Mermera: Non-Coherent Distributed Shared Memory for Parallel Computing

The proliferation of inexpensive workstations and networks has prompted several researchers to use such distributed systems for parallel computing. Attempts have been made to offer a shared-memory programming model on such distributed memory computers. Most systems provide a shared-memory that is coherent in that all processes that use it agree on the order of all memory events. This dissertation explores the possibility of a significant improvement in the performance of some applications when they use non-coherent memory. First, a new formal model to describe existing non-coherent memories is developed. I use this model to prove that certain problems can be solved using asynchronous iterative algorithms on shared-memory in which the coherence constraints are substantially relaxed. In the course of the development of the model I discovered a new type of non-coherent behavior called Local Consistency. Second, a programming model, Mermera, is proposed. It provides programmers with a choice of hierarchically related non-coherent behaviors along with one coherent behavior. Thus, one can trade-off the ease of programming with coherent memory for improved performance with non-coherent memory. As an example, I present a program to solve a linear system of equations using an asynchronous iterative algorithm. This program uses all the behaviors offered by Mermera. Third, I describe the implementation of Mermera on a BBN Butterfly TC2000 and on a network of workstations. The performance of a version of the equation solving program that uses all the behaviors of Mermera is compared with that of a version that uses coherent behavior only. For a system of 1000 equations the former exhibits at least a 5-fold improvement in convergence time over the latter. The version using coherent behavior only does not benefit from employing more than one workstation to solve the problem while the program using non-coherent behavior continues to achieve improved performance as the number of workstations is increased from 1 to 6. This measurement corroborates our belief that non-coherent shared memory can be a performance boon for some applications.

AIDA-Based Real-Time Fault-Tolerant Broadcast Disks

The proliferation of mobile computers and wireless networks requires the design of future distributed real-time applications to recognize and deal with the significant asymmetry between downstream and upstream communication capacities, and the significant disparity between server and client storage capacities. Recent research work proposed the use of Broadcast Disks as a scalable mechanism to deal with this problem. In this paper, we propose a new broadcast disks protocol, based on our Adaptive Information Dispersal Algorithm (AIDA). Our protocol is different from previous broadcast disks protocols in that it improves communication timeliness, fault-tolerance, and security, while allowing for a finer control of multiplexing of prioritized data (broadcast frequencies). We start with a general introduction of broadcast disks. Next, we propose broadcast disk organizations that are suitable for real-time applications. Next, we present AIDA and show its fault-tolerance and security properties. We conclude the paper with the description and analysis of AIDA-based broadcast disks organizations that achieve both timeliness and fault-tolerance, while preserving downstream communication capacity.

ImageRover: A Content-Based Image Browser for the World Wide Web

ImageRover is a search by image content navigation tool for the world wide web. To gather images expediently, the image collection subsystem utilizes a distributed fleet of WWW robots running on different computers. The image robots gather information about the images they find, computing the appropriate image decompositions and indices, and store this extracted information in vector form for searches based on image content. At search time, users can iteratively guide the search through the selection of relevant examples. Search performance is made efficient through the use of an approximate, optimized k-d tree algorithm. The system employs a novel relevance feedback algorithm that selects the distance metrics appropriate for a particular query.

Forewarding in Mobile Opportunistic Networks

Recent advances in processor speeds, mobile communications and battery life have enabled computers to evolve from completely wired to completely mobile. In the most extreme case, all nodes are mobile and communication takes place at available opportunities – using both traditional communication infrastructure as well as the mobility of intermediate nodes. These are mobile opportunistic networks. Data communication in such networks is a difficult problem, because of the dynamic underlying topology, the scarcity of network resources and the lack of global information. Establishing end-to-end routes in such networks is usually not feasible. Instead a store-and-carry forwarding paradigm is better suited for such networks. This dissertation describes and analyzes algorithms for forwarding of messages in such networks. In order to design effective forwarding algorithms for mobile opportunistic networks, we start by first building an understanding of the set of all paths between nodes, which represent the available opportunities for any forwarding algorithm. Relying on real measurements, we enumerate paths between nodes and uncover what we refer to as the path explosion effect. The term path explosion refers to the fact that the number of paths between a randomly selected pair of nodes increases exponentially with time. We draw from the theory of epidemics to model and explain the path explosion effect. This is the first contribution of the thesis, and is a key observation that underlies subsequent results. Our second contribution is the study of forwarding algorithms. For this, we rely on trace driven simulations of different algorithms that span a range of design dimensions. We compare the performance (success rate and average delay) of these algorithms. We make the surprising observation that most algorithms we consider have roughly similar performance. We explain this result in light of the path explosion phenomenon. While the performance of most algorithms we studied was roughly the same, these algorithms differed in terms of cost. This prompted us to focus on designing algorithms with the explicit intent of reducing costs. For this, we cast the problem of forwarding as an optimal stopping problem. Our third main contribution is the design of strategies based on optimal stopping principles which we refer to as Delegation schemes. Our analysis shows that using a delegation scheme reduces cost over naive forwarding by a factor of O(√N), where N is the number of nodes in the network. We further validate this result on real traces, where the cost reduction observed is even greater. Our results so far include a key assumption, which is unbounded buffers on nodes. Next, we relax this assumption, so that the problem shifts to one of prioritization of messages for transmission and dropping. Our fourth contribution is the study of message prioritization schemes, combined with forwarding. Our main result is that one achieves higher performance by assigning higher priorities to young messages in the network. We again interpret this result in light of the path explosion effect.