1.1 Research Background and Motivation
The design of modern wireless mobile networks is based on a cellular architecture that allows efficient use of the limited available spectrum, such as GSM, GPRS, and 3G. The cellular architecture consists of a number of base stations, and mobile units; the mobile units communicate with other ones through the base stations via wireless links. The cell is a geographical region that a base station can serve, and the procedure when a mobile user moves from one cell to another is called handoff.
How to maintain the continuity and the required Quality of Service (QoS) when handoff occurs is a focal issue.
We can simply classify the incoming calls of a base station into two categories, new calls, and handoff calls. No matter what type of call, before a mobile user wants to communicate with other ones, it must first obtain a channel from one of base stations. If a base station doesn’t have available channel when the call is arriving, this call would be dropped. The action of dropping a handoff call causes abruptly force terminating an on-going call, and dropping a new call means rejecting a call in the first place. In users’ perspective, it can’t be tolerated that stopping a conversation rather than disallowing a session initialization. In wireless mobile circumstance, blocking probability of new calls and dropping probability of handoff calls are critical QoS parameters. Therefore, we need to keep the dropping probability of handoff calls under a required threshold to guarantee the QoS of handoff calls. In order to achieve the goal, handoff prioritization scheme is most often adopted.
There are two generic handoff prioritization schemes, Handoff Queuing Scheme (HQS) and Guard Channel Scheme (GCS) [10]. In HQS, handoff calls are queued in the new base station if no channel is available at the time of arrival. As soon as one channel is released in the new base station, it is offered to the handoff call in the queue. Besides, new calls are served only when a channel is available and no handoff calls in the queue. HQS decreases the dropping probability of handoff calls, but results in starvation of new calls and high system complexity.
1
Call Admission Control (CAC) provide the users with access to a wireless network for services. On the other hand, they are the decision making part of the network carriers with the objectives of providing services to users with guaranteed quality and increasing resource utilization as much as possible at the same time.
Therefore, it is conceivable that CAC policy is one of the critical design considerations in any wireless networks [33]. Another scheme is GCS, one kind of CAC, which assign higher priority to handoff calls by reserving a fixed number of channels for them, which called Guard Channel. In other words, handoff calls are served when any channel in base station is available, but new calls are accepted only when a channel except for reserved channels is available. GCS is developed under the assumption of stationary of call arrivals; it causes QoS degradation under non-stationary traffic patterns due to fluctuation of mobility.
As the variation of traffic condition, how to reserve channels for handoff calls is an important issue. Many researchers studied on it [1]-[5][7][8] [11]-[17][26]-[32]. In the present thesis, we proposed a two-phase adaptive call admission control scheme to dynamically deploy the number of guard channels in response to traffic fluctuations.
The phase one is to allocate guard channels based on a non-linear programming model subject to minimize the absolute difference between theoretical handoff call dropping probability and QoS threshold of it, then to operate in coordination with a channel adjustment mechanism in phase two for guaranteeing the QoS agreement of handoff call dropping probability. The main objective of our research is to guarantee the required dropping probability of handoff calls while keeping blocking probability of new calls as low as possible.
2
1.2 Approach and Result
The proposed scheme is built upon Guard Channel Scheme to prioritize handoff calls, and adaptively adjust the number of reserved channels as network condition changes. Due to limitation of fixed channels in a base station, reserving more channels for handoff calls will lead to a base station can serve less new call requests. So how to determine the number of guard channels reserved for handoff calls is a key problem.
In the thesis, we make use of a non-linear programming model to solve an adequate parameter for the number of guard channels. The proposed non-linear programming model is to minimize the absolute difference between theoretical handoff call dropping probability and QoS threshold. The effect of this soft constraint design may result in a slight violation of QoS agreement, but relatively provide more channel capacity for new calls. For making up for the QoS violation, another channel adjustment mechanism is adopted to automatically increase the number of guard channels when measured dropping rate exceeds a predetermined threshold.
The former mechanism, we call it channel allocation, and the latter is channel adjustment. By integrating these two components, a two-phase adaptive call admission control scheme is formed. The mechanism of channel allocation runs in phase one to lower the new call blocking probability, and channel adjustment resides in phase two to guarantee the QoS agreement.
Comparing with static GCS, present scheme can guarantee the dropping probability of handoff calls under a predefined QoS level, though this would result in the degradation of new call blocking probability. However, comparing with another adaptive CAC scheme [33], the handoff call dropping probability of our scheme is 5.28% lower than it in average, and the new call blocking probability of our scheme is 4.92% lower than it in average. The experiment result shows that the proposed scheme not only diminishes the dropping probability, but also keeps the blocking probability lower than it, which means that the proposed scheme can serve more new calls and handoff calls than another adaptive scheme.
3
1.3 Thesis Outline
In this chapter, we have generally introduced the research background, motivation, and the proposed scheme. The rest of the thesis is organized as follows. In Chapter 2, Call Admission Control, and related researches on CAC are introduced.
Details of the proposed two-phase adaptive CAC scheme are presented in Chapter 3.
In Chapter 4, we show the simulation result compared with Guard Channel scheme and another adaptive CAC scheme to point out the contribution of our research. At the last, the future work discussion and conclusion are made in Chapter 5.
4