應用在網頁應用程式的身分驗證流程的自動化模型

Abstract

登入，是一個需要向應用程式輸入自己的身分證明並提交給系統的流程。系統會記錄此行為並提供相應的服務和操作權限。在現今的網頁應用程式中，擁有登入與註冊模組的身分驗證系統已經非常常見。而在應用程式開發的過程中，進行測試是一個必要的環節。現今的測試方式主要分為兩種：人工手動測試、撰寫腳本進行測試。兩種都會需要花大量的人力資源來完成。在本論文中，我們提出了一個叫 AAPA的自動化的方法來通過登入與註冊的流程。透過自動建構的 Automata，AAPA能夠實時的分析當前遇到的頁面來選擇最有可能完成流程的行動，不再只是一連串的固定觸發流程。並且，我們整合了讀取電子郵件和簡訊的認證碼之流程，以擴大自動化的涵蓋範圍。最後，我們進行了實驗來評估我們方法的有效性，實驗證明了這樣的方式能夠讓流程更為彈性。

Logging in is a process that requires you to submit your identity to the system. The system records this action and provides the corresponding service and operation privileges. In today's web applications, identity verification systems with login and registration modules are very common. In the process of application development, testing is necessary. Nowadays, there are two mainstream testing methods: manual testing and scripting. Both of them require a lot of human resources to complete. In this paper, we propose an automated method named AAPA to go through the login and registration process. With the automatically constructed Automata, AAPA can analyze the current page encountered in real time to select the most likely action to complete the process, instead of just a series of fixed triggers. Moreover, we integrate the process of reading the verification code and URL of e-mail and SMS to expand the scope of automation. Finally, we conduct experiments in order to evaluate the effectiveness of our approach. This approach proved to make the process more robust.