Matching Items (29)
Filtering by
- All Subjects: Computer Science
Description
This dissertation studies routing in small-world networks such as grids plus long-range edges and real networks. Kleinberg showed that geography-based greedy routing in a grid-based network takes an expected number of steps polylogarithmic in the network size, thus justifying empirical efficiency observed beginning with Milgram. A counterpart for the grid-based model is provided; it creates all edges deterministically and shows an asymptotically matching upper bound on the route length. The main goal is to improve greedy routing through a decentralized machine learning process. Two considered methods are based on weighted majority and an algorithm of de Farias and Megiddo, both learning from feedback using ensembles of experts. Tests are run on both artificial and real networks, with decentralized spectral graph embedding supplying geometric information for real networks where it is not intrinsically available. An important measure analyzed in this work is overpayment, the difference between the cost of the method and that of the shortest path. Adaptive routing overtakes greedy after about a hundred or fewer searches per node, consistently across different network sizes and types. Learning stabilizes, typically at overpayment of a third to a half of that by greedy. The problem is made more difficult by eliminating the knowledge of neighbors' locations or by introducing uncooperative nodes. Even under these conditions, the learned routes are usually better than the greedy routes. The second part of the dissertation is related to the community structure of unannotated networks. A modularity-based algorithm of Newman is extended to work with overlapping communities (including considerably overlapping communities), where each node locally makes decisions to which potential communities it belongs. To measure quality of a cover of overlapping communities, a notion of a node contribution to modularity is introduced, and subsequently the notion of modularity is extended from partitions to covers. The final part considers a problem of network anonymization, mostly by the means of edge deletion. The point of interest is utility preservation. It is shown that a concentration on the preservation of routing abilities might damage the preservation of community structure, and vice versa.
ContributorsBakun, Oleg (Author) / Konjevod, Goran (Thesis advisor) / Richa, Andrea (Thesis advisor) / Syrotiuk, Violet R. (Committee member) / Czygrinow, Andrzej (Committee member) / Arizona State University (Publisher)
Created2011
Description
Contention based IEEE 802.11MAC uses the binary exponential backoff algorithm (BEB) for the contention resolution. The protocol suffers poor performance in the heavily loaded networks and MANETs, high collision rate and packet drops, probabilistic delay guarantees, and unfairness. Many backoff strategies were proposed to improve the performance of IEEE 802.11 but all ignore the network topology and demand. Persistence is defined as the fraction of time a node is allowed to transmit, when this allowance should take into account topology and load, it is topology and load aware persistence (TLA). We develop a relation between contention window size and the TLA-persistence. We implement a new backoff strategy where the TLA-persistence is defined as the lexicographic max-min channel allocation. We use a centralized algorithm to calculate each node's TLApersistence and then convert it into a contention window size. The new backoff strategy is evaluated in simulation, comparing with that of the IEEE 802.11 using BEB. In most of the static scenarios like exposed terminal, flow in the middle, star topology, and heavy loaded multi-hop networks and in MANETs, through the simulation study, we show that the new backoff strategy achieves higher overall average throughput as compared to that of the IEEE 802.11 using BEB.
ContributorsBhyravajosyula, Sai Vishnu Kiran (Author) / Syrotiuk, Violet R. (Thesis advisor) / Sen, Arunabha (Committee member) / Richa, Andrea (Committee member) / Arizona State University (Publisher)
Created2013
Description
Communication networks, both wired and wireless, are expected to have a certain level of fault-tolerance capability.These networks are also expected to ensure a graceful degradation in performance when some of the network components fail. Traditional studies on fault tolerance in communication networks, for the most part, make no assumptions regarding the location of node/link faults, i.e., the faulty nodes and links may be close to each other or far from each other. However, in many real life scenarios, there exists a strong spatial correlation among the faulty nodes and links. Such failures are often encountered in disaster situations, e.g., natural calamities or enemy attacks. In presence of such region-based faults, many of traditional network analysis and fault-tolerant metrics, that are valid under non-spatially correlated faults, are no longer applicable. To this effect, the main thrust of this research is design and analysis of robust networks in presence of such region-based faults. One important finding of this research is that if some prior knowledge is available on the maximum size of the region that might be affected due to a region-based fault, this piece of knowledge can be effectively utilized for resource efficient design of networks. It has been shown in this dissertation that in some scenarios, effective utilization of this knowledge may result in substantial saving is transmission power in wireless networks. In this dissertation, the impact of region-based faults on the connectivity of wireless networks has been studied and a new metric, region-based connectivity, is proposed to measure the fault-tolerance capability of a network. In addition, novel metrics, such as the region-based component decomposition number(RBCDN) and region-based largest component size(RBLCS) have been proposed to capture the network state, when a region-based fault disconnects the network. Finally, this dissertation presents efficient resource allocation techniques that ensure tolerance against region-based faults, in distributed file storage networks and data center networks.
