智慧型網路服務控制點大容量資料庫環境之研究

Abstract

傳統交換機的軟體與機器相關（machine dependent），當欲開放一種新 服務時，因交換機的種類不一，更改方式各異，須耗時頗長才得以完成， 不利高度自由化、競爭的社會。如果將分散於各交換機之資料儘量集中於 一處，由電腦中心處理，並由全體網路分享.資料便於管理,更新 ,也易於 開發新服務,這就是智慧型網路.智慧型網路的智慧所在即是服務控制點（ ServiceControl Point），它具有幾個特點：大容量資料庫、各門號資料 互相獨立、快速的資料存取。為了滿足服務控制點大容量資料庫資料的快 速存取,本文提出六個主題討論:資料庫大小,區塊邊界,檔案長度,叢集( cluster),緩衝區,資料壓縮.為了提高緩衝區命中率，本文提出了下列方 法：指標專用、指標共用、緩衝區串列、位元映射、緩衝區共用、緩衝群 。經實驗證實，以緩衝群的緩衝區命中率最高。為了突破磁碟存取速度的 緩慢，未來研究方向可朝分散式系統及資料壓縮進行。 Software of traditional switches in telephone networks varies depending on switching hardware. It takes long time to update or modify the software in order to create a new service. If the service data are collected in a central database, it could be more convenient to manage and update.It is easier to create new services, too. In Intelligent Network, service data are collected on Service Control Point(SCP). The volume of data in the database is very large and data is independent of each other. Fast access of data is required to meet the users' expected response. For fast acces of data in the large database on Service Control Point, in this thesis we explore six design issues on size of database , boundary of blocks, size of file, cluster, buffering and data compressing. For prompting the hit ratio of buffering, we present six buffering mechanism: private pointer, shared pointer , buffer list, bitmap, shared buffer and buffer group. The hit ratio using buffer group is the best by simulation. To improve the scheme of disk access, distributed system and data compression can be further investigated in the future work.