Browsing by Author "Carla D. Savage, Committee Member"

Now showing 1 - 5 of 5

Hierarchical Traffic Grooming in Large-Scale WDM Networks
(2005-06-14) Chen, Bensong; George N. Rouskas, Committee Chair; Matthias F. M. Stallmann, Committee Member; Carla D. Savage, Committee Member; Rudra Dutta, Committee Co-Chair
The advances in fiber optics and wavelength division multiplexing (WDM) technology are viewed as the key to satisfying the data-driven bandwidth demand of today's Internet. The mismatch of bandwidths between user needs and wavelength capacity makes it clear that some multiplexing should be done to use the wavelength capacity efficiently, which will result in reduction on the cost of line terminating equipment (LTE). The technique is referred to as traffic grooming. Previous studies have concentrated on different objectives, or on some special network topologies such as rings. In our study, we aim at minimizing the LTE cost to directly target on minimizing the network cost. We look into the grooming problem in elemental topologies as a starting point. First, we conduct proofs to show that traffic grooming in path, ring and star topology networks with the cost function we consider is NP-Complete. We also show the same complexity results for a Min-Max objective that has not been considered before, on the two elementary topologies. We then design polynomial-time heuristic algorithms for the grooming problem in rings (thus implicitly paths) and stars for networks of larger size. Experiments on various network sizes and traffic patterns show the effectiveness of our algorithms. For general topology networks, we design a hierarchical approach which first partitions a large network into several clusters, then applies the method we use in star networks to each cluster, by selecting a hub node to groom traffic within the cluster. At the second hierarchy, we apply the star grooming method again only on the hub nodes. The polynomial-time hierarchical approach scales well and can cope with large networks of general topology efficiently, both for minimizing LTE cost and for lowering wavelength requirements. We also design a clustering algorithm that can generate good results for subsequent steps in the hierarchical grooming method. Numerical results from experiments show that both the hierarchical grooming approach and our clustering algorithm generate satisfying results for the grooming on various network topologies and traffic demand patterns.
The Internet Topology: Illusion and Reality
(2008-12-04) Kim, Sangmin; Khaled Harfoush, Committee Chair; Arne A. Nilsson, Committee Member; Carla D. Savage, Committee Member; George N. Rouskas, Committee Member
Research studies targeted at unveiling the Internet topology are essential for understanding the performance of the Internet and its resilience to failures or distributed attacks, and for generating realistic topologies to simulate Internet protocols and applications. An accurate understanding of the complex Internet structure and behavior, while very rewarding, is very challenging and in fact is a source of controversy in the networking research community. Till now, there is a lack of understanding of the Internet complexity In this thesis, we make the following contributions. First, we propose an efficient tool, AROMA, to unveil Layer-3 maps of the Internet and use it to reveal ISP maps. AROMA reveals the same number of routers and links as existing tools such as Rocketfuel after sending less than 5.1% of the number of probes used by Rocketfuel, and reveals at least 100% more links and routers than Rocketuel while using the same number of probe packets. Second, we study the limitations of existing layer-3 tools such as traceroute in unveiling the details of the Internet structure and identify that the power law connectivity observed in the Internet topology is not an illusion as suggested by some researchers. It is mainly manifested due to the blindness of traceroute to layer-2 devices, and this manifestation will persist independent of the nature of the underlying physical topology. Third, we provide a realistic Internet topology model, HINT, which captures the Internet structure and features. HINT is based on economical, performance and security constraints that are typically used to construct networks. Matching HINT topologies to known ISP topologies confirms its superiority to existing Internet topology models.
Modeling, Predicting, And Optimizing Parallel Performance Of Grid Stuctured Problems
(2004-06-04) Smith, Frank Anderson; Jeffrey S. Scroggs, Committee Member; Edward W. Davis, Committee Member; Rex A. Dwyer, Committee Member; Robert E. Funderlic, Committee Chair; Carla D. Savage, Committee Member
The performance of parallel computers can be greatly affected by a user's choices of data distribution and logical processor configuration. Selecting optimal choices for such user specifiable parameters may be easier if the performance of the target machine can be predicted by a performance model. Models for parallel performance on the IBM SP for grid structured problems are considered. Such problems are ubiquitous in scientific computing and frequently are characterized by a nearest neighbor communication pattern. Bounds are derived for the size of the solution space of data distributions and logical processor configurations for problems with nearest neighbor communication. Proofs are derived that exclude a substantial number of non-optimal choices of data distribution and logical processor configuration. Algorithms are given that are intended to predict parallel performance for a model application and allow the user to select optimal choices for parameters that can be specified. Experimental evidence is presented that suggests that performance on the SP is characterized fairly accurately by a specific model. Experimental evidence also suggests that an algorithm exists to optimize a user's choices of data distribution and logical processor configuration for grid structured problems on the SP.
