Our physical environment is a natural storage where we can store and share information. For instance, a stop sign is a piece of information implanted in a particular location. In this paper, we propose a storage underlay platform called “ELIAS” (Every Location Is A Storage) for a mobile peer-to-peer system. ELIAS enables user applications to transparently save/retrieve data items to/from a location without concerning with low-level detailed implementation. The underlying platform chooses appropriate mobile nodes to store/retrieve data items. We discuss an efficient implementation of ELIAS, and present a detailed analytical model that is validated by simulation. Finally, we show results of sensitivity analysis on a variety of factors that impact the performance of ELIAS.

For safety and privacy protection, nodes participating in ad hoc networks must keep their precise location in secret. When having to report its location (for routing purposes, etc.), a node can report only a cloaking region, a spatial region that contains its current position. Reducing location resolution has a significant impact on the applications that rely on nodes' location information. These applications can become less efficient. Some may not even function. Moreover, operations such as packet forwarding may allow an adversary to refine a node's location resolution, thus reversing the effect of location cloaking in safety and privacy protection. In this paper, we investigate these problems in the context of geographic ad hoc routing and propose a novel concept called safe link. A network link is considered safe if sending and receiving of a data packet do not allow an adversary to refine the sender and receiver's location. With this concept in place, we develop a new secure routing protocol that can work with inaccurate location and construct a routing path using only safe links. We evaluate the performance of the proposed technique under various network conditions using simulation. Detailed results are presented to provide the insight of the impact of safety and privacy protection on packet delivery.

Anonymous location information may be correlated with restricted spaces such as home and office for subject re-identification. This makes it a great challenge to provide location privacy protection for users of location-based services. Existing work adopts traditional K-anonymity model and ensures that each location disclosed in service requests is a spatial region that has been visited by at least K users. This strategy requires a user to specify an appropriate value of K in order to achieve a desired level of privacy protection. This is problematic because privacy is about feeling, and it is awkward for one to scale her feeling using a number. In this paper, we propose a feeling-based privacy model. The model allows a user to express her privacy requirement by specifying a public region, which the user would feel comfortable if the region is reported as her location. The popularity of the public region, measured using entropy based on its visitors' footprints inside it, is then used as the user's desired level of privacy protection. With this model in place, we present a novel technique that allows a user's location information to be reported as accurate as possible while providing her sufficient location privacy protection. The new technique supports trajectory cloaking and can be used in application scenarios where a user needs to make frequent location updates along a trajectory that cannot be predicted. In addition to evaluating the effectiveness of the proposed technique under various conditions through simulation, we have also implemented an experimental system for location privacy-aware uses of location-based services.

Location information is crucial to design efficient and scalable ad hoc networks, yet the exposure of such information presents them significant safety threats. This paper investigates the problem of preventing an adversary from locating (and thus destroying) nodes based on their location information revealed explicitly in communications. Our idea is to reduce location resolution to achieve a desired level of safety protection. We define the safety level of a geographic region to be the ratio of its area and the number of nodes inside it. The higher safety level a region has, the less attractive for an adversary to search over it for the nodes. Thus, when a node has to disclose its location, it can compute a cloaking box that meets a desired level of safety requirement and report that as its current location information. Although the basic idea is simple, there are several challenges to implement it. First, each cloaking box must be as small as possible in order to minimize the impact of reduced location resolution on the efficiency of network operating and applications. Second, nodes must be able to compute their cloaking boxes without having to reveal their accurate position. Finally, given a sequence of cloaking boxes, they must not be correlated to refine an area whose safety level is less than the required. This paper addresses these challenges with cost-effective solutions in the context of both stationary and mobile ad hoc networks. Our extensive performance evaluation indicates that our proposed technique is efficient in node safety protection, and does not have a significant impact on the performance of networks and applications.

We present a new approach for if-anonymity protection in Location-Based Services (LBSs). Specifically, we depersonalize location information by ensuring that each location reported for LBSs is a cloaking area that contains K different footprints-historical locations of different mobile nodes. Therefore, the exact identity and location of the service requestor remain anonymous from LBS service providers. Existing techniques, on the other hand, compute the cloaking area using current locations of K neighboring hosts of the service requestor. Because of this difference, our approach significantly reduces the cloaking area, which in turn decreases query processing and communication overhead for returning query results to the requesting host. In addition, existing techniques also require frequent location updates from all nodes, regardless of whether or not these nodes are requesting LBSs. Most importantly, our approach is the first practical solution that provides K-anonymity trajectory protection needed to ensure anonymity when a mobile host requests LBSs continuously as it moves. Our solution depersonalizes a user's trajectory (a time-series of the user's locations) based on the historical trajectories of other users. We evaluate our techniques under various conditions using location data synthetically generated based on real road maps. The results show that our techniques can provide K-anonymity trajectory protection using a minimized cloaking area.

