IP Infusion Elevates AI Data Center Networking with Cutting-Edge Ethernet-Based Disaggregated Solutions
In a groundbreaking move to address the growing demands of artificial intelligence (AI) and machine learning (ML) workloads, IP Infusion, a global leader in open networking solutions, has announced the release of OcNOS® Data Center (OcNOS-DC) 6.6.1. This latest version of its disaggregated networking software is purpose-built to support the rigorous requirements of AI-driven data centers, offering unmatched performance, scalability, and cost efficiency. Designed to operate on Broadcom’s Tomahawk 5 chipset, OcNOS-DC 6.6.1 powers high-performance Ethernet-based AI/ML fabrics on ONIE-enabled white-box switches such as the Edgecore AS9817-64D and UfiSpace S9321-64E, boasting an impressive switching capacity of 51.2Tbps.
As AI and ML training clusters continue to scale exponentially, traditional networking solutions are increasingly unable to meet the unique demands of these workloads. The intensive east-west traffic patterns, combined with stringent latency and packet loss requirements, pose significant challenges for distributed GPU workloads. OcNOS-DC 6.6.1 addresses these pain points with advanced features designed to deliver high-throughput, lossless connectivity. These include Priority-based Flow Control (PFC) over Layer 3, Enhanced Transmission Selection (ETS), Data Center Bridging Capabilities Exchange Protocol (DCBX), and Dynamic Load Balancing (DLB). Together, these capabilities ensure that AI-driven networks can handle the immense computational demands of modern AI applications while maintaining optimal performance.
One of the standout features of OcNOS-DC 6.6.1 is its ability to enhance AI workload orchestration through real-time telemetry. By exporting gNMI telemetry to orchestrators like Kubernetes, the software enables dynamic allocation and reallocation of jobs when network paths become overloaded. This ensures efficient utilization of resources and minimizes downtime, even during peak workloads. Additionally, integration with Ansible automates critical tasks such as reserving lossless queues, mapping priority-to-traffic-class configurations, and activating dynamic load balancing on job-specific links. These automation capabilities not only streamline operations but also optimize the performance of GPU server racks, ensuring seamless execution of AI/ML tasks.
The disaggregated nature of OcNOS-DC 6.6.1 offers significant advantages in terms of cost efficiency and flexibility. By decoupling hardware and software, organizations can select best-of-breed components, enabling competitive bidding and component-level price tracking. This approach allows for agile data center buildouts and provides the freedom to choose optics tailored to specific needs, further enhancing cost savings without compromising performance.
“OcNOS Data Center sets a new standard for AI/ML networking, delivering unparalleled performance and scalability for next-generation data centers,” said Kiyo Oishi, CEO of IP Infusion. “By combining Broadcom’s Tomahawk 5 chipset with our advanced software features, we’re enabling customers like HYPER SCALERS to build agile, cost-efficient networks that meet the rigorous demands of GPU-intensive AI workloads.”
HYPER SCALERS, a leading provider of high-performance GPU-based services, has already embraced OcNOS-DC 6.6.1, leveraging its capabilities to revolutionize their infrastructure. “Partnering with IP Infusion to deploy OcNOS-DC 6.6.1 has transformed our ability to deliver exceptional GPU-based services,” said George Cvetanovski, Founder and CEO of HYPER SCALERS. “The AI-optimized features, seamless orchestration integration, and disaggregated architecture allow us to scale efficiently while maintaining the low-latency, lossless connectivity our customers require for their AI/ML workloads.”
The partnership between IP Infusion and HYPER SCALERS exemplifies how cutting-edge technology can be harnessed to meet the evolving needs of AI-driven enterprises. As AI and ML applications become increasingly integral to industries ranging from healthcare to finance, the ability to process massive datasets quickly and accurately is paramount. OcNOS-DC 6.6.1 empowers organizations to achieve this by providing a robust, scalable, and cost-effective networking solution that supports the most demanding AI workloads.
Moreover, the disaggregated model championed by IP Infusion represents a paradigm shift in data center networking. Traditional proprietary solutions often come with high costs and limited flexibility, making it challenging for organizations to adapt to rapidly changing requirements. In contrast, OcNOS-DC 6.6.1’s open networking approach allows businesses to future-proof their infrastructure, ensuring they can scale and evolve alongside advancements in AI and ML technologies.
The impact of OcNOS-DC 6.6.1 extends beyond just improving performance and reducing costs. By enabling efficient resource utilization and minimizing latency, the software plays a crucial role in accelerating AI innovation. Researchers and developers can focus on pushing the boundaries of what’s possible without being constrained by network limitations. This, in turn, drives progress across various fields, from autonomous vehicles to drug discovery.
As the demand for AI and ML continues to surge, the need for specialized networking solutions will only grow. IP Infusion’s OcNOS-DC 6.6.1 positions itself at the forefront of this trend, offering a comprehensive solution that addresses the unique challenges of AI-driven data centers. With its advanced features, seamless integration capabilities, and commitment to cost efficiency, OcNOS-DC 6.6.1 is poised to become the go-to choice for organizations looking to build agile, high-performance networks for their GPU-intensive workloads.
About IP Infusion
IP Infusion develops open network software solutions for carriers, service providers, and data centers. With hundreds of customers and thousands of deployments, IP Infusion leads the market in Network Operating Systems. Its flagship platform, OcNOS®, empowers network operators to disaggregate their networks, streamline operations, and reduce total cost of ownership (TCO). Headquartered in Santa Clara, Calif., IP Infusion is a wholly owned subsidiary of ACCESS CO., LTD.
