Our long-standing research partnership is now a strategic opportunity to advance open source AI, and everyone is invited.

If you’ve followed Red Hat Research over the years, chances are you’ve heard of the Mass Open Cloud Alliance (MOC-A). The MOC is a production cloud service with 150 petabytes of storage, supporting containers through Red Hat OpenShift, an AI platform through Red Hat OpenShift AI, and bare metal access through Elastic Shared Infrastructure (ESI). When the MOC launched in 2014, it aimed to support research and education with a transparent alternative to opaque public commercial clouds. That goal was a perfect fit for Red Hat Research’s initial focus: working with university partners to innovate in the hybrid cloud space and keep those innovations open.

We have an unprecedented opportunity to bring new AI technologies from academic groups to the wider research community, and to grow and improve AI implementations at scale with real workloads.

The success of this relationship means we now have an unprecedented opportunity to bring new AI technologies from academic groups to the wider research community, and to grow and improve AI implementations at scale with real workloads. The MOC has become a strategic platform for Red Hat to partner with academic researchers and evolve open source software with research and educational users. Red Hat Research is working with the MOC-A to further develop this platform so researchers and engineers can improve open source AI tools and solutions, ensure that dominant AI technologies are open source, and build an ecosystem of partners and collaborators providing AI solutions and services. In addition, the MOC-A can enable access to AI infrastructure to researchers and new research startup projects that could not otherwise access open transparent AI implementations. 

What’s in it for you?

Great question, and we have a lot of answers, depending on where you operate in the technology landscape. To name a few:

Domain researchers

Domain researchers can leverage infrastructure to use in ways that won’t happen with traditional public clouds. Red Hat and the MOC-A empower domain specialists to run AI workloads, support facilitation with the technology, and work with system researchers and industry partners to rapidly evolve the platform. Domain researchers also benefit from the experience and engineering collaboration of Red Hat’s Office of the CTO. These experts bring the flexibility and transparency of the whole open source ecosystem to bear on projects that range from designing satellites to detecting disease by analyzing medical images. Red Hat’s new machine learning tools— InstructLab and Red Hat OpenShift AI—are also conveniently accessible to all MOC users.

Technology vendors

Technology vendors have an opportunity to reach influential audiences. Hardware providers can partner to expose new technology to a wide community of users. For example, Lenovo partnered with the MOC-A to provide 64 NVIDIA A100 GPUs under a new business model of GPU core-hour lease, and is now expanding this partnership with 192 H100 GPUs. Software developers can benefit both from building relationships with collaborating universities and driving adoption and by working with MOC-A resources, like telemetry, to optimize their solutions based on real usage.

The MOC-A is an active member of the AI Alliance, supporting forums and working groups and providing technical leadership to some of the AI Alliance’s most important initiatives.

All MOC users benefit from connections to the broader AI and cloud communities. Red Hat Research engineers are actively involved in community projects where they contribute their software to the community and also bring the latest new ideas back to implement in MOC-A joint science, engineering, and start-up incubation projects. For example, the MOC-A is an active member of the AI Alliance, supporting forums and working groups and providing technical leadership to some of the AI Alliance’s most important initiatives. The MOC-A also works with multiple nonprofit organizations and government agencies to make the best of high-tech available to everyone, one project at a time.

What’s happening now?

To galvanize our expanded focus on the MOC, in June 2024 Orran Krieger, Professor of Electrical and Computer Engineering, Boston University, and Mass Open Cloud PI, joined Red Hat Research while on sabbatical to help focus the research team—not just in Boston, but globally—on the goal of making the MOC an open, distributed platform for AI/ML workloads. With Heidi Dempsey, US Director for Red Hat Research, we are engaging industry and university partners, open source communities, and Red Hat business units to collaborate in making the MOC the platform of choice for open source AI development. 

We’ve already begun to exploit the combined capabilities of the MOC and Red Hat OpenShift AI. One powerful example is the AI for Drug Discovery Forum, held in October to launch a new AI Alliance working group focused on drug discovery using new open source AI foundation models (Heidi is an interim co-lead for the working group). The forum also gave users hands-on experience in running inference and using notebooks to submit various types of protein combinations to the models. A joint effort of Boston University, IBM, Red Hat, the MOC-A, the Cleveland Clinic, and the AI Alliance, the AI for Drug Discovery Forum demonstrates the power of interdisciplinary collaboration and open source AI models to break through research bottlenecks while ensuring that solutions are ethical, accessible, and transparent. An essential step in building this platform is building the ecosystem of contributors, users, advocates, collaborators and partners who will benefit from it. If you are interested in partnering with Red Hat around the MOC-A, or would like more information, please contact Heidi Dempsey (hdempsey@redhat.com) or Orran Krieger (okrieger@redhat.com).