Prediction Based Resource Allocation
In a dynamic environment, call arrivals and departures in different service areas may trigger reallocations for all MTs in service. In a decentralized architecture, this is translated to a heavy signalling overhead between the MTs and different BSs/APs wi
- PDF / 584,787 Bytes
- 21 Pages / 439.37 x 666.142 pts Page_size
- 106 Downloads / 247 Views
Prediction Based Resource Allocation
In a dynamic environment, call arrivals and departures in different service areas may trigger reallocations for all MTs in service. In a decentralized architecture, this is translated to a heavy signalling overhead between the MTs and different BSs/APs with every call arrival and/or departure in any service area. Hence, the main challenge is how to develop an efficient decentralized radio resource allocation mechanism that reduces the associated signalling overhead with call arrivals and departures. In this chapter, concepts of call traffic load prediction and network cooperation are introduced to address the challenges that face the decentralized resource allocation in a dynamic environment.
3.1 Introduction In Chap. 2, the DORA algorithm is presented to support MTs with multi-homing capabilities in a heterogeneous wireless access medium. The DORA algorithm mainly identifies the role of different entities in the heterogeneous wireless access medium in order to enable a decentralized architecture. Specifically, the main role of a network, n ∈ N , BS/AP, s ∈ Sn , in the decentralized architecture is to update a link access price value (λns ) that indicates the capacity limitation experienced by this BS/AP. On the other hand, the main role of an MT, m, is to update its coor(1) (2) dination parameter(s) (νm for MT with CBR service, or μm − μm for MT with VBR service) in order to satisfy its required bandwidth. Both link access price values for different BSs/APs and coordination parameter(s), together with the priority parameter pnms , determine the allocated resources from each network BS/AP so as to satisfy the MT total required bandwidth. The DORA algorithm is an iterative one that relies on signalling exchange between an MT and different BSs/APs in order to reach the optimal resource allocation from each BS/AP to the MT. This includes the exchange of the current iteration MT coordination parameter (from MT to BSs/APs) and the corresponding BS/AP resource allocation bnms (from each BS/AP to MT).
M. Ismail and W. Zhuang, Cooperative Networking in a Heterogeneous Wireless Medium, SpringerBriefs in Computer Science, DOI: 10.1007/978-1-4614-7079-3_3, © The Author(s) 2013
37
38
3 Prediction Based Resource Allocation
The DORA algorithm is proposed for a static environment without arrival of new calls or departure of existing ones. As illustrated in Fig. 2.9, due to the complimentary slackness condition for (2.8), we have the following observations: 1. When the total call traffic load ( m∈Mns bnms ) carried by network n BS/AP s is less than the BS/AP transmission capacity limitation Cn , the corresponding optimal link access price value λ∗ns = 0. This results in allocating the maximum required bandwidth for all VBR calls under this BS/AP jurisdiction. 2. When the carried call traffic load reaches the BS/AP transmission capacity limitation, λ∗ns > 0. Hence, the allocated bandwidth to each of the VBR calls in service is reduced towards the call minimum required bandwidth so as t
Data Loading...
