Bu Gra Gedik’s research while affiliated with Georgia Institute of Technology and other places

The peer to peer (P2P) file sharing systems such as Gnutella have been widely acknowledged as the fastest growing Internet applications ever. The P2P model has many potential advantages due to the design flexibility of overlay networks and the server-less management of cooperative sharing of information and resources. However, these systems suffer from the well-known performance mismatch between the randomly constructed overlay network topology and the underlying IP layer topology for packet routing. This paper proposes to structure the P2P overlay topology using a capacity-aware multi-tier topology to better balance load at peers with heterogeneous capacities and to prevent low capacity nodes from downgrading the performance of the system. To study the benefits and cost of the multi-tier capacity aware topology with respect to basic and advanced routing protocols, we also develop a probabilistic broadening scheme for efficient routing, which further utilizes capacity-awareness to enhance the P2P routing performance of the system. We evaluate our design through simulations. The results show that our multi-tier topologies alone can provide eight to ten times improvements in the messaging cost, two to three orders of magnitude improvement in terms of load balancing characteristics, and seven to eight times lower topology construction and maintenance costs when compared to Gnutellas random power-law topology.

The advances in wireless communication and decreasing costs of mobile devices have enabled users to access desired information at any time. Coupled with positioning technologies like GPS, this opens up an exciting domain of location based services, allowing a mobile user to query for objects based on its current position. Main bottlenecks in such infrastructures are the draining of power of the mobile devices and the limited network bandwidth available. To alleviate these problems, broadcasting spatial information about relevant objects has been widely accepted as an efficient mechanism. An important class of queries for such an infrastructure is the k-nearest neighbor (kNN) queries, in which users are interested in k closest objects to their position. Most of the research in kNN queries, use unconventional broadcast indexes and provide only approximate kNN search [30, 15]. In this paper, we describe mechanisms to perform exact kNN search on conventional sequential-access R-trees, and optimize established kNN search algorithms. We also propose a novel use of histograms for guiding the search and derive analytical results on maximum queue size and node access count. In addition, we discuss the effects of different broadcast organizations on search performance and challenge the traditional use of Depth-First (df s ) organization. We also extend our mechanisms to support kNN search with non-spatial constraints. While we demonstrate our ideas using a broadcast index, they are equally applicable to any kind of sequential access medium like tertiary tape storage. We validate our mechanims through an extensive experimental analysis and present our findings.

Location monitoring is an important issue for real time management of mobile object positions. Significant research efforts have been dedicated to techniques for efficient processing of spatial continuous queries on moving objects in a centralized location monitoring system. Surprisingly, very few have promoted a distributed approach to real-time location monitoring. In this paper we present a distributed and scalable solution to processing continuously moving queries on moving objects and describe the design of MobiEyes, a distributed real-time location monitoring system in a mobile environment. Mobieyes utilizes the computational power at mobile objects, leading to significant savings in terms of server load and messaging cost when compared to solutions relying on central processing of location information at the server. We introduce a set of optimization techniques, such as Lazy Query Propagation, Query Grouping, and Safe Periods, to constrict the amount of computations handled by the moving objects and to enhance the performance and system utilization of Mobieyes. We also provide a simulation model in a mobile setup to study the scalability of the MobiEyes distributed location monitoring approach with regard to server load, messaging cost, and amount of computation required on the mobile objects.

A key challenge in peer-to-peer computing systems is to provide a decentralized and yet reliable service on top of a network of loosely coupled, weakly connected and possibly unreliable peers. This paper presents an effective dynamic passive replication scheme designed to provide reliable service in PeerCQ, a decentralized and self-configurable peer-to-peer Internet information monitoring system. We first describe the design of a distributed replication scheme, which enables reliable processing of long-running information monitoring requests in an environment of inherently unreliable peers. Then we present an analytical model to discuss its fault tolerance properties. A set of initial experiments is reported, showing the feasibility and the effectiveness of the proposed approach.

Connectivity based node clustering has wide ranging applications in decentralized Peer-to-Peer (P2P) networks such as P2P file sharing systems, mobile ad-hoc networks, P2P sensor networks and so forth. This paper describes a Connectivity-based Distributed Node Clustering scheme (CDC). This scheme presents a scalable and an efficient solution for discovering connectivity based clusters in peer networks. In contrast to centralized graph clustering algorithms, the CDC scheme is completely decentralized and it only assumes the knowledge of neighbor nodes, instead of requiring a global knowledge of the network (graph) to be available. An important feature of the CDC scheme is its ability to cluster the entire network automatically or to discover clusters around a given set of nodes. We provide detailed experimental evaluations of the CDC scheme, addressing its effectiveness in discovering good quality clusters. Our experiments show that utilizing message-based connectivity structure can considerably reduce the messaging cost, and provide better utilization of resources, which in turn improves the quality of service of the applications executing over decentralized peer-to-peer networks.

Location monitoring is an important issue for real time management of mobile object positions.

Continued advances in mobile networks and positioning technologies have created a strong market push for location-based services (LBSs). Examples include location-aware emergency services, location based service advertisement, and location sensitive billing. One of the big challenges in wide deployment of LBS systems is the privacy-preserving management of location-based data. Without safeguards, extensive deployment of location based services endangers location privacy of mobile users and exhibits significant vulnerabilities for abuse.

PeerCQ is a totally decentralized system that performs information monitoring tasks over a network of peers with heterogeneous capabilities. It uses Continual Queries (CQs) as its primitives to express information-monitoring requests. A primary objective of the PeerCQ system is to build a decentralized Internet scale distributed information-monitoring system, which is highly scalable, self-configurable and supports efficient and robust way of processing CQs. This paper describes the basic architecture of the PeerCQ system and focuses on the mechanisms used for service partitioning at the P2P protocol layer. A set of initial experiments is reported, demonstrating the sensitiveness of the PeerCQ approach to large scale P2P information monitoring and the effectiveness of the PeerCQ service-partitioning algorithms with respect to load balancing and system utilization.

UsulDSM is a page-based recoverable software distributed shared memory system designed for network of computers that don't have access to a physically shared memory. In this report we describe architecture of the UsulDSM and discuss its design and implementation. We also evaluate its performance through a simple parallel application that uses UsulDSM. We also analyze UsulDSM's scalability and the overhead in-duced by its recoverability property.