自主式社群網路身分的設計與實作

Abstract

近年來社群網站蓬勃發展，Google 和 Facebook 等公司透過搜集分析與操作網際網路上的使用者資料，提供了許多智慧化的服務與資訊。不過，這些本來是私人的資料逐漸被這些公司所把持，其中包括許多個人的隱私。於是，不僅難以確保隱私，也難以在不同公司間的各種網路服務分享與修改個人資料。另一方面，社群網站上存在許多假的甚至惡意的資訊，不僅對使用者造成困擾，也浪費了許多社會成本來處理。因此，本論文提出了自主式的社群網路身分 (Aid)，讓使用者能夠完全掌握自己的資料和隱私。我們將 Aid 從網路服務中獨立出來，讓大部份的資訊儲存在本地，並透過數位簽章的機制進行身分驗證。此外，使用者能夠真正自主的前提是使用者要能先為自己的社會活動負責。因此，我們也提出了一個自主的機制來評估如信用度或真實度等 Aid 的度量，讓使用者可以利用這些度量彼此量化評估。目前我們已經實作出 Aid 的雛型以及給服務使用的應用介面，提供多個學校的實驗室開發相關的自主網路服務。透過『自主量化』的 Aid，相信網路身分能變得更加自主可靠，進而提升社群網路活動的品質。

As the booming of social network recently, companies such as Google and Facebook provide intelligent services and information by collecting, analyzing, and manipulating user data on the Internet. However, these originally private data which contain lots of privacy gradually become dominated by these companies. Thus, not only might the privacy not be ensured but it is also difficult to share and update personal data among services of different companies. On the other hand, much information is fake or even malicious on social network. It not only confuses users but also wastes lots of social cost to handle. Hence, this thesis proposes an autonomous identity (Aid) for social network to provide users with total control of own data and privacy. Aid is designed independent from network services, most of Aid information is stored locally, and it is authenticated by digital signature. Moreover, users need to be responsible for their social activities before they can be really autonomous. Therefore, we also propose an autonomous mechanism to evaluate metrics of Aid, such as credit or reality, so that users can evaluate one another by these metrics quantitatively. We have implemented the prototype of Aid and application programming interfaces for services, offered some laboratories of some universities to develop autonomous network services. By Aid of 'autonomous quantification', we believe network identities become more autonomous and reliable to increase the quality of activities in social network.