Summer 2008

Participants

 REU 091

Summer 2008 REU students (the two front rows, from left to right: Marlon Bright, Camilo Silva, Jairo Pava, Mansi Gupta, Paola Boettner, Robert Parks, Seychelles Martinez, Mike Torchio, Courtney Enoex, and Elias Rodriguez) and faculty mentors (the back row, from left to right: Peter Clarke, Raju Rangaswami, and Masoud Sadjadi)

Name

Home Institution

Year

Mentor

Paola Boettner

Wellesley College

Junior

Dr. Peter J. Clarke

Marlon Bright

Florida International University

Junior

Dr. Masoud Sadjadi

Courtney Enoex

Southern Illinois University

Junior

Dr. Peter J. Clarke

Mansi Gupta

Bryn Mawr

Sophomore

Dr. Peter J. Clarke

Seychelles Martinez

Florida International University

Junior

Dr. Masoud Sadjadi

Robert Parks

California State University, Fresno

Junior

Dr. Raju Rangaswami

Jairo Pava

Florida International University

Sophomore

Dr. Peter J. Clarke

Elias Rodriguez

Polytechnic Univ. of P.R.

Junior

Dr. Masoud Sadjadi

Camilo Silva

Florida International University

Junior

Dr. Masoud Sadjadi

Mike Torchio

The University of Vermont

Sophomore

Dr. Raju Rangaswami

Activities

Research and Education:

Dates

Activities

Coordinator(s)

May 19-23

  • Welcome
  • Tour of School and University Facilities
  • Presentation - Introduction to Autonomic Computing
  • Presentation - What is Research?
  • Assignment of Projects

Dr. Milani, Ms. Dona Dorsett, Ms. Catherine Hernandez, and faculty mentors



May 26

Start of Research Activities

Faculty Mentors

June 19

Progress Presentation 1

Dr. Milani

July 14

Progress Presentation 2

Dr. Milani

August 1

Final Project Presentation

Dr. Milani

Other Activities:

Dates

Activities

Coordinator(s)

Regularly

Visit to Bowling Alley

 

 

Visit to Coconut Grove

Dr. Milani

July 4

Independence Day Party

Dr. Clarke

July 12

Snorkeling Trip

Drs. Milani, Rangaswami, and Sadjadi

Projects

  1. Title: Developing User Policies for Configuring Communication Frameworks

Participants: Paola Boettner, Mansi Gupta, Andrew Allen (Ph.D. Student), and Dr. Peter Clarke (faculty mentor)

Project Description: This work was motivated by the need for communication customers to have a way to specify a goal oriented policy that states their preferences with respect to the type of communication service provided by the underlying multiple frameworks.   This project investigated how user-defined communication policies may be used to guide self-configuration of communication frameworks for the Communication Virtual Machine (CVM).  This investigation involved performing a detailed feature analysis on communication frameworks such as Skype, GoogleTalk, Smack, Windows Live Messenger and Yahoo! Messenger.    The initial survey for this work was performed by the students in the REU summer 2007 program.  Based on the feature analysis of the communication frameworks the structure of an xml-based policy was developed to capture user-defined preferences.  In addition a graphical user interface was developed to support the easy capture of the preferences from the user.

Presentations:

·         First Presentation, June 19, 2008

·         Second Presentation, July 14, 2008

·         Third Presentation, August 1, 2008

Publications:

·         Paola Boettner, Mansi Gupta, Yali Wu and Andrew Allen. Towards Policy Driven Self-Configuration of User-Centric Communication.  Under preparation for the 47th ACM Southeast Conference, ACMSE '09 (Student paper).

 

  1. Title: Applying the Principles of Autonomic Computing to Regression Testing of Web-Based Applications

Participants: Jairo Pava, Courtney Enoex, Yanelis Hernandez (Ph.D. Student), and Dr. Peter Clarke (faculty mentor)

Project Description: Regression testing of software applications continues to be a very expensive endeavor during maintenance of these applications.  This project investigated how the principles of autonomic computing may be applied to regression testing of web-based applications.  The research resulted in an autonomic web application test harness that dynamically configures its test suites to generate and execute platform specific test scripts based on the web technologies implemented by a web application. An autonomic test manager and the structure that is needed to automatically port regression tests to various platforms were defined.  The approach used to develop the autonomic test manager uses a model-driven engineering approach where meta-models for the web applications and the test scripts used during regression testing were developed.

A prototype of the test harness for an online e-commerce web application was also created to validate the practicality and feasibility of the approach.  Currently, the prototype implements this approach for web applications written in (1) Ruby on Rails and (2) PHP on the server side.  Both these applications use HTML and Java script on the client side.

Presentations:

·         First Presentation, June 19, 2008

·         Second Presentation, July 14, 2008

·         Third Presentation, August 1, 2008

Publications:

·         Jairo Pava, Courtney Enoex and Yanelis Hernandez.  An Autonomic Testing Harness For Web Applications.  Under preparation for the 47th ACM Southeast Conference, ACMSE '09 (Student paper).

