Kyriakos Mouratidis

Preventing Location-Based Identity Inference in Anonymous Spatial Queries

The increasing trend of embedding positioning capabilities (for example, GPS) in mobile devices facilitates the widespread use of location-based services. For such applications to succeed, privacy and confidentiality are essential. Existing privacy-enhancing techniques rely on encryption to safeguard communication channels, and on pseudonyms to protect user identities. Nevertheless, the query contents may disclose the physical location of the user. In this paper, we present a framework for preventing location-based identity inference of users who issue spatial queries to location-based services. We propose transformations based on the well-established K-anonymity concept to compute exact answers for range and nearest neighbor search, without revealing the query source. Our methods optimize the entire process of anonymizing the requests and processing the transformed spatial queries. Extensive experimental studies suggest that the proposed techniques are applicable to real-life scenarios with numerous mobile users.

Conceptual partitioning: an efficient method for continuous nearest neighbor monitoring

Given a set of objects P and a query point q, a k nearest neighbor (k-NN) query retrieves the k objects in P that lie closest to q. Even though the problem is well-studied for static datasets, the traditional methods do not extend to highly dynamic environments where multiple continuous queries require real-time results, and both objects and queries receive frequent location updates. In this paper we propose conceptual partitioning (CPM), a comprehensive technique for the efficient monitoring of continuous NN queries. CPM achieves low running time by handling location updates only from objects that fall in the vicinity of some query (and ignoring the rest). It can be used with multiple, static or moving queries, and it does not make any assumptions about the object moving patterns. We analyze the performance of CPM and show that it outperforms the current state-of-the-art algorithms for all problem settings. Finally, we extend our framework to aggregate NN (ANN) queries, which monitor the data objects that minimize the aggregate distance with respect to a set of query points (e.g., the objects with the minimum sum of distances to all query points).

Aggregate nearest neighbor queries in spatial databases

Given two spatial datasets <i>P</i> (e.g., facilities) and <i>Q</i> (queries), an <i>aggregate nearest neighbor</i> (ANN) query retrieves the point(s) of <i>P</i> with the smallest aggregate distance(s) to points in <i>Q</i>. Assuming, for example, <i>n</i> users at locations <i>q</i><inf>1</inf>,…<i>q</i><inf><i>n</i></inf>, an ANN query outputs the facility <i>p</i> ∈ <i>P</i> that minimizes the <i>sum</i> of distances |<i>pq</i><inf>i</inf>| for 1 ≤ <i>i</i> ≤ <i>n</i> that the users have to travel in order to meet there. Similarly, another ANN query may report the point <i>p</i> ∈ <i>P</i> that minimizes the <i>maximum</i> distance that any user has to travel, or the <i>minimum</i> distance from some user to his/her closest facility. If <i>Q</i> fits in memory and <i>P</i> is indexed by an R-tree, we develop algorithms for aggregate nearest neighbors that capture several versions of the problem, including weighted queries and incremental reporting of results. Then, we analyze their performance and propose cost models for query optimization. Finally, we extend our techniques for disk-resident queries and approximate ANN retrieval. The efficiency of the algorithms and the accuracy of the cost models are evaluated through extensive experiments with real and synthetic datasets.

Continuous monitoring of top-k queries over sliding windows

Given a dataset P and a preference function f, a top-k query retrieves the k tuples in P with the highest scores according to f. Even though the problem is well-studied in conventional databases, the existing methods are inapplicable to highly dynamic environments involving numerous long-running queries. This paper studies continuous monitoring of top-k queries over a fixed-size window W of the most recent data. The window size can be expressed either in terms of the number of active tuples or time units. We propose a general methodology for top-k monitoring that restricts processing to the sub-domains of the workspace that influence the result of some query. To cope with high stream rates and provide fast answers in an on-line fashion, the data in W reside in main memory. The valid records are indexed by a grid structure, which also maintains book-keeping information. We present two processing techniques: the first one computes the new answer of a query whenever some of the current top-k points expire; the second one partially pre-computes the future changes in the result, achieving better running time at the expense of slightly higher space requirements. We analyze the performance of both algorithms and evaluate their efficiency through extensive experiments. Finally, we extend the proposed framework to other query types and a different data stream model.

Group nearest neighbor queries

Given two sets of points P and Q, a group nearest neighbor (GNN) query retrieves the point(s) of P with the smallest sum of distances to all points in Q. Consider, for instance, three users at locations q/sub 1/ q/sub 2/ and q/sub 3/ that want to find a meeting point (e.g., a restaurant); the corresponding query returns the data point p that minimizes the sum of Euclidean distances |pq/sub i/| for 1/spl les/i/spl les/3. Assuming that Q fits in memory and P is indexed by an R-tree, we propose several algorithms for finding the group nearest neighbors efficiently. As a second step, we extend our techniques for situations where Q cannot fit in memory, covering both indexed and nonindexed query points. An experimental evaluation identifies the best alternative based on the data and query properties.

