Virtual Repository Testbed Server Project<br/>Fall 2008

The Software Studio projects are always large, complex, distributed systems. Each year we experiment with some interesting way of building these kinds of systems. When appropriate, we may use or adapt some idea from previous projects, but always have some new thing to try.

This year we built the project concept around a set of virtual servers that support teams collaborating on development of software systems. The Repository, Testbed, and Build servers are each instances of a virtual server. The idea of a virtual server is that it provides, within its own infrastructure, the means to replicate itself on any machine registered with the system. So, for example, if a team wants to have its own local copy of a Repository, that is just a few clicks away. Virtual servers support a common set of communication and data management protocols that support the many activities required for software development.

All the servers and their tools communicate through Windows Communication Foundation services, that define message-passing channels to provide a message handling core, used by all the tools. Use and Administration of the system is orchestrated through a Windows Presentation Foundation based Client, that uses the power of this framework to manage a complex process (creation, management, and testing of a very large number of software components) in a user friendly way.

The Developers of this system are the students enrolled in CSE784 - Software Studio, during the Fall of 2008. This system was specified, designed, implemented, tested, integrated, and qualified in a period of 9 weeks. The qualification test consisted of the verification of approximately 250 requirements via more than 100 carefully written test procedures. The system was qualified, in a little more than eight hours, on December 8th, 2008.

I have been fortunate over the years to work with many bright, enthusiastic, and dedicated students who work very hard to bring these projects to a successful conclusion. This year is no exception. The team was very well managed, some interesting ideas were shown to be feasible and worth the investment to make them part of a large useful system. And everyone worked effectively to make that happen.

I am proud of this class and have enjoyed being part of their experience with the Virtual Repository Testbed System. Project Manager - Phil Pratt-Szeliga Software Architect - Dusan Palider Test Team - Madhumita Nagel, Leader
Devaprem Banerjee (also Virtal Build Server)
Gauri Bindu
Hitendrakumar Modi
Sonam Neema
Deepashree Prabhakar
Somashekhar Kotagi (also Comm Team)
Zhe Zhao VRTS Tools & Testing Tools - Rajaa Alqudah, Leader
Archana Pilaipakkam
Paul Petzke
Mark Webb VRTS Client and Portal - Mohammed Koni, Leader
Manasavalli Vidali
Shashikiran Srinivasa Virtual Repository Server - Gokul Sivajii, Leader
Nidhi Solanki
Sarvesh Devi
Robert Durham
Anway Shahane Virtual Testbed Server - Devendra Chauhan, Leader
Siddharth Damani
Anuhya Kodali
Sohil Shah Virtual Build Server - Devaprem Banerjee, Leader
Sudheer Anantha
Nikila Iyer
Mitali Patel Communication - Somashekhar Kotagi, Leader
Sagar Gupta
Sambhav Jain
Pavana Reddy Browsing Service - Saikiran Permula, Leader
Sumanth Anand
Akshara Chaturvedi
Jay Sachaniya

When Dr. Fawcett first handed out the project statement in class I think I can speak for the whole class and say that we have never seen such a large assignment. That weekend Dusan Palider, the Software Architect, and I worked in the computer lab in Link making sense of the requirements and the preliminary architecture. Dusan was very good at this and I can see why Dr. Fawcett chose him for the job! We worked together and by the end of the weekend we had a better grasp of what was going on.

Dusan allocated the A-Specification requirements to each team and presented them at the Architectural Review. We then created an integration schedule that featured frequent, early integrations. At first we didn't really know how to integrate and often a team would not meet all of the planned features in the first integration. So, we made a policy that if integration was not met on the due date the teams would work in the lab everynight from 6pm to around 11pm or midnight until every feature was working and tested. I would like to thank every team leader (Madhumita Nagel, Devaprem Banerjee, Rajaa Alqudah, Mohammed Koni, Gokul Sivajii, Devendra Chauhan, Somashekhar Kotagi and Saikiran Permula) as well as all the team members for working this hard in the labs.

Each student in the class had his or her own specific task that they hopefully enjoyed. The task that I enjoyed the most was to reason about problems in a system as complex as this and attempt to suggest solutions. I found it very empowering to be put in this situation and I would like to thank Dr. Fawcett for the opportunity. Overall I think we learned what it takes to work on a large software development project (having a process, working together and working hard) and how to survive while doing so.

Qualification Start and Near the End

Repository and Testbed Teams

Build Server and Client teams

Browsing Service and Communication Teams

Part of Tools Team and Tools Testing

Testing the Testbed Server and Comm

Last Hurrah at 4:30am!