Skip to content

Roadmap

Vishwanath Grid is being built in three phases. The first phase is the pilot — one site, one scheduler, real researchers using it. The next phase brings two more universities on as partners. The third phase is about everything that sits on top of the cluster.

If you are landing on this page first, the short version: Vishwanath Grid is a research-computing cluster that several Indian universities run together. Everything in the stack is open-source. The first site is live at MIT-WPU Pune.

Phase 0 · Pilot, running today

The pilot is a small HTCondor pool at MIT World Peace University, Pune. HTCondor is the scheduler software — the program that takes a queue of researcher jobs and sends them to free machines. It is the same scheduler that runs at the CERN LHC computing grid, the NASA High-End Computing programme, and a long list of universities and research labs, so the day-to-day experience for a researcher is familiar.

The point of the pilot is not scale; it is learning. Running the real workflow — login, submit, wait, get output, repeat — at one site tells us what we will hit when we add the next two.

Phase 1 · In progress

Two more Indian universities are joining as partner sites. The work of this phase is the connecting tissue:

  • One identity across sites. A researcher who logs in at campus A can submit a job that runs on hardware at campus B.
  • Operations. Hardened scripts and dashboards so a single campus-side administrator can keep the partner site running without needing a dedicated team.
  • A partner-onboarding playbook. A documented, repeatable set of steps for adding the next university.

This is the phase where "federation" becomes real, rather than a single site with the federation software switched off.

Phase 2 · Planned

Once the federation is stable, the platform becomes the home for the everyday tools a research group uses:

  • A web page to submit and monitor jobs — no SSH terminal needed.
  • Interactive Jupyter notebooks and remote desktops, running on the same hardware as the batch jobs.
  • Code hosting, project tracking, collaborative documents, group chat — all under the same single login.
  • Dashboards that partner administrators can use to audit their own site's usage.

These tools are well-established open-source projects in their own right. The work in Phase 2 is connecting them properly — one login, one stable URL per tool, one playbook to install them at a new partner site.

How this is funded

Phase 0 is funded and operated by the MIT-WPU research-computing group.

Conversations about Phase 1 funding are in progress with academic partners and research-infrastructure funders. If you would like to be part of those conversations, get in touch. We are happy to walk through the project in detail on a call.