Application Of Alpha Stable Distribution Model In The Classification Guarantee Of Mobile Data Service

As the coming of 3G era, the development of mobile communications and the rapid growing of data service traffic, the development of communication technology has lagged behind. To some extend, there are not enough available network resources, which will has some adverse effect to QoS (quality of service). Currently, Mobile Operators has set the same QoS level for different APNs (CMWAP, CMNET, CMMM and other APN), different types of service (high-value business and general business) and different brands (Gotone, M-Zone and Easyown). There is no distinction between them, so there is no reasonable disposal of existing network resources.In order to achieve different QoS of different APN, types of services and brands, and use the network resources more efficiently. This paper has conducted the following research:1. Establish multi-dimensional QoS system for GPRS data services. First, provide different quality of service to users of different brands; in the same time, protect the quality of service of high-value traffic, and use network resources more rationally and efficiently.2. Explore flow characteristics of mobile data services in depth. In the premise of taking full account of the pulse characteristics and heavy tail of mobile core network data traffic, we adopt Alpha stable distribution model as a self-similar source model to propose a bandwidth allocation model for mobile data network.3. Propose a for multi-dimensional data services in the view of APN, traffic types and user brands respectively. And taking into account the implementation difficulty of the three dimensions and the effect of users, specific implementation strategies of hierarchical security plan is given:(1) Implement the security programs based on APN and set different QoS for different APN in HLR.(2) Implement the security programs based on user brands and set different QoS contract value for user brands in HLR.(3) Implement the security programs based on traffic types and limit the bandwidth of ordinary business. In busy hours, give priority to the bandwidth needs of high-value business, and the traffic peak of ordinary business.This research achievement of the paper is specific strategy which can be deployed in the GPRS network to reduce processing resources consumption of GGSN. This method can effectively distinguish the quality of different services and guarantee the bandwidth demand of the high-value service and improve GPRS network bandwidth utilization.