線上收費系統於分享方案下之動態預約方法

Abstract

為了簡化計費流程與減少資料用量，近年來電信業者們所提出的分享方案蔚為風潮。基於第三代合作伙伴計畫(3GPP)所制定之標準，線上收費系統(OCS)必須於服務網路發送一段服務至使用者設備之前先計算並決定好一個核准量(Granted Unit, GU)，並為此次工作階段預留該核准量。然而固定的核准量並不合適於動態的使用者行為變化，因此需要動態地調整 每次給予的核准量。當給予的核准量太小，可能會造成訊息交換次數過於頻繁而在預留時產生大量的訊號量；而當給予的核准量太大，可能會造成不平均的資源分配問題。上述這些問題於分享方案底下會更為明顯。直至目前為止，如何決定核准量仍為一尚未解決的問題，電信業者可自行設計核准量的決定方法。本文基於同一分享方案底下使用者們的過去使用記 錄，提出一個新的方法來動態地分配每位使用者的核准量。實驗結果顯示，在不可預期之使用者行為使用環境底下，本文所提出之方法可較其他方法節省14%之訊號量。

For simplifying billing and limiting data usage, Shared Plan (SP) is now a popular plan provided by telecommunications operators. Based on the 3rd Generation Partnership Project (3GPP) specifications, Online Charging System (OCS) has to determine the Granted Unit (GU) and create reservations per session before a service is delivered by the serving network to a User Equipment (UE). However, fixed GUs cannot adapt to dynamic consumer behaviors, which usually require dynamic GUs. Clearly, giving small GUs will incur too many signaling messages for making reservations, but giving excessive GUs will cause unbalanced distribution of resources, especially in SP cases. How to determine GUs is an open issue that can be self-defined by telecommunications operators. In this thesis, we propose a new scheme for dynamically assigning GUs to UEs belonging to the same SP, based on their historical data usage and total monthly data allowance. Simulation results show that our scheme can substantially save signalings by 14% compared to other schemes under unpredictable behaviors.