COLLABORATIVE TESTING ACROSS SHARED SOFTWARE COMPONENTS

Large component-based systems are often built from many of the same components. As individual component-based software systems are developed, tested and maintained, these shared components are repeatedly manipulated. As a result there are often significant overlaps and synergies across and among the different test efforts of different component-based systems. However, in practice, testers of different systems rarely collaborate, taking a test-all-by-yourself approach. As a result, redundant effort is spent testing common components, and important information that could be used to improve testing quality is lost. The goal of this research is to demonstrate that, if done properly, testers of shared software components can save effort by avoiding redundant work, and can improve the test effectiveness for each component as well as for each component-based software system by using information obtained when testing across multiple components. To achieve this goal I have developed collaborative testing techniques and tools for developers and testers of component-based systems with shared components, applied the techniques to subject systems, and evaluated the cost and effectiveness of applying the techniques. The dissertation research is organized in three parts. First, I investigated current testing practices for component-based software systems to find the testing overlap and synergy we conjectured exists. Second, I designed and implemented infrastructure and related tools to facilitate communication and data sharing between testers. Third, I designed two testing processes to implement different collaborative testing algorithms and applied them to large actively developed software systems. This dissertation has shown the benefits of collaborative testing across component developers who share their components. With collaborative testing, researchers can design algorithms and tools to support collaboration processes, achieve better efficiency in testing configurations, and discover inter-component compatibility faults within a minimal time window after they are introduced.

UNDERSTANDING CUSTOMER CHOICES IN SERVICE OUTSOURCING AND REVENUE MANAGEMENT

This dissertation investigates customer behavior modeling in service outsourcing and revenue management in the service sector (i.e., airline and hotel industries). In particular, it focuses on a common theme of improving firms’ strategic decisions through the understanding of customer preferences. Decisions concerning degrees of outsourcing, such as firms’ capacity choices, are important to performance outcomes. These choices are especially important in high-customer-contact services (e.g., airline industry) because of the characteristics of services: simultaneity of consumption and production, and intangibility and perishability of the offering. Essay 1 estimates how outsourcing affects customer choices and market share in the airline industry, and consequently the revenue implications from outsourcing. However, outsourcing decisions are typically endogenous. A firm may choose whether to outsource or not based on what a firm expects to be the best outcome. Essay 2 contributes to the literature by proposing a structural model which could capture a firm’s profit-maximizing decision-making behavior in a market. This makes possible the prediction of consequences (i.e., performance outcomes) of future strategic moves. Another emerging area in service operations management is revenue management. Choice-based revenue systems incorporate discrete choice models into traditional revenue management algorithms. To successfully implement a choice-based revenue system, it is necessary to estimate customer preferences as a valid input to optimization algorithms. The third essay investigates how to estimate customer preferences when part of the market is consistently unobserved. This issue is especially prominent in choice-based revenue management systems. Normally a firm only has its own observed purchases, while those customers who purchase from competitors or do not make purchases are unobserved. Most current estimation procedures depend on unrealistic assumptions about customer arriving. This study proposes a new estimation methodology, which does not require any prior knowledge about the customer arrival process and allows for arbitrary demand distributions. Compared with previous methods, this model performs superior when the true demand is highly variable.

Optical Wireless Communication for Mobile Platforms

The past few decades have witnessed the widespread adaptation of wireless devices such as cellular phones and Wifi-connected laptops, and demand for wireless communication is expected to continue to increase. Though radio frequency (RF) communication has traditionally dominated in this application space, recent decades have seen an increasing interest in the use of optical wireless (OW) communication to supplement RF communications. In contrast to RF communication technology, OW systems offer the use of largely unregulated electromagnetic spectrum and large bandwidths for communication. They also offer the potential to be highly secure against jamming and eavesdropping. Interest in OW has become especially keen in light of the maturation of light-emitting diode (LED) technology. This maturation, and the consequent emerging ubiquity of LED technology in lighting systems, has motivated the exploration of LEDs for wireless communication purposes in a wide variety of applications. Recent interest in this field has largely focused on the potential for indoor local area networks (LANs) to be realized with increasingly common LED-based lighting systems. We envision the use of LED-based OW to serve as a supplement to RF technology in communication between mobile platforms, which may include automobiles, robots, or unmanned aerial vehicles (UAVs). OW technology may be especially useful in what are known as RF-denied environments, in which RF communication may be prohibited or undesirable. The use of OW in these settings presents major challenges. In contrast to many RF systems, OWsystems that operate at ranges beyond a few meters typically require relatively precise alignment. For example, some laser-based optical wireless communication systems require alignment precision to within small fractions of a degree. This level of alignment precision can be difficult to maintain between mobile platforms. Additionally, the use of OW systems in outdoor settings presents the challenge of interference from ambient light, which can be much brighter than any LED transmitter. This thesis addresses these challenges to the use of LED-based communication between mobile platforms. We propose and analyze a dual-link LED-based system that uses one link with a wide transmission beam and relaxed alignment constraints to support a more narrow, precisely aligned, higher-data-rate link. The use of an optical link with relaxed alignment constraints to support the alignment of a more precisely aligned link motivates our exploration of a panoramic imaging receiver for estimating the range and bearing of neighboring nodes. The precision of such a system is analyzed and an experimental system is realized. Finally, we present an experimental prototype of a self-aligning LED-based link.

BUILDING EFFICIENT AND COST-EFFECTIVE CLOUD-BASED BIG DATA MANAGEMENT SYSTEMS

In today’s big data world, data is being produced in massive volumes, at great velocity and from a variety of different sources such as mobile devices, sensors, a plethora of small devices hooked to the internet (Internet of Things), social networks, communication networks and many others. Interactive querying and large-scale analytics are being increasingly used to derive value out of this big data. A large portion of this data is being stored and processed in the Cloud due the several advantages provided by the Cloud such as scalability, elasticity, availability, low cost of ownership and the overall economies of scale. There is thus, a growing need for large-scale cloud-based data management systems that can support real-time ingest, storage and processing of large volumes of heterogeneous data. However, in the pay-as-you-go Cloud environment, the cost of analytics can grow linearly with the time and resources required. Reducing the cost of data analytics in the Cloud thus remains a primary challenge. In my dissertation research, I have focused on building efficient and cost-effective cloud-based data management systems for different application domains that are predominant in cloud computing environments. In the first part of my dissertation, I address the problem of reducing the cost of transactional workloads on relational databases to support database-as-a-service in the Cloud. The primary challenges in supporting such workloads include choosing how to partition the data across a large number of machines, minimizing the number of distributed transactions, providing high data availability, and tolerating failures gracefully. I have designed, built and evaluated SWORD, an end-to-end scalable online transaction processing system, that utilizes workload-aware data placement and replication to minimize the number of distributed transactions that incorporates a suite of novel techniques to significantly reduce the overheads incurred both during the initial placement of data, and during query execution at runtime. In the second part of my dissertation, I focus on sampling-based progressive analytics as a means to reduce the cost of data analytics in the relational domain. Sampling has been traditionally used by data scientists to get progressive answers to complex analytical tasks over large volumes of data. Typically, this involves manually extracting samples of increasing data size (progressive samples) for exploratory querying. This provides the data scientists with user control, repeatable semantics, and result provenance. However, such solutions result in tedious workflows that preclude the reuse of work across samples. On the other hand, existing approximate query processing systems report early results, but do not offer the above benefits for complex ad-hoc queries. I propose a new progressive data-parallel computation framework, NOW!, that provides support for progressive analytics over big data. In particular, NOW! enables progressive relational (SQL) query support in the Cloud using unique progress semantics that allow efficient and deterministic query processing over samples providing meaningful early results and provenance to data scientists. NOW! enables the provision of early results using significantly fewer resources thereby enabling a substantial reduction in the cost incurred during such analytics. Finally, I propose NSCALE, a system for efficient and cost-effective complex analytics on large-scale graph-structured data in the Cloud. The system is based on the key observation that a wide range of complex analysis tasks over graph data require processing and reasoning about a large number of multi-hop neighborhoods or subgraphs in the graph; examples include ego network analysis, motif counting in biological networks, finding social circles in social networks, personalized recommendations, link prediction, etc. These tasks are not well served by existing vertex-centric graph processing frameworks whose computation and execution models limit the user program to directly access the state of a single vertex, resulting in high execution overheads. Further, the lack of support for extracting the relevant portions of the graph that are of interest to an analysis task and loading it onto distributed memory leads to poor scalability. NSCALE allows users to write programs at the level of neighborhoods or subgraphs rather than at the level of vertices, and to declaratively specify the subgraphs of interest. It enables the efficient distributed execution of these neighborhood-centric complex analysis tasks over largescale graphs, while minimizing resource consumption and communication cost, thereby substantially reducing the overall cost of graph data analytics in the Cloud. The results of our extensive experimental evaluation of these prototypes with several real-world data sets and applications validate the effectiveness of our techniques which provide orders-of-magnitude reductions in the overheads of distributed data querying and analysis in the Cloud.

User Behavioral Modeling of Web-based Systems for Continuous User Authentication

Authentication plays an important role in how we interact with computers, mobile devices, the web, etc. The idea of authentication is to uniquely identify a user before granting access to system privileges. For example, in recent years more corporate information and applications have been accessible via the Internet and Intranet. Many employees are working from remote locations and need access to secure corporate files. During this time, it is possible for malicious or unauthorized users to gain access to the system. For this reason, it is logical to have some mechanism in place to detect whether the logged-in user is the same user in control of the user's session. Therefore, highly secure authentication methods must be used. We posit that each of us is unique in our use of computer systems. It is this uniqueness that is leveraged to "continuously authenticate users" while they use web software. To monitor user behavior, n-gram models are used to capture user interactions with web-based software. This statistical language model essentially captures sequences and sub-sequences of user actions, their orderings, and temporal relationships that make them unique by providing a model of how each user typically behaves. Users are then continuously monitored during software operations. Large deviations from "normal behavior" can possibly indicate malicious or unintended behavior. This approach is implemented in a system called Intruder Detector (ID) that models user actions as embodied in web logs generated in response to a user's actions. User identification through web logs is cost-effective and non-intrusive. We perform experiments on a large fielded system with web logs of approximately 4000 users. For these experiments, we use two classification techniques; binary and multi-class classification. We evaluate model-specific differences of user behavior based on coarse-grain (i.e., role) and fine-grain (i.e., individual) analysis. A specific set of metrics are used to provide valuable insight into how each model performs. Intruder Detector achieves accurate results when identifying legitimate users and user types. This tool is also able to detect outliers in role-based user behavior with optimal performance. In addition to web applications, this continuous monitoring technique can be used with other user-based systems such as mobile devices and the analysis of network traffic.