ContributorsBanerjee, Sujogya (Author) / Sen, Arunabha (Thesis advisor) / Xue, Guoliang (Committee member) / Richa, Andrea (Committee member) / Hurlbert, Glenn (Committee member) / Arizona State University (Publisher)
Created2013
Description
While network problems have been addressed using a central administrative domain with a single objective, the devices in most networks are actually not owned by a single entity but by many individual entities. These entities make their decisions independently and selfishly, and maybe cooperate with a small group of other entities only when this form of coalition yields a better return. The interaction among multiple independent decision-makers necessitates the use of game theory, including economic notions related to markets and incentives. In this dissertation, we are interested in modeling, analyzing, addressing network problems caused by the selfish behavior of network entities. First, we study how the selfish behavior of network entities affects the system performance while users are competing for limited resource. For this resource allocation domain, we aim to study the selfish routing problem in networks with fair queuing on links, the relay assignment problem in cooperative networks, and the channel allocation problem in wireless networks. Another important aspect of this dissertation is the study of designing efficient mechanisms to incentivize network entities to achieve certain system objective. For this incentive mechanism domain, we aim to motivate wireless devices to serve as relays for cooperative communication, and to recruit smartphones for crowdsourcing. In addition, we apply different game theoretic approaches to problems in security and privacy domain. For this domain, we aim to analyze how a user could defend against a smart jammer, who can quickly learn about the user's transmission power. We also design mechanisms to encourage mobile phone users to participate in location privacy protection, in order to achieve k-anonymity.
ContributorsYang, Dejun (Author) / Xue, Guoliang (Thesis advisor) / Richa, Andrea (Committee member) / Sen, Arunabha (Committee member) / Zhang, Junshan (Committee member) / Arizona State University (Publisher)
Created2013
Description
Broadcast Encryption is the task of cryptographically securing communication in a broadcast environment so that only a dynamically specified subset of subscribers, called the privileged subset, may decrypt the communication. In practical applications, it is desirable for a Broadcast Encryption Scheme (BES) to demonstrate resilience against attacks by colluding, unprivileged subscribers. Minimal Perfect Hash Families (PHFs) have been shown to provide a basis for the construction of memory-efficient t-resilient Key Pre-distribution Schemes (KPSs) from multiple instances of 1-resilient KPSs. Using this technique, the task of constructing a large t-resilient BES is reduced to finding a near-minimal PHF of appropriate parameters. While combinatorial and probabilistic constructions exist for minimal PHFs with certain parameters, the complexity of constructing them in general is currently unknown. This thesis introduces a new type of hash family, called a Scattering Hash Family (ScHF), which is designed to allow for the scalable and ingredient-independent design of memory-efficient BESs for large parameters, specifically resilience and total number of subscribers. A general BES construction using ScHFs is shown, which constructs t-resilient KPSs from other KPSs of any resilience ≤w≤t. In addition to demonstrating how ScHFs can be used to produce BESs , this thesis explores several ScHF construction techniques. The initial technique demonstrates a probabilistic, non-constructive proof of existence for ScHFs . This construction is then derandomized into a direct, polynomial time construction of near-minimal ScHFs using the method of conditional expectations. As an alternative approach to direct construction, representing ScHFs as a k-restriction problem allows for the indirect construction of ScHFs via randomized post-optimization. Using the methods defined, ScHFs are constructed and the parameters' effects on solution size are analyzed. For large strengths, constructive techniques lose significant performance, and as such, asymptotic analysis is performed using the non-constructive existential results. This work concludes with an analysis of the benefits and disadvantages of BESs based on the constructed ScHFs. Due to the novel nature of ScHFs, the results of this analysis are used as the foundation for an empirical comparison between ScHF-based and PHF-based BESs . The primary bases of comparison are construction efficiency, key material requirements, and message transmission overhead.
