Cisco has unveiled the 8223-routing system, designed to efficiently and securely connect data centres and support the next generation of AI workloads.
As AI adoption increases, data centres are experiencing higher demand, greater power constraints, and evolving security challenges. The Cisco 8223 is a 51.2-terabit-per-second (Tbps) Ethernet fixed router designed to manage the heavy data traffic associated with AI workloads between data centres.
The Silicon One P200 chip powers the 8223. These technologies aim to improve network performance and prepare infrastructure for large-scale AI operations.
“AI compute is outgrowing the capacity of even the largest data centre, driving the need for reliable, secure connection of data centres hundreds of miles apart,” says Martin Lund, EVP for Cisco common hardware group.
“With the Cisco 8223, powered by the new Cisco Silicon One P200, we’re delivering the massive bandwidth, scale and security needed for distributed data centre architectures.”
In many of the world’s data centres, AI workloads are stretching power and space limitations. Hyperscalers can’t scale-up (add more capacity into each individual system) or scale-out (connect multiple systems within a data centre) any further.
This dynamic puts increasing demand on data centre interconnects, as the industry must scale-across by distributing AI workloads across multiple data centres. Without addressing the connection points between data centre, organisations could face performance challenges, capacity bottlenecks and suboptimal processing that wastes time, power and money.
The Cisco 8223 system aims to provide organisations the flexibility and programmability necessary to build these networks, with deep buffering to provide the cross-site security and reliability necessary for crucial workloads.
Power Efficient. Scalable. Programmable. Secure.
Cisco says it is essential to build a future-ready backbone AI network that can meet the power consumption challenges while remaining scalable, flexible and secure. The Cisco 8223 is designed to provide customers the capacity to handle surging workloads, the necessary flexibility with full programmability, and the power efficiency to directly address power consumption challenges.
Cisco says the 8223 is:
- Power efficiency: The 8223 is designed for scale-across networking and uses a deep-buffer routing architecture optimised for fixed deployments. It aims to reduce energy consumption while managing the high power demands of AI workloads. The 3RU system is built to maximise space efficiency, a factor expected to become increasingly important as AI clusters expand.
- Scalability: The 8223 supports 64 ports of 800G, providing high bandwidth and routing density for fixed systems. It can process more than 20-billion packets per second and can scale up to 3 exabits per second. With 800G coherent optics support, it can be used for data centre interconnect and metro applications up to 1,000 km. The P200’s deep-buffering capabilities help manage traffic surges during AI training to maintain network performance.
- Adaptability: The 8223 is built to respond to real-time network conditions. Powered by the programmable P200 silicon, it can support emerging network protocols and standards without requiring hardware replacement, allowing networks to remain flexible and efficient as AI traffic patterns evolve.
- Security: The 8223 includes hardware and software-based protection, with line-rate encryption using post-quantum resilient algorithms, integrated safeguards, and continuous monitoring tools. It connects with Cisco’s observability platforms to provide detailed network insights, helping maintain data security and reliability.
Flexibility
Cisco says the demands of AI networking are not only increasing but also continuously evolving. According to the company, organisations require infrastructure that can adapt as business needs change and offer flexibility in deployment to suit different network designs. The 8223 will first be available for open-source SONiC deployments, with IOS XR support planned for future releases.
The P200 silicon, which powers the 8223, will be available for use in modular platforms and disaggregated chassis, providing consistent architecture across networks of various sizes. Cisco’s Nