In a concise statement, provide a summary of your ...



Explain the circumstances such as a change in discipline, advisors, or school, which you believe qualify you for being considered to be in the early stages of their graduate studies. Your response will be limited to 1,500 characters, including spaces.

I emailed the committee for clarification on this question.

In a concise statement, provide a summary of your educational program objectives and your long-range professional goals. As part of this statement, we are interested in your ideas about: (1) the kinds of research in which you would like to be engaged during your graduate study or in the longer term; and/or (2) specific research questions that interest you and how you became interested in them. Please discuss these research interests in sufficient detail for an expert who is technically competent in your field to judge your understanding of the questions to be addressed, relevant hypotheses and approaches one might take to answering the questions, and other research principles required to investigate in the research area you identify. Your response will be limited to 3,000 characters, including spaces.

Traditional methods for building distributed, real-time and embedded (DRE) systems, such as the total ship computing environment (TSCE) or avionics mission planning, involved building the system as stovepipes directly on top of hardware, operating system and network protocols. This methodology created many accidental complexities that made applications brittle, inflexible, non-portable and incapable of evolving to overcome obsolete technology. To address these limitations, traditional methods were replaced with various layers of reusable middleware solutions, e.g., host and distributed middleware, where each layer provided a specific service to shield developers from the arising complexities arising out of network and OS heterogeneity, such as communicating between hosts with different endianness and messaging formats. When these layers are combined, they permit the creation of reusable components under what is known as a Service-Oriented Architecture (SOA). SOAs eliminate many complexities associated with traditional methods because technologies do not have to be re-invented, and system architectures become more adaptive, evolvable and flexible. Consequently, SOAs shift some complexities to system engineers, who are responsible for deploying - placing components on nodes - and configuring - choosing parameters to ensure quality of service (QoS) - components of DRE systems. This shift, however, has also prolongeds the integration phase of large-scale DRE systems and had adversely impacts on project costs and schedules. Therefore, as part of my graduate research and long-term R&D activities, I am interested in investigatingresearching methods and processes to assist system developers within minimizing cost, effort and time spent integrating systems by developing techniques for qualitative and quantitative evaluation of system and infrastructure QoS at development time.

I became interested in this area when I was a visiting researcher at Raytheon in Portsmouth, RI in summer 2005 working on a DARPA ARMS funded project called the Component Workload Emulator (CoWorkEr) Utilization Test Suite (CUTS). The main objective of CUTS is to combine model-driven development tools with real-time SOAs, e.g., TAO and CIAO, which are real-time object broker and component middleware technologies used in many DoD projects, to emulate and evaluate DRE systems QoS. I realized the need for such athis technology when developing large-scaled DRE systems, so I committed full-time. My preliminary experience and continued involvement on the DARPA ARMS project raises I have also raised several questions that motivate a questions that show strong need for continuing R&D on CUTS as a result of my preliminary experience and continued involvement on the DARPA ARMS project. First, how can CUTS be extended to evaluate real implementations in conjunction with emulated component behavior? Second, how well do emulated real implementations performance compare to real ones? Third, how can CUTS be extended to test concerns, such as fault tolerance and security of DRE systems? These are some questions I want to answer with the goals of developing a tool for minimizing cost, effort and time spent integrating large-scale systems that depend on and to perform according to requirements.

In a concise statement provide a summary of any educational and personal experiences you consider relevant to your stated goals and objectives, such as any scientific research activities, participation in relevant volunteer and extracurricular activities, professional work experience or work-training experience. Your response will be limited to 2,000 characters, including spaces.

My experience as a visiting researcher at NASA’s Jet Propulsion Laboratory (JPL) in Pasadena, CA and Raytheon Integrated Defense Systems (IDS) in Portsmouth, RI, and my decision to attend Vanderbilt University for graduate studies are most relevant to my stated objectives and goals.

When I was at NASA’s JPL in summer 2004, I was introduced to distributed, real-time and embedded (DRE) systems and CORBA programming. I also had committed to Vanderbilt for graduate school and met several professors, including Dr. Douglas C. Schmidt and Dr. Aniruddha (Andy) Gokhale, who were involved in path breaking R&D activities supported by DoD funded projects (DARPA PCES, DARPA ARMS) to apply model-driven development techniques and real-time middleware to defense systems. While I was at JPL, I learned that my supervisor, Tom Lockhart, and Doug were colleagues. This connection allowed Andy, Doug, and Tom to design a summer project that would be beneficial to my research experience since I was a prospect for Andy and Doug’s research group at Vanderbilt.

My experience at Raytheon is relevant because the research I started in summer 2005 has become the main research topic for my graduate studies. While at Raytheon, I began developing a tool called the Component Workload Emulator (CoWorkEr) Utilization Test Suite (CUTS). The main objective of CUTS is to allow system developers and engineers to evaluate component-based DRE systems prior to system integration to minimize cost, effort and time spent in full integration. Our preliminary efforts have shown proof-of-concept for this technique to evaluate DRE systems and we are continuing its R&D efforts. I have also co-authored a peer-reviewed paper that has been accepted to 13th IEEE International Conference and Workshop on the Engineering of Computer Based Systems, March 2006; and co-authored a paper submitted for peer-review to the 18th Euromicro Conference on Real-Time Systems, July 2006.

................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download