You are here

Open Research Topics

This page provides a list o open reasearch topics for students interested in joining the projects. This list is by no means exaustive and is meant to provide some guidance for new students serching for a research theme. If you would like to work on a research topic not listed here, you are welcome to suggest this topic to the professors involved in the project.

  • TCC or Undergraduate Research

    • Parallel applications for InteGrade
      • Development of computacionally intensive parallel applications that uses and produces large amounts of data. The student is expected to perform a number of experiments over InteGrade using the developed application.
  • Masters

    • Design and implementation of fault-tolerance mechanisms for the InteGrade system, i.e., the middleware should continue to work seamlessly when machines running the servers (GRM, AR, GSM, etc.) fail. Our assumption here is that once in a while (for example, once a week, the machines could fail). Note: the cluster federation should continue working.

    • Support for additional parallel programming models on top of InteGrade and comparative experimental evaluation of different models.
    • Use the knowledge of the inter-cluster grid topology and of the heterogeneity of available resources to support smart schedulling strategies on the grid.
    • Support for wireless and mobile access to the grid.
    • Integration with other grid architectures (e.g. OGSA).
    • Support for parallel applications written on other programming languages (e.g. Fortran, C#); maybe support for multi-language parallel applications.
  • Doctoral

    • Design and implementation of a faul-tolerance mechanism for the InteGrade servers that would be able to cope with (non-availability) "failures" that happen in opportunistic grids, i.e., the servers (GRM, GSM, etc.) would execute on shared machines and several times per day the machine would become unavailable. The system would need to continue running seamlessly with tens (thousands) of grid machines becoming unavailable all the time. Note: the cluster federation should continue working.

    • Support for adaptive applications on top of the grid by using dynamic reconfiguration and reflection.
    • Adding adaptive capabilities to the middleware elements to cope with fluctuations on the wide-area grid environment.