UWSpace

UWSpace is the University of Waterloo’s institutional repository for the free, secure, and long-term home of research produced by faculty, students, and staff.

Depositing Theses/Dissertations or Research to UWSpace

Are you a Graduate Student depositing your thesis to UWSpace? See our Thesis Deposit Help and UWSpace Thesis FAQ pages to learn more.

Are you a Faculty or Staff member depositing research to UWSpace? See our Waterloo Research Deposit Help and Self-Archiving pages to learn more.

Photo by Waterloo staff

Communities in DSpace

Selecteer een community om diens collecties te verkennen.

Recent Submissions

  • Thumbnail Image
    Item type: Item ,
    Hermes: Dynamic partitioning for distributed social network graph databases
    (University of Waterloo, 2015-02) Nicoara, Daniel; Kamali, Shahin; Daudjee, Khuzaima; Chen, Lei
    Social networks are large graphs that require multiple graph database servers to store and manage them. Each database server hosts a graph partition with the objectives of balancing server loads, reducing remote traversals (edge-cuts), and adapting the partitioning to changes in the structure of the graphic in the face of changing workloads. To achieve these objectives, a dynamic repartitioning algorithm is required to modify an existing partitioning to maintain good quality partitions while not imposing a significant overhead to the system. In this paper, we introduce a lightweight repartitioner, which dynamically modifies partitioning using a small amount of resources. In contrast to the exist repartitioning algorithms, our lightweight repartitioner is efficient, making it suitable for use in a real system. We integrated our lightweight repartitioner into Hermes, which we designed as an extension of the open source Neo4j graph database system, to support workloads over partitioned graph data distributed over multiple servers. Using real-world social network data, we show that Hermes leverages the lightweight repartitioner to maintain high quality partitions and provides a two to three times performance improvement over the de-facto standard random hash-based partitioning.
  • Thumbnail Image
    Item type: Item ,
    Improving time-of-use electricity pricing in Ontario
    (University of Waterloo, 2016-09-20) Adepetu, Adedamola; Keshav, Srinivasan
    Time-of-Use (ToU) electricity pricing is an electricity pricing scheme where consumers are charged at a rate that is dependent on the time of electricity consumption. This pricing scheme is often implemented to match the cost of generating and supplying electricity, and to make consumers defer appliance usage; this would reduce the daily electricity consumption peak that can both reduce the cost of generation and carbon footprints. We first critique the current ToU scheme in Ontario and make recommendations to improve it. Subsequently, we create an Agent-Based Model (ABM) to study ToU pricing and its effectiveness in reducing peak loads, which allows us to evaluate the benefit of our recommendations. We find that while ToU is effective in incentivizing load deferral, improvements can be made in the Ontario ToU scheme.
  • Thumbnail Image
    Item type: Item ,
    Feature-oriented modelling in BIP: A case study
    (University of Waterloo, 2016-09-20) Bocovich, Cecylia; Atlee, Joanne
    In this paper, we investigate the usage of Behaviour-Interaction-Priority version 2 (BIP2), a component-based modelling framework, for specifying feature-oriented systems. We evaluate BIP2 in the context of the Feature Interaction Problem and quantify the amount of work needed to add features to an existing system (i.e., in terms of rework to existing features, and work to identify and specify interactions). We present the results of a case study on a telephony system with five optional features where we found that the amount of work depends heavily on how features are interconnected. We identify a number of different strategies for interconnecting features, and propose one that reduces the amount of work and rework needed to add new features to an existing system.
  • Thumbnail Image
    Item type: Item ,
    On referring expressions in information systems derived from conceptual models
    (University of Waterloo, 2016-04-28) Borgida, Alexander; Toman, David; Weddell, Grant
    We apply recent work on referring expression types to the issue of identification in conceptual modelling. In particular, we consider how such types yield a separation of concerns in a setting where an information system based on a conceptual schema is to be mapped to a relational schema plus SQL queries. We start from a simple object-centered representation (as in semantic data models), where naming is not an issue because everything is self-identified (possibly using surrogates). We then allow the analyst to attach to every class a preferred "referring expression type", and to specify uniqueness constraints in the form of generalized functional dependencies. We show (1) how a number of well-formedness conditions concerning an assignment of referring expressions can be efficiently diagnosed, and (2) how a concrete relational schema and SQL queries over this schema are derived from a combination of the conceptual schema and queries over it, once identification issues have been separately resolved as above.
  • Thumbnail Image
    Item type: Item ,
    Distributed data deduplication
    (University of Waterloo, 2016-02-01) Chu, Xu; Ilyas, Ihab; Koutris, Paraschos
    Data deduplication refers to the process of identifying tuplets in a relation that refers to the same real world entity. The complexity of the problem is inherently quadratic with respect to the number of tuples, since a similarity value must be computed for every pair of tuples. In order to avoid comparing tuple pairs that are obviously non-duplicates, matching algorithms using blocking techniques that divide the tuples into blocks and compare only tuples within the same block. However, even with the use of blocking, data deduplication remains a costly problem for large datasets. In this paper, we show how to further speed up data deduplication by leveraging parallelism in a shared-nothing computing environment. Our main contribution is a distribution strategy, called \disdedup, that minimizes the maximum workload across all worker nodes and provides strong theoretical guarantees. We demonstrate the effectiveness of our proposed strategy by performing extensive experiments on both synthetic datasets with varying block size distributions, as well as real world datasets.