Datacenter tech: Here’s where the Open Compute Project is going next | ZDNet

The Open Compute Project (OCP), a computer engineering project launched in 2011 to create better datacenter hardware by sharing designs and ideas between its members, has announced the next stage of its project. 

“Over the last 10 years, significant advancements have been made in open compute standards with the formation of multiple working groups delivering over 350 collaborations,” said Rebecca Weekly, chair of OCP and senior director of hyperscale Strategy and Execution at Intel.

SEE: Google’s new cloud computing tool helps you pick the greenest data centers

Weekly’s stated goals for OCP 2.0 remain largely the same as the original purpose, including modularity, scalability, sustainability, and the ability to integrate across the stack. OCP also plans to “seed” innovation in optics, open silicon, AI, and cooling. 

“While OCP continues to cover all aspects of modular hardware design – compute, storage, switches, accelerators, and racks, there has been growing interest in forward-looking initiatives such as open hardware, chiplets, cooling and software solutions for broad community collaboration, to accelerate innovation and enable scale through ecosystem adoption,” says Weekly. 

OCP set out its seed plans as:

• Optics: Lead in defining market requirements for process and technology transitions for optimal convergence
• Open Silicon: Define interfaces for future co-packaging to enable best-in-class components from the silicon up, and drive tools and reference platform standards
• AI: Scale AI in the marketplace and lead innovation and market adoption of AI/ML by building standardized infrastructure solutions for AI training at scale, AI inferencing, and ubiquity AI
• Cooling: Deliver best-in-class, advanced cooling solutions, sustainable immersion cooling and cold plate designs that support use cases from cloud to edge

The open computing community includes the entire supply chain, from tech and data center equipment vendors, to cloud and communications companies, enterprises, system integrators and semiconductor manufacturers.

SEE: Supercomputers are becoming another cloud service. Here’s what it means

Since it was formed out of a Facebook engineering project, OCP has gradually gained industry momentum. Besides its founding members Facebook, Intel and Rackspace and Goldman Sachs, Google Cloud joined OCP in 2016, and Chinese cloud giant Alibaba joined in 2017.  

Retail giant Target, a one-time AWS customer, joined OCP this year with contributions around edge-computing hardware. Microsoft uses OCP specs in its Azure cloud, while Google contributed a 48V power distribution rack so it can use OCP technology in its data center.