Scalable and Efficient Solutions with OCP

GIGABTYE is an active member of the OCP, regularly attending the OCP's annual summits and continuously designing and releasing new server nodes, node trays, and server racks based on the OCP's Open Rack Standard specifications as well as providing the best performing add-on cards for your OCP solution.

The Future of Open Source Ecosystem

The Open Compute Project (OCP) is a collaborative community focused on redesigning hardware technology to efficiently support the growing demands on compute infrastructure. In 2011, the OCP Foundation was initiated with a mission to apply the benefits of open source and open collaboration to hardware and rapidly increase the pace of innovation. Its collaboration model is now being widely applied in fields like data centers, telecom industry, and edge infrastructure.

GIGABYTE One Stop OCP Solutions

GIGABYTE aims to provide the most diverse product lines to fit all use cases. With that in mind, GIGABYTE offers OCP solutions, including computational components such as add-on cards, compute nodes, and storage nodes, and further expands the product lineup to cover trays and racks. With all necessary parts made by a single company, GIGABYTE ensures that product performance is optimized and provides a complete rack solution for easy adoption of the OCP ecosystem.

Add-on Card

JBOD

GPU Server

Compute Node

Node Tray

Power Shelf

OCP Rack

Taking another step forward, GIGABYTE has also introduced an immersion cooling total solution. Similar to its air cooling and direct liquid cooling counterparts, the single-phase immersion cooling solution covers everything, from add-on cards to the immersion tank. This next-generation cooling solution is perfect for companies interested in adopting more environmentally friendly and cost-effective alternatives while wanting to avoid the complicated and time-consuming verification process.

Immersion Ready Server

Immersion Tank

Why the Open Compute Project

By openly sharing ideas, specifications, and other intellectual properties, the OCP aims to design, use, and enable mainstream delivery of the most efficient designs for scalable computing. It maximizes innovation and reduces complexity in tech components. The OCP provides a structure in which individuals and organizations can share their intellectual properties with others and encourage the IT industry to evolve, throwing off the shackles of proprietary and redefining tech infrastructure by designing commodity hardware that is more efficient, flexible, and scalable.

What is ORv3 and Why Adopting it

ORv3 stands for Open Rack v3, a specification within the broader OCP concept that focuses on rack design and power supply regulation, as its name suggests. Unlike the more common EIA 19” server standard, the larger 21” specification allows for a significant increase in airflow. This is particularly valuable as server components generate more and more heat. The enhanced airflow is achieved with the same cabinet external width, ensuring compatibility with already existing infrastructure. And this is just one of the reasons for adopting ORv3.

The advantages of ORv3 can generally be divided into the following three categories:

Enhanced Airflow

Keeping the same cabinet width, the wider and taller chassis design allows better airflow, accommodating the increasing heat from server components.

Improved Power Efficiency

The unified power shelf design simplifies the maintenance and reduces power consumption by utilizing PSU at the capacity with better efficiency.

Easy Management

I/Os are regulated to face towards the front and be hot-pluggable. Bringing convenience to users and improving overall system manageability.

OCP Improves Flexibility and Scalability in Data Centers

ML & AI

Training algorithms for AI applications requires a large amount of parallel processing power, delivered by dense GPU deployments. A large number of high-powered GPUs leads to massive energy consumption and produces a lot of heat. With the focus on thermal efficiency and scalability, GIGABYTE OCP servers provide the best solutions for large scale ML and AI adoption in data centers.

HPC

High performance computing (HPC) relies on high compute performance from CPU and/or accelerators to process large volumes of data or run various simulations. For the amount of data, they too need high speed or high-capacity storage paired with fast networking for clusters. GIGABYTE OCP servers, along with GIGABYTE OCP add-on cards line-up, provide various options to customers looking for configurations for specific applications.

Cloud Computing

As computing transitions toward the cloud, more and more requirements start to be emphasized in cloud infrastructures, other than a stable and powerful system. The ability to quickly process and transmit data, the ease of managing massive-scale systems, and the right storage architecture and networking capabilities are all essential. GIGABYTE OCP servers can handle all of these requirements while ensuring scalability for future scale-up and scale-out to meet the exponentially increasing need in cloud computing.