Name
Title | ||
|---|---|---|
Wireless Networks Revenue Optimization through Overload Control with Priority Services |
Abstract | ||
|---|---|---|
Wireless networks are currently experiencing more overload situations than their wireline counterparts because of explosive mobile traffic growth, unpredictable traffic behavior, service differentiated traffic shedding, etc. Even though extensive research work on network overload control in general is going on, the economic aspect of the overload problem has received very little attention. Managing the incoming traffic in a way that generates the maximum possible revenue under overload warrants special attention. In this paper, we study a realistic wireless switch overload model where message exchange and message discarding at multiple nodes are considered. We propose a distributed overload control framework considering dif- ferent service types to obtain the optimal revenue while maintaining the switch's capability to handle call attempts during overload situation. This is achieved by exchanging overload information among the nodes that can have a global view of the switch-wide overload situation and its impact on revenue, and hence can adjust their own traffic shedding to improve the revenue generation. Next, we extend the proposed framework to incorporate different call priorities and discuss the conditions for reaching the optimal revenue that ensures preferential treatment to the high priority service types. In most vendors' implementations, a wireless switch is no longer a single node, but consists of multiple nodes that work together to handle the user traffic and the signalling. A call1 setup procedure includes several rounds of message exchanges among these nodes, while overload traffic shedding could happen in any of these nodes. However, in most of the existing literature, a call setup process in a node is only modeled as a single message processing. There has been little analysis on the overload performance where the message flows are considered. In this paper, we address the overload control in the context of a multi-node wireless switch, with the purpose of optimizing the revenue generation while maintaining a fixed bound on the delay at each node. In particular, we propose a distributed overload control mechanism that can be applied to wireless (and wireline) switches. The rest of the paper is organized as follows: Section II briefly discusses the state-of-the-art in single/multi-class overload control algorithms. Section III outlines a multi- node wireless switch model, upon which our overload control framework is applied. Section IV gives detailed description of the proposed overload control algorithms where we do not consider call priorities. Section V discusses the implementation issues and VI presents the simulation results to show that our proposed framework maintains key performance metrics of a switch under overload while improving the revenue generation. In Section VII, we present an extension of the distributed framework that will support call types having different priorities. Section VIII concludes the paper. |
Year | DOI | Venue |
|---|---|---|
2006 | 10.4304/jcm.1.4.22-30 | JCM |
Keywords | Field | DocType |
nonlinear optimization,priority services.,index terms— distributed control framework,revenue maximization,indexing terms,wireless network | Revenue,Wireless network,Wireline,Wireless,Mobile traffic,Computer science,Nonlinear programming,Computer network,Overload control,Revenue optimization,Distributed computing | Journal |
Volume | Issue | Citations |
1 | 4 | 1 |
PageRank | References | Authors |
0.38 | 14 | 3 |
Name | Order | Citations | PageRank |
|---|---|---|---|
| Hai-Tao Lin | 1 | 225 | 23.38 |
| Preetam Ghosh | 2 | 349 | 43.69 |
| Prabir Das | 3 | 16 | 5.20 |