 

 

  1. Title: Building Self-managing Storage Systems with Active Block Layer Extensions

Participants: Robert Parks, Mike Torchio, Jorge Guerra (PhD student), Dr. Raju Rangaswami (faculty mentor)

Project Description: The Active Block Layer Extensions (ABLE) project proposes to develop a systematic approach to the development, deployment and management of self-managing storage system extensions. During the summer, we worked on two complementary components of the ABLE project – simulated storage devices, and persistent data structures. Robert designed and developed a simulated storage device abstraction inside the Linux operating system kernel. This abstraction simplifies the development (specifically debugging and testing) of self-management extensions. We designed and developed three classes of simulated block-level devices which is a component within the larger ABLE project. These devices aid developers of storage extensions in testing their extensions in a non-intrusive way. The three classes of devices that Robert developed were (i) fully simulated, (ii) memory simulated, and (iii) emulated. Additionally, we worked on developing a model for the recently introduced Solid-state Storage Devices to be integrated within the three classes of device simulation alongside disk drives. Mike designed and developed a persistent data structures library inside the operating system that allows easy development of stateful self-management extensions that can much more easily survive system and power failures. With persistent data structures, operating system extensions can use traditional memory allocation and management primitives for managing data structures which they wish to be persistent. Further this capability also provides extensions with knobs for tuning the synchronization policies for each persistent data structure.

Presentations:

·         Robert Parks’ Presentation, Simulated Storage Devices, August 1, 2008

·         Mike Torchio’s Presentation, Persistent Data Structures, August 1, 2008

Publications:

·          

 

  1. Title: Application Performance Profiling and Prediction in Grid Environment

Participants: Marlon Bright, Javier Delgado (PhD student), Dr. Masoud Sadjadi (faculty mentor)

Project Description: In a Grid computing environment, resources are shared among a large number of applications. Brokers and schedulers find matching resources and schedule the execution of the applications by monitoring dynamic resource availability and employing policies such as first-come- first-served and back-filling. To support applications with timeliness requirements in such an environment, brokering and scheduling algorithms must address an additional problem - they must be able to estimate the execution time of the application on the currently available resources. In this project, we investigate to formulate a model that can estimate the execution time of long-running scientific applications. The modeling approach we propose is generic; models can be constructed by merely observing the application execution “externally” without using intrusive techniques such as code inspection or instrumentation. The model is cross-platform; it enables prediction without the need for the application to be profiled first on the target hardware. We focused on modeling the behavior of one specific scientific application, namely WRF, as a starting point but we believe that our approach can be used to model the execution time of time-sensitive scientific applications; thereby, enabling the development of more intelligent brokering and scheduling algorithms.

Presentations:

·         First Presentation, June 19, 2008

·         Second Presentation, July 14, 2008

·         Third Presentation, August 1, 2008

Publications:

·          

 

  1. Title: Weather Research and Forecasting (WRF) Portal

Participants: Seychelles Martinez, Elias Rodriguez, Eric Meyer (PhD Student), and Dr. Masoud Sadjadi (faculty mentor)

Project Description: The motivation for creating this portal is to provide a comprehensive, easy to use, high-level tool for meteorologist to simulate and perform ensemble runs using the WRF forecasting model. Likewise the idea is to help scientists visualize the output statically or dynamically via MPEG/GIF or Google maps with KML layered over it with relevant information where one can zoom in and out, respectively. This will help the meteorologists, who are currently occupied with the particulars of the computer science aspect of weather forecasting, concentrate on their own science. With a readily available tool created for them, a meteorologist does not have to be bothered by the configuration of the software; much in the same way a person who uses a telephone does not need to be aware of the circuitry inside to use it to his or her advantage. In this way, the science of meteorology will not be hindered by the cumbersome task of configuring tools themselves, but rather can concentrate on their own science and achieve research and other scientific goals quicker and more efficiently.

Presentations:

·         First Presentation, June 19, 2008

·         Second Presentation, July 14, 2008

·         Third Presentation, August 1, 2008

Publications:

·          

 

  1. Title: Finding Discriminating DNA Probe Sequences by Implementing a Parallelized Solution in a Cluster

Participants: Camilo Silva, Michael Robinson (PhD Student), Dr. Masoud Sadjadi (faculty mentor), and Dr. Giri Narasimhan (faculty mentor)

Project Description: The problem of identifying subsequences that are either repeated or unique has many applications in Bioinformatics. A number of algorithms and implementations exist for these problems, but are compute-intensive and on low-end machines would take a long time to respond. To accelerate the rate of findings in Biology, in this project we investigate an approach that speeds up the online response time. In our approach, we utilize an algorithm based on suffix arrays for these problems, which was previously developed by Michael Robinson, to preprocess the data and create necessary indexes off-line. This preprocessing allows for much faster response time at runtime. We then investigate an enhancement of this algorithm to run on MPI clusters to reach to acceptable response time using many compute nodes available in a cluster computing environment.

Presentations:

·         First Presentation, June 19, 2008

·         Second Presentation, July 14, 2008

·         Third Presentation, August 1, 2008

Publications:

·          

 

 

More Photos

REU 028

Summer 2008 REU Participants (from the back row and from left to right: Dr. Peter Clarke, Dr. Raju Rangaswami, Dr. Masoud Sadjadi, Paola Boettner, Mansi Gupta, Seychelles Martinez, Courtney Enoex, Robert Parks, Elias Rodriguez, Dr. Masoud Milani, and Jairo Pava)

 

REU 071

Marlon Bright’s live video is projected on the screen, attending and presenting his work from Barcelona Supercomputing Center, Barcelona, Spain. Camilo Silva is attending the meeting remotely also from University de Guadalajara, Mexico.

 

REU 086

2008 REU Participants including the REU students, graduate students, and faculty mentors (we are missing Dr. Milani and Dr. Tao Li in this picture).