Date of Award
Information Security, Location Privacy, Mobile Computing, Mobile Database, Private Information Retrieval
Usage of mobile services is growing rapidly. Most Internet-based services targeted for PC based browsers now have mobile counterparts. These mobile counterparts often are enhanced when they use user's location as one of the inputs. Even some PC-based services such as point of interest Search, Mapping, Airline tickets, and software download mirrors now use user's location in order to enhance their services. Location-based services are exactly these, that take the user's location as an input and enhance the experience based on that. With increased use of these services comes the increased risk to location privacy. The location is considered an attribute that user's hold as important to their privacy. Compromise of one's location, in other words, loss of location privacy can have several detrimental effects on the user ranging from trivial annoyance to unreasonable persecution.
More and more companies in the Internet economy rely exclusively on the huge data sets they collect about users. The more detailed and accurate the data a company has about its users, the more valuable the company is considered. No wonder that these companies are often the same companies that offer these services for free. This gives them an opportunity to collect more accurate location information. Research community in the location privacy protection area had to reciprocate by modeling an adversary that could be the service provider itself. To further drive this point, we show that a well-equipped service provider can infer user's location even if the location information is not directly available by using other information he collects about the user.
There is no dearth of proposals of several protocols and algorithms that protect location privacy. A lot of these earlier proposals require a trusted third party to play as an intermediary between the service provider and the user. These protocols use anonymization and/or obfuscation techniques to protect user's identity and/or location. This requirement of trusted third parties comes with its own complications and risks and makes these proposals impractical in real life scenarios. Thus it is preferable that protocols do not require a trusted third party.
We look at existing proposals in the area of private information retrieval. We present a brief survey of several proposals in the literature and implement two representative algorithms. We run experiments using different sizes of databases to ascertain their practicability and performance features. We show that private information retrieval based protocols still have long ways to go before they become practical enough for local search applications.
We propose location privacy preserving mechanisms that take advantage of the processing power of modern mobile devices and provide configurable levels of location privacy. We propose these techniques both in the single query scenario and multiple query scenario. In single query scenario, the user issues a query to the server and obtains the answer. In the multiple query scenario, the user keeps sending queries as she moves about in the area of interest. We show that the multiple query scenario increases the accuracy of adversary's determination of user's location, and hence improvements are needed to cope with this situation. So, we propose an extension of the single query scenario that addresses this riskier multiple query scenario, still maintaining the practicability and acceptable performance when implemented on a modern mobile device. Later we propose a technique based on differential privacy that is inspired by differential privacy in statistical databases. All three mechanisms proposed by us are implemented in realistic hardware or simulators, run against simulated but real life data and their characteristics ascertained to show that they are practical and ready for adaptation.
This dissertation study the privacy issues for location-based services in mobile environment and proposes a set of new techniques that eliminate the need for a trusted third party by implementing efficient algorithms on modern mobile hardware.
Eltarjaman, Wisam Mohamed, "Leveraging Client Processing for Location Privacy in Mobile Local Search" (2016). Electronic Theses and Dissertations. 1218.
Recieved from ProQuest
Wisam Mohamed Eltarjaman