A Price Trajectory Algorithm for Solving Iterative Auction Problems
(2006-12-11) Zhong, Jie; Carla D. Savage, Committee Member; Yahya Fathi, Committee Member; Shu-Cherng Fang, Committee Member; Peter R. Wurman, Committee Chair
A variety of auctions exist in the literature such as the English auction, the Dutch auction, and the Vickrey auction. The underlying problem in an auction is to find the winners and the corresponding payments. Proxy bidding has proven useful in solving auction problems in many real–world auction formats, most notably eBay. It has been proposed for several iterative combinatorial auctions, such as the Ascending Package auction, the Ascending k-Bundle auction, and the iBundle auction. In this dissertation, a new type of iterative auction called the Simple Combinatorial Proxy auction is proposed. The winners of the new auction are the same as that of the Ascending k-Bundle auction. Simulating the incremental bidding decisions of the agents is a popular method to solve proxy-enabled version of the auction problems. This approach has some disadvantages. First, the outcome depends upon implementation details. Second, the accuracy of the outcome relies on the bid increment. Third, the running time is sensitive to the magnitude of values, the ordering of agents, and the tie–breaking rules. In this dissertation, a new approach called the Price Trajectory Algorithm is presented to solve iterative combinatorial auctions with proxy bidding. This approach computes the agents' allocation of their attention across the bundles only at "inflection points" — the points at which agents change their behavior. Inflections are caused by one the following reasons: (1) an introduction of a new bundle into an agent's demand set, (2) a change in the set of current competitive allocations, or (3) a withdrawal of an agent from the set of active agents. The proposed algorithm tracks the behavior of agents and the competitive allocations of items to establish a connection between the demand set and competitive allocations. With the allocation of agents' attention, one can compute the slopes of price curves to get the bundle prices and speed up the computation by jumping from one inflection point to the next. The price trajectory algorithm can solve the Simple Combinatorial Proxy Auction and the Ascending Package Auction. It has several advantages over alternatives: (1) it computes exact solutions; (2) the solutions are independent of the bid increment or tie-breaking rules; and (3) the solutions are invariant to the magnitude of the bids. For the security consideration, a cryptographic protocol is presented for the price trajectory algorithm. It guarantees that only the auctioneer obtains the correct and necessary information from the agents and there is no leak of private information between agents. The detection of fraud by the auctioneer is also discussed.
Security Mechanisms for Wireless Sensor Networks
(2005-06-28) Liu, Donggang; Mladen A. Vouk, Committee Member; Carla D. Savage, Committee Member; Douglas S. Reeves, Committee Member; Peng Ning, Committee Chair
Wireless sensor networks have received a lot of attention recently due to its wide applications in military and civilian operations. Example applications include target tracking, scientific exploration, and data acquisition in hazardous environments. Security becomes one of the main concerns when there are malicious attacks against the network. However, providing security services in such networks turns out to be a challenging task due to the resource constraints on sensor nodes and the node compromise attacks. These features and challenges motivate the research on security mechanisms for wireless sensor networks. This dissertation includes three studies on security mechanisms for wireless sensor networks. The first study extends the capabilities of $mu$TESLA, a broadcast authentication technique for wireless sensor networks, so that it can cover long time period and support a large number of sensor nodes as well as potential senders in the network. The second study addresses how to establish pairwise keys between sensor nodes in a wireless sensor network. A key pre-distribution framework based on bivariate polynomial pool is developed for this purpose. Two efficient instantiations of this framework are also provided: a {em random subset assignment} scheme and a {em hypercube-based} key pre-distribution scheme. To further improve the pairwise key establishment in static sensor networks, prior deployment knowledge, post deployment knowledge and group-based deployment knowledge are used to facilitate key pre-distribution. The third study investigates how to enhance the security of location discovery in sensor networks. An attack-resistant MMSE method and a voting-based method are developed to tolerate malicious attacks against location discovery. Both methods can survive malicious attacks even if the attacks bypass traditional cryptographic protections such as authentication, as long as the benign beacon signals constitute the majority of the 'consistent' beacon signals. In addition, a number of techniques are proposed to detect and revoke malicious beacon nodes that supply malicious beacon signals to sensor nodes.