雲端儲存連結閘道

Abstract

隨著網路技術與雲端技術的發展，Web 2.0 服務已經轉化成使用者角色為資訊產生者，使用者將上傳的資料交給不同的網路服務，藉由每家網路服務的伺服器端統一管理使用者的資料，並且儲存於公司內部的儲存設備中，每家公司必須花大量成本購買可確保使用者資料可靠性的硬體設備，對公司是一種負擔。 現在雲端儲存服務愈來愈流行，例如Dropbox、Google Drive及微軟的OneDrive，對於使用者而言，能隨時隨地透過網路即可存取自己的資料，而對於網路服務提供者而言，可以使用目前已提供的雲端儲存API(Application Programming Interface)，來將資料儲存到使用者的雲端空間中，減少自身在儲存空間上的成本支出，但每家雲端儲存空間都有各自的API連接方式，導致網路服務提供者需要額外支出成本去了解與設計不同的連接方式。 本研究發展出一個基於CDMI(Cloud Data Management Interface)標準格式的雲端儲存連結閘道，可以解決Dropbox、 Google Drive及微軟的OneDrive三家雲端儲存空間使用不同API形式的困擾。一般網路服務後端都會有資料庫管理使用者的資料，本研究也發展出透過連結閘道可創立與操作雲端儲存空間上的關聯式資料庫檔案與非關聯式資料庫檔案的API，能讓使用者將資料儲存到資料庫檔案中並且儲存至使用者自己的雲端儲存空間中。除了能夠讓網路服務提供者減少維護儲存空間的成本，使用者也能管理自己的資料。

Web 2.0 services have been let users become an information producer and distributor. Users upload their data to different web services. The data are managed by server-side of each web service and then they are stored in the business's internal storage devices. Every businesses must spend a lot of cost purchasing storage devices that can ensure reliability of data. Therefore, this situation lays big burden on them. Cloud storage services (CSSs) such as Dropbox, Google Drive and Microsoft’s OneDrive are more and more popular. For users, anytime and anywhere they can access their data through the Internet. On the other hand, for web service providers (WSPs), they can store data on user’s cloud storage by using CSSs’ API and reduce their costs on storage devices. But each CSS has its own connection method of cloud storage’s API, WSPs must spend additional time understanding their APIs. In this thesis, we propose a CDMI-compliant cloud storage access gateway to solve Dropbox, Google Drive and OneDrive three different connection methods of APIs. Moreover, WSPs usually use a database to manage user-generated data, so we propose two APIs on the access gateway that can operate the relational database file and non-relational database file on CSSs. By using them, user-generated data are stored on the database file and then it can be uploaded to users’ cloud storage. WSPs can reduce the cost of maintaining storage space and users can manage their own data.