Skip to content

Refuse For Research

Refuse For Research was a long-running, self-operated distributed computing project that repurposed decommissioned hardware to support public scientific research. The project ran from 2008 to 2013 and partnered with Saddle Creek Logistics Services, Inc. to contribute computing resources to humanitarian research projects using the World Community Grid.

This page provides context for the project and why it belongs here.

Why it existed

The project began as a practical response to a recurring problem: usable hardware being discarded simply because it no longer met production standards. Rather than treat this equipment as waste, Refuse For Research explored whether it could be repaired, pooled, and operated reliably enough to contribute meaningful computational work.

The goal was not experimentation for its own sake, but sustained participation in real research efforts under real constraints.

How it was run

Refuse For Research was not a simulation or a one-off demonstration. It was an operational system:

  • Decommissioned and surplus machines were repaired and standardized
  • Power, cooling, and network capacity were provisioned deliberately
  • Systems were monitored and maintained over multiple years
  • Failures were expected, logged, and corrected rather than hidden
  • Infrastructure was sized and adjusted as conditions changed

Most supporting infrastructure - racks, cabling, power distribution - was also repurposed or recovered, reinforcing the project’s emphasis on reuse and operational discipline.

Constraints and trade-offs

Operating the project required working within constraints that would be familiar in other production environments:

  • Non-uniform hardware
  • Limited physical space
  • Power and cooling boundaries
  • Equipment aging and attrition
  • The need to remain unattended for long periods

Design decisions favored stability and recoverability over optimization. Systems were kept simple enough to be repaired, reimaged, or replaced without specialized intervention.

Outcomes

Over its lifetime, Refuse For Research ran continuously for several years, contributing sustained compute time to multiple research projects through World Community Grid. The specific statistics matter less than the fact that the system remained operational, useful, and trusted for an extended period.

The project was formally shut down in 2013 when continued operation no longer made sense within available constraints.

Why it belongs here

Refuse For Research reflects how I approach systems outside formal job roles:

  • Treat side projects as real systems, not toys
  • Design for longevity, not novelty
  • Respect physical and operational limits
  • Plan for failure rather than perfection
  • Shut systems down cleanly when their purpose is complete

Those priorities are consistent across my professional work and other projects documented on this site.

Refuse For Research is no longer active. Historical details, statistics, and archived content remain available on the project site for reference.