A threshold-based algorithm for continuous monitoring of k nearest neighbors

Assume a set of moving objects and a central server that monitors their positions over time, while processing continuous nearest neighbor queries from geographically distributed clients. In order to always report up-to-date results, the server could constantly obtain the most recent position of all objects. However, this naive solution requires the transmission of a large number of rapid data streams corresponding to location updates. Intuitively, current information is necessary only for objects that may influence some query result (i.e., they may be included in the nearest neighbor set of some client). Motivated by this observation, we present a threshold-based algorithm for the continuous monitoring of nearest neighbors that minimizes the communication overhead between the server and the data objects. The proposed method can be used with multiple, static, or moving queries, for any distance definition, and does not require additional knowledge (e.g., velocity vectors) besides object locations.

Scalable verification for outsourced dynamic databases

Query answers from servers operated by third parties need to be verified, as the third parties may not be trusted or their servers may be compromised. Most of the existing authentication methods construct validity proofs based on the Merkle hash tree (MHT). The MHT, however, imposes severe concurrency constraints that slow down data updates. We introduce a protocol, built upon signature aggregation, for checking the authenticity, completeness and freshness of query answers. The protocol offers the important property of allowing new data to be disseminated immediately, while ensuring that outdated values beyond a pre-set age can be detected. We also propose an efficient verification technique for ad-hoc equijoins, for which no practical solution existed. In addition, for servers that need to process heavy query workloads, we introduce a mechanism that significantly reduces the proof construction time by caching just a small number of strategically chosen aggregate signatures. The efficiency and efficacy of our proposed mechanisms are confirmed through extensive experiments.

Partially materialized digest scheme: an efficient verification method for outsourced databases

In the outsourced database model, a data owner publishes her database through a third-party server; i.e., the server hosts the data and answers user queries on behalf of the owner. Since the server may not be trusted, or may be compromised, users need a means to verify that answers received are both authentic and complete, i.e., that the returned data have not been tampered with, and that no qualifying results have been omitted. We propose a result verification approach for one-dimensional queries, called Partially Materialized Digest scheme (PMD), that applies to both static and dynamic databases. PMD uses separate indexes for the data and for their associated verification information, and only partially materializes the latter. In contrast with previous work, PMD avoids unnecessary costs when processing queries that do not request verification, achieving the performance of an ordinary index (e.g., a B+-tree). On the other hand, when an authenticity and completeness proof is required, PMD outperforms the existing state-of-the-art technique by a wide margin, as we demonstrate analytically and experimentally. Furthermore, we design two verification methods for spatial queries. The first, termed Merkle R-tree (MR-tree), extends the conventional approach of embedding authentication information into the data index (i.e., an R-tree). The second, called Partially Materialized KD-tree (PMKD), follows the PMD paradigm using separate data and verification indexes. An empirical evaluation with real data shows that the PMD methodology is superior to the traditional approach for spatial queries too.

Anonymous Query Processing in Road Networks

The increasing availability of location-aware mobile devices has given rise to a flurry of location-based services (LBSs). Due to the nature of spatial queries, an LBS needs the user position in order to process her requests. On the other hand, revealing exact user locations to a (potentially untrusted) LBS may pinpoint their identities and breach their privacy. To address this issue, spatial anonymity techniques obfuscate user locations, forwarding to the LBS a sufficiently large region instead. Existing methods explicitly target processing in the euclidean space and do not apply when proximity to the users is defined according to network distance (e.g., driving time through the roads of a city). In this paper, we propose a framework for anonymous query processing in road networks. We design location obfuscation techniques that: (1) provide anonymous LBS access to the users and (2) allow efficient query processing at the LBS side. Our techniques exploit existing network database infrastructure, requiring no specialized storage schemes or functionalities. We experimentally compare alternative designs in real road networks and demonstrate the effectiveness of our techniques.

Authenticating the query results of text search engines

The number of successful attacks on the Internet shows that it is very difficult to guarantee the security of online search engines. A breached server that is not detected in time may return incorrect results to the users. To prevent that, we introduce a methodology for generating an integrity proof for each search result. Our solution is targeted at search engines that perform similarity-based document retrieval, and utilize an inverted list implementation (as most search engines do). We formulate the properties that define a correct result, map the task of processing a text search query to adaptations of existing threshold-based algorithms, and devise an authentication scheme for checking the validity of a result. Finally, we confirm the efficiency and practicality of our solution through an empirical evaluation with real documents and benchmark queries.