ContributorsO'Brien, Devon James (Author) / Colbourn, Charles J (Thesis advisor) / Bazzi, Rida (Committee member) / Richa, Andrea (Committee member) / Arizona State University (Publisher)
Created2012
Description
The Dual Marching Tetrahedra algorithm is a generalization of the Dual Marching Cubes algorithm, used to build a boundary surface around points which have been assigned a particular scalar density value, such as the data produced by and Magnetic Resonance Imaging or Computed Tomography scanner. This boundary acts as a skin between points which are determined to be "inside" and "outside" of an object. However, the DMT is vague in regards to exactly where each vertex of the boundary should be placed, which will not necessarily produce smooth results. Mesh smoothing algorithms which ignore the DMT data structures may distort the output mesh so that it could incorrectly include or exclude density points. Thus, an algorithm is presented here which is designed to smooth the output mesh, while obeying the underlying data structures of the DMT algorithm.
ContributorsJohnson, Sean (Author) / Farin, Gerald (Thesis advisor) / Richa, Andrea (Committee member) / Nallure Balasubramanian, Vineeth (Committee member) / Arizona State University (Publisher)
Created2011
Description
A primary goal in computer science is to develop autonomous systems. Usually, we provide computers with tasks and rules for completing those tasks, but what if we could extend this type of system to physical technology as well? In the field of programmable matter, researchers are tasked with developing synthetic materials that can change their physical properties \u2014 such as color, density, and even shape \u2014 based on predefined rules or continuous, autonomous collection of input. In this research, we are most interested in particles that can perform computations, bond with other particles, and move. In this paper, we provide a theoretical particle model that can be used to simulate the performance of such physical particle systems, as well as an algorithm to perform expansion, wherein these particles can be used to enclose spaces or even objects.
ContributorsLaff, Miles (Author) / Richa, Andrea (Thesis director) / Bazzi, Rida (Committee member) / Computer Science and Engineering Program (Contributor) / Barrett, The Honors College (Contributor) / School of Mathematical and Statistical Sciences (Contributor)
Created2015-05
Description
Many programmable matter systems have been proposed and realized recently, each often tailored toward a particular task or physical setting. In our work on self-organizing particle systems, we abstract away from specific settings and instead describe programmable matter as a collection of simple computational elements (to be referred to as particles) with limited computational power that each perform fully distributed, local, asynchronous algorithms to solve system-wide problems of movement, configuration, and coordination. In this thesis, we focus on the compression problem, in which the particle system gathers as tightly together as possible, as in a sphere or its equivalent in the presence of some underlying geometry. While there are many ways to formalize what it means for a particle system to be compressed, we address three different notions of compression: (1) local compression, in which each individual particle utilizes local rules to create an overall convex structure containing no holes, (2) hole elimination, in which the particle system seeks to detect and eliminate any holes it contains, and (3) alpha-compression, in which the particle system seeks to shrink its perimeter to be within a constant factor of the minimum possible value. We analyze the behavior of each of these algorithms, examining correctness and convergence where appropriate. In the case of the Markov Chain Algorithm for Compression, we provide improvements to the original bounds for the bias parameter lambda which influences the system to either compress or expand. Lastly, we briefly discuss contributions to the problem of leader election--in which a particle system elects a single leader--since it acts as an important prerequisite for compression algorithms that use a predetermined seed particle.
