適地性服務的分散式隱私保護策略

A Decentralized Privacy Protection Strategy for Location-based Services

陳智群(Chih-Chun Chen)

指導教授 ： 薛幼苓

國立中正大學/工學院/資訊工程學系/碩士(2014年)

摘要

鑒於雲端計算整合適地性服務的快速發展下，近來開始引發了許多的注意。科技的進步使得有越來越多的使用者擁有智慧型手機，也因為如此越來越多使用者周遭的訊息和感興趣的事物已經變得廣泛為人所使用。然而，大眾開始會顯著注意及聚焦個人隱私的議題。現有的資料管理方法主要強調於集中式管理方法，但這將會引發極大的安全疑慮。集中式管理的方法只利用單一查詢處理器 (亦稱之為適地性服務提供者)，這將會引發安全上的顧慮，因為單一個服務提供者是非常可能被惡意的使用者入侵。為了保護查詢處理器及避免此種威脅的發生，我們在這篇論文提出了一個分散式隱私保護策略 (A Decentralized Privacy Protection Strategy (DPPS))，來保護這些實際向適地性服務伺服器 (location-based services) 發送查詢的敏感資訊。在模糊區域的階段，我們提出了兩個模糊區域機制的方法: 一個是以隨機 (Random-based) 為基準，而另一個是以希爾伯特曲線 (Hilbert-curve-based) 為基準。我們的系統以一個區域來模糊化使用者的查詢點。 除此之外，我們切割原本的模糊區域成好幾個子區域，並分散這些子區域至不同查詢處理器，以增進隱私保護。實驗的結果顯示，我們的演算法產生了一個合理的遮蔽區域，並且保持了一個具備有競爭性的效能和均勻的分散度。

關鍵字

安全性與隱私 ； 匿名化 ； 使用者隱私保護 ； 適地性服務

並列摘要

The rapid development of the integration of cloud computing and location-based services has drawn much attention recently. With the increasing number of users who own smart phones, a large amount of data that describe user surrounding information and interests have become widely available. However, significant attention has been focused on personal privacy issues. The existing approaches to data management mainly focus on a centralized approach which raises tremendous security concerns. A centralized approach that only utilizes a location-based services(LBS) provider(also termed as a services provider) presents serious security concerns, because a single service provider is very likely to be hacked by malicious users. To prevent a centralized query processor from this threat, we propose a decentralized privacy protection strategy (DPPS) with k-anonymity and dummy techniques to protect the sensitive location information of users who request location-based services. In the cloaked region generation phase, we propose two cloaked region mechanisms: a random-based and a Hilbert-curve-based approaches. Our system obfuscates the query position of a user into a cloaked region. Furthermore, to enhance privacy protection, we partition the cloaked region into several subregions and distribute them to different service providers. Our experimental results show that our algorithm retains a uniform subregion distribution among all server providers and meanwhile, maintains a competitive performance.

並列關鍵字

security and privacy ； k-anonymity ； user privacy protection ； location-based services

參考文獻

[1] Platform for privacypreferences (p3p) project. In W3C, 2011.

[2] R. Agrawal, J. Kiernan, R. Srikant, and Y. Xu. Order-preserving encryption for numeric data. In SIGMOD Conference, pages 563–574, 2004.

[3] A. R. Beresford andF. Stajano. Location privacy in pervasive computing. IEEE Pervasive Computing, 2(1):46–55, Jan. 2003.

[4] A.R. BeresfordandF. Stajano. Mix zones: Userprivacyin location-aware services. In PerCom Workshops, pages 127–131, 2004.

[5]R.Cheng,Y.Zhang,E. Bertino,andS. Prabhakar. Preserving user locationprivacy in mobile data management infrastructures. In Privacy Enhancing Technologies, pages 393–412, 2006.