PhD, University of Cincinnati, 2013, Engineering and Applied Science: Computer Science and Engineering

Multi-hop network has been envisioned to be a key technology in the next generation wireless networks. It offers the most flexible characteristic of the networks, requires no centralized or very small control, and very little configuration, thereby enjoying the quick deployment property. Yet, it has been named as the "next generation" for so many times that its usefulness is being questioned. The difficulty mainly comes from the inferiority in the performance being poor enough to make such networks overshadowed by other type of networks, unless quick deployment is a requirement, or performance is not a major concern. In this dissertation, we investigate the performance issues in multi-hop wireless networks. Some new architectures of multi-hop wireless networks with network coding are thoroughly explored. We first revisit network coding briefly. Being the latest revolution in the wireless world, network coding has evolved into a more practical shape, and has matured to a level so as to be adopted. A general introduction of network coding (network information theory) is covered for better understanding of how this technology would change the computer networks, and why we pick this topic for our research. The first work this dissertation covers is Murco, an opportunistic framework that brings the benefits of both multi-radio multi-channel technology and network coding to the multi-hop wireless networks. This combination, though it seems natural, faces many challenges. We address them with a loose collaboration between network coding and multi-radio technology coordinated by our framework. Our framework requires few changes or compromises on either side, and the simulation results demonstrate enhanced throughput. Following this work, we get into a more complex problem. Coding-aware routing in multi-hop wireless networks is vital for network coding's possible boom. We address this problem with a heuristic routing metric ETOX and a hybrid routing protocol HyCare. (open full item for complete abstract)

Committee: Dharma Agrawal D.Sc. (Committee Chair); Yizong Cheng Ph.D. (Committee Member); Chia Han Ph.D. (Committee Member); Yiming Hu Ph.D. (Committee Member); John Schlipf Ph.D. (Committee Member) Subjects: Computer Science

PhD, University of Cincinnati, 2008, Engineering : Computer Science and Engineering

Currently, Peer-to-Peer overlays can be classified into two main categories: unstructured and structured ones. Unstructured overlays are simple, robust, and powerful in keyword search. Structured ones can scale to very large systems in terms of node number and geography, and guarantee to locate an object within O(Log N) hops. However, both of them face difficulties in efficiency and security of overlays. For unstructured ones, the efficiency problem presented is poor scalability. For structured ones, it is long routing latency and enormous overhead on handling system churn. Moreover, both of them are vulnerable to malicious attacks. Peer-to-Peer overlays belong to application-level network. To a great extension, overlay network designs ignore physical characteristics. As the result, their structures are far from underlying physical network or the distribution pattern of overlay peers. These inconsistencies induce system common operations costly, such as routing and lookup. On the other hand, most peers are assumed to have uniform resources and similar behaviors. Thus, Peer-to-Peer protocols were designed to be symmetric. However, in the realistic environment, peers' resources and behaviors are highly skewed. Symmetric protocols actually compromise system performance. Frequently joining and leaving of peers generates enormous traffic. The significant fraction of peers with high latency/low bandwidth links increase lookup latency. Moreover, under the environment without mutual trust, Peer-to-Peer systems are very vulnerable for varied attacks because they lack a practical authentication mechanism. From a different perspective, this dissertation proposes to construct a highly efficient and secure Peer-to-Peer overlay based on the physical network structure of the Internet and network locality of overlay peers. By naturally integrating different network-aware techniques into the Peer-to-Peer overlay, a novel SNSA (Scalable Network Structure Aware) technique has been dev (open full item for complete abstract)

Committee: Dr. Yiming Hu (Advisor) Subjects: Computer Science

Doctor of Philosophy, The Ohio State University, 2016, Computer Science and Engineering

The fastest supercomputer in the world as of July 2016 is the Sunway TaihuLight. It can achieve a staggering performance of 93 PetaFlops. This incredible performance is achieved via massive parallelism. Today's supercomputers and compute clusters have tens of thousands of distributed memory nodes with each node comprised of several shared memory multi/many core processors. Scaling on these massively parallel systems is not an easy task. A major performance and scalability bottleneck is the limited data movement bandwidth, which can be orders of magnitude smaller than the computation bandwidth. Developing applications to scale on these massively parallel systems requires minimizing data movement volume at different levels of memory hierarchy using locality optimization techniques. Locality optimization aims to reduce the data movement between slow and fast memory by rescheduling/remapping the original computation to reuse the data once it is in fast memory, thereby avoiding subsequent movement of the same data from slow memory. This dissertation explores multiple aspects of locality optimizations for enhancing scalability and performance of various regular and irregular applications on massively parallel computing environment. It develops distributed algorithms, lower bound techniques, and compiler and runtime frameworks for optimizing Tensor Contractions, Four-Index Transform, Convolutional Neural Networks (CNNs), and Recursive Tree Traversal on k-d trees. Each of these application domains is limited in performance and scalability primarily by data movement costs at a particular level of memory hierarchy. To be specific, on a massively parallel system, distributed Tensor Contractions can have limited scalability due to the cost of communication between distributed memory nodes. The Four-Index Transform, on the other hand, can be limited in the size of the largest problem that can be completed in a reasonable amount of time due to data transfer cost from disk to (open full item for complete abstract)

Committee: P. Sadayappan (Advisor) Subjects: Computer Engineering; Computer Science

MS, University of Cincinnati, 2010, Engineering and Applied Science: Computer Engineering

Future Network-on-Chip (NoC) designs no longer map single cores to each cache slice but rather multiple cores in layouts known as hybrid architectures. Additional proposals have suggested creating reconfigurable hybrid architectures where the OS can revise core-to-cache mappings as required. However, these designs will still be measured by their ability to reduce the average L2 cache delay. Denser core placements with varying core mappings require cache policies with intelligent data placement schemes otherwise there will be no gain to overall system performance as a result of the networked architecture. Solutions such as OS-directed page placement can reduce some of this delay by placing pages in caches local to the initial requestor. However, due to the page-level allocation granularity compared to line-level data accesses, this policy can still result in shared data existing in remote locations during highly parallelized applications. The most effective network delay reduction alternative is line-level data migration. Data migration policies are designed to take advantage of data temporal locality by assuming data recently used by a processor will be used again in the future. Several variations of migration policies have been proposed to address this demand. However, the physical costs, high computation demands and poor scalability of these methods have reduced their effectiveness in future layouts with hundreds of cores. Additionally, many proposals fail to consider migrating data to a centralized location with even latencies for multiple active cores instead they reduce latency for a single core at the expense of all others. This best average placement is also known as the nearest-neighbor search or the “Two-Dimensional Post Office Problem”. The proposed Directional Migration solution attempts to solve these problems by providing an autonomous, line-level migration that is responsive to multiple cores with varying access patterns. This design maintains two (open full item for complete abstract)

Committee: Ranganadha Vemuri PhD (Committee Chair); Carla Purdy, C PhD (Committee Member); Wen Ben Jone PhD (Committee Member) Subjects: Electrical Engineering