Today's Internet is incapable of handling information like a stop sign. Such information stays in a geographic region, consistently beams to the people who move into from certain direction. This paper presents a system aimed at bridging the Internet with the physical world. The hardware platform of the system consists of a central server and a set of position-aware mobile clients that communicate with the server through means of wireless networks. Each client has a capable zone, a circular region centered on its current position with a user-defined radius. Each piece of information managed by the server is a geopage, an HTML page associated with a geographic region. As clients move, they load with the geopages implanted in their surroundings. This paper identifies the technical challenges of managing a large number of geopages that spans over a large terrain with a large number of mobile clients, and proposes a cost-effective solution that minimizes mobile communication and server processing costs. We evaluate the scalability of the proposed technique with a detailed mathematical model. The accuracy of the model is verified using simulation. We have also implemented a prototype and field tested the system in several cities.

A major concern for large-scale deployment of location-based services (LBSs) is the potential abuse of their client location data, which may imply sensitive personal information. Location privacy protection is challenging because a location itself may reveal a subject's identity. To support location anonymity, existing research reduces location resolution by ensuring each location reported to a service provider is a cloaking area that contains at least K mobile nodes. This strategy is effective when each location update can be considered as an independent event. In this paper, we investigate location anonymity in the context of continuous LBSs, which require frequent location updates from service users. Knowing that a user is inside a cloaking area constrains its position in the next cloaking area. Thus, simply ensuring each cloaking area contains at least K users does not give a user K-anonymity protection. We propose to measure the anonymity degree of a cloaking area using entropy, which takes into account not only the number of the entities inside, but also their anonymity probability distribution. To find a cloaking area that can provide a given level of anonymity protection and is also as small as possible, we present a novel technique with a polynomial time complexity. The effectiveness of our techniques is studied under various conditions using location data synthetically generated using real road maps and traffic volume data. The results show that our techniques can indeed protect user anonymity at a desired level, and at the same time, minimize the size of each cloaking area, allowing users to receive high quality services.

Our physical environment is a natural storage where we are used to store and share information. A stop sign, for instance, is such a piece of information that is stored on a particular location. In this paper, we assume a fully distributed mobile networking system and investigate the challenges of enabling the system to treat its underlying geographic environment as a virtual storage disk. Specifically, our research aims at developing a platform that allows one to conceptually save/retrieve a data item to/from a location. Since this capability can be used in a broad range of applications, the proposed platform can serve as a common underlay for developing and running these applications. We propose a cost-effective implementation of this platform. For the purpose of performance evaluation, we develop a detailed analytical model and verify its accuracy using simulation.

Multimedia streaming applications can significantly boost the value of mobile ad hoc networks (MANET). Live streaming, however, means continuous data delivery, which is a major challenge in MANET. Because of host mobility, a streaming path may be broken easily, causing streaming interruption. In this paper, we address this problem with a light-weighted yet robust streaming protocol. Our technique distinguishes itself from existing ones with two innovative features: proactive link protection (PLP) and receiver-oriented adaptation (ROA). PLP allows a mobile host in a streaming path to find an alternative link before its current one becomes broken. This feature minimizes the chance of having to discover a new path in urgent to replace a broken link. While PLP protects the streaming continuity, ROA ensures the streaming efficiency by minimizing the hop number of a streaming path. Specifically, ROA dynamically adjusts the path of a live stream to make it as straight as possible along the direction from the source to the receiver. We evaluate the proposed technique through simulation, and our extensive study indicates that the new technique can support robust streaming applications with a minimal control overhead.

Unlike conventional range queries, a range-monitoring query is a continuous query. It requires retrieving mobile objects inside a user-defined region and providing continuous updates as the objects move into and out of the region. In this paper, we present an efficient technique for real-time processing of such queries. In our approach, each mobile object is associated with a resident domain, and when an object moves, it monitors its spatial relationship with its resident domain and the monitoring areas inside it. An object reports its location to the server when it crosses over some query boundary or moves out of its resident domain. In the first case, the server updates the affected query results accordingly, while in the second case, the server determines a new resident domain for the object. This distributive approach achieves an accurate and real-time monitoring effect with minimal mobile communication and server processing costs. Our approach also allows a mobile object to negotiate a resident domain based on its computing capability. By having a larger resident domain, a more capable object has less of a chance of moving out of it and having to request a new one. As a result, both communication and server processing costs are reduced. Our comprehensive performance study shows that the proposed technique can be highly scalable in supporting location-based services in a wireless environment that consists of a large number of mobile devices.