competitive schemes. The results indicate that our schemes perform very well while, in addition, achieving significantly reduced complexity and signaling load.

Cell Throughput Analysis of the Proportional Fair Scheduling Policy

The fairness concept has been widely studied in the area of data networks. The most well-known fairness criterion, max-min fairness, gives priority to the minimum rate session. Kelly questioned its appropriateness in his works on the bandwidth sharing among the end-to-end flows and proposed another fairness criterion preferring short distance flows to enhance the overall throughput, which is called the proportional fairness (PF). A simple scheduler achieving this objective was introduced in wireless access networks and revealed that it can achieve a good compromise between cell throughput and user fairness. Though it has received much attention for some time, research on its performance mainly depended on computer simulations. In this paper, we analyze the PF scheduler to obtain the cell throughput which is a primary performance metric.

Multiclass call admission control in QoS-sensitive CDMA networks

In this paper, we consider a CDMA cellular network that has signal power as its scare resource and uses the multicode transmission scheme to support multiclass calls with different QoS requirements. By simply extending the number-based scheme to accommodate multiclass traffic, we present a power-based call admission control scheme. Contrary to the general fact that complete sharing (CS) achieves better utilization than complete partitioning (CP), our finding in this paper is somewhat unique. That is, the CP of signal power for each traffic class can accommodate a reasonably high number of users when compared with CS. This result allows us to use CP which is much easier than CS in terms of resource management. By adopting the CP technique for resource reservation in wireless CDMA networks, we can guarantee QoS of calls in terms of handoff drops with minimal complexity.

Tier Based Anycast to Achieve Maximum Lifetime by Duty Cycle Control in Wireless Sensor Networks

In a wireless sensor network, the routing control overhead is prohibitive because mostly many nodes are involved in routing. To overcome the overhead problem, tier based anycast protocol was proposed in (Kulkarni et al., 2006). However, it has a shortcoming of using much more energy compared to other competitors of deterministic routing protocols in transmitting data packets towards the sink node. In this paper, we visited the tier based anycast protocol in depth and presented a new analytic framework by using the concept of sub-tiering. Firstly, we formulated the problem for finding an optimal duty cycle for each tier as a minimax optimization problem, and proved that the solution exists and it is unique. From the analysis results, we found that the network lifetime can be improved very much by allocating a duty cycle adaptively for each tier. Through simulations, we also confirmed that our duty cycle control algorithm increases the network lifetime by approximately 30%.

Hierarchical Markov chain analysis of an adaptive bandwidth reservation algorithm in wireless communication systems

Admission control for handoff drops is a key resource management technique in wireless multimedia networks. While several adaptive schemes have been studied in the literature, little effort has been made for their analysis due to difficulty in mathematical modeling. Our main contribution is the proposal of a methodology for analyzing the Markov chain that is created when modeling a simple adaptive bandwidth reservation mechanism. For this purpose, we decompose the Markov chain into two levels: intra-domain for the analysis of equilibrium states and inter-domain for the analysis of transient states. This hierarchical decomposition provides an exact solution with significantly reduced computational complexity.

A backward-compatible multiple-round collision avoidance scheme for contention based medium access control

Random-access mechanisms play an important role in wireless networks, and have been extensively studied in recent years. Although many previous studies have proposed enhanced algorithms, each one has only considered either throughput or fairness. In this paper, we propose an efficient random-access mechanism called Multi-round Collision Avoidance (MrCA) that considers throughput and fairness together. The key idea in MrCA is to avoid collisions through multiple contentions, each with a smaller sized contention window. With this simple modification, we can significantly reduce the collision probability as well as the access delay, in addition to increasing fairness index. We find the collision probability and throughput analytically. Through simulation, we validate our analytical model and find appropriate parameters for achieving good performance. We also demonstrate that, compared to the IEEE 802.11 DCF, MrCA makes the collision probability extremely low, so that it increases throughput by 25% as well as short-term fairness by 50% with 50 contending nodes. When MrCA and 802.11 DCF schemes are combined with the auto rate fallback scheme, the performance gain of MrCA over 802.11 DCF increases because MrCA lowers the collision probability significantly, which makes channel error estimation more accurate. We also discuss the issues of implementation and backward compatibility.

Channel aware MAC scheme based on CSMA/CA

The knowledge of channel states can improve the performance of wireless communications systems, and its benefits are outstanding where the channel states vary rapidly with time. The fast feedback of channel information was realized in the downlink of cdma2000 1/spl times/EV-DO system and it shows great enhancement in terms of total average throughput. However, it does not scale well since the burden of feedback channels increases proportionally with the increase of users in number. We propose a multiple access scheme based on CSMA/CA, which gives high priority to users in the good channel by assigning small backoff counters. Since the channel gain is used to choose the backoff counter locally, it is not sent back to the central controller, thereby achieving low overhead and complexity. Our proposed channel aware MAC scheme adapts to traffic conditions dynamically by changing the size of contention window when collision occurs.

Opportunistic scheduling for utility maximization under QoS constraints

Opportunistic scheduling provides an important capability of resource management for wireless systems by taking advantage of multiuser diversity, allowing delay variation in delivering data packets. Our opportunistic scheduling aims at maximizing the utility which is usually expressed as a function of user throughput. In this paper we derive an off-line optimal scheduling policy by optimization theory, and show that the policy is also optimal in opportunistic scheduling. We extend the scheduling policy by adding QoS requirements for each user. To do so, we classify opportunistic schedulers according to QoS constraints. Their optimalities are also proved through mathematical analysis. To obtain optimal scheduling policy, the parameters related to QoS should be computed, so we propose a practical algorithm for parameter adaptation and verify its performance through simulations. The results confirm that our considered opportunistic schedulers show good fairness performance and guarantee each users QoS without any difficulty in implementation

Utility-based downlink power allocation in multicell wireless packet networks

We introduce a utility-based radio resource management technique for multicell wireless packet networks. Assuming each base station operates to maximize its own cell utility, we try to maximize the system utility which is defined as the sum of each cell utility. By using the downlink transmit power of each base station as a control variable, we formulate a problem of system utility maximization. However, the problem is intractable because it contains non-convex functions, and, accordingly, the complexity is too high. Therefore, we first analyze a simple two-cell problem and take an intuitive approach to find its optimal solution. We extend this intuition to a general multi-cell problem and devise a simple heuristic algorithm that requires a light overhead of signaling between neighboring base stations. Through simulations, we compare, in terms of utility gain, our proposed heuristic algorithm with two competing schemes, optimal power allocation and maximum power allocation. As expected, the optimal allocation scheme shows the best performance but cannot be applied to real multi-cell problems due to its intractable complexity. Our proposed heuristic algorithm has very low complexity and its performance lies between the optimal allocation and maximum allocation schemes.