ContributorsDaymude, Joshua Jungwoo (Author) / Richa, Andrea (Thesis director) / Kierstead, Henry (Committee member) / Computer Science and Engineering Program (Contributor) / School of Mathematical and Statistical Sciences (Contributor) / Barrett, The Honors College (Contributor)
Created2016-05
Description
We consider programmable matter as a collection of simple computational elements (or particles) that self-organize to solve system-wide problems of movement, configuration, and coordination. Here, we focus on the compression problem, in which the particle system gathers as tightly together as possible, as in a sphere or its equivalent in the presence of some underlying geometry. Within this model a configuration of particles can be represented as a unique closed self-avoiding walk on the triangular lattice. In this paper we will examine the bias parameter of a Markov chain based algorithm that solves the compression problem under the geometric amoebot model, for particle systems that begin in a connected configuration with no holes. This bias parameter, $\lambda$, determines the behavior of the algorithm. It has been shown that for $\lambda > 2+\sqrt{2}$, with all but exponentially small probability, the algorithm achieves compression. Additionally the same algorithm can be used for expansion for small values of $\lambda$; in particular, for all $0 < \lambda < \sqrt{\tau}$, where $\lim_{n\to\infty} {(p_n)^{1
}}=\tau$. This research will focus on improving approximations on the lower bound of $\tau$. Toward this end we will examine algorithmic enumeration, and series analysis for self-avoiding polygons.
}}=\tau$. This research will focus on improving approximations on the lower bound of $\tau$. Toward this end we will examine algorithmic enumeration, and series analysis for self-avoiding polygons.
ContributorsLough, Kevin James (Author) / Richa, Andrea (Thesis director) / Fishel, Susanna (Committee member) / School of Mathematical and Statistical Sciences (Contributor, Contributor) / Computer Science and Engineering Program (Contributor) / Barrett, The Honors College (Contributor)
Created2019-05
Description
Implementing a distributed algorithm is more complicated than implementing a non-distributed algorithm. This is because distributed systems involve coordination of different processes each of which has a partial view of the global system state. The only way to share information in a distributed system is by message passing. Task that are straightforward in a non-distributed system, like deciding on the value of a global system state, can be quite complicated to achieve in a distributed system [1]. On top of the difficulties caused by the distributed nature of the computations, distributed systems typically need to be able to operate normally even if some of the nodes in the system are faulty which further adds to the uncertainty that processes have about the global state. Many factors make the implementation of a distributed algorithms difficult. Design patterns [2] are useful in simplifying the development of general algorithms. A design pattern describes a high level solution to a common, abstract problem that many systems may face. Common structural, creational, and behavioral problems are identified and elegantly solved by design patterns. By identifying features that an algorithm uses, and framing each feature as one of the common problems that a specific design pattern solves, designing a robust implementation of an algorithm becomes more manageable. In this way, design patterns can aid the implementation of algorithms. Unfortunately, design patterns are typically not discussed when developing distributed algorithms. Because correctly developing a distributed algorithm is difficult, many papers (eg. [1], [3], [4]) focus on verifying the correctness of the developed algorithm. Papers that are more practical ([5], [6]) establish the correctness of their algorithm and that their algorithm is efficient enough to be practical. However, papers on distributed algorithms usually make little mention of design patterns. The goal of this work was to gain experience implementing distributed systems including learning the application of design patterns and the application of related technical topics. This was achieved by implementing a currently unpublished algorithm that is tentatively called Bakery Consensus. Bakery Consensus is a replicated state-machine protocol that can tolerate servers with Byzantine faults, but assumes non-faulty clients. The algorithm also establishes non-skipping timestamps for each operation completed by the replicated state-machine. The design of the structure, communication, and creation of the different system parts depended heavily upon the book Design Patterns [2]. After implementing the system, the success of the in implementing its various parts was based upon their ability to satisfy the SOLID [7] principles as well as their ability to establish low coupling and high cohesion [8]. The rest of this paper is organized as follows. We begin by providing background information about distributed algorithms, including replicated state-machine protocols and the Bakery Consensus algorithm. Section 3 gives a background on several design patterns and software engineering principles that were used in the development process. Section 4 discusses the well designed parts of the system that used design patterns, and how these design patterns were chosen. Section 5 discusses well designed system parts that relied upon other technical topics. Section 6 discusses system parts that need redesign. The conclusion summarizes what was accomplished by the implementation process and the lessons learned about design patterns for distributed algorithms.
ContributorsStoutenburg, Tristan Kaleb (Author) / Bazzi, Rida (Thesis director) / Richa, Andrea (Committee member) / Computer Science and Engineering Program (Contributor) / Barrett, The Honors College (Contributor)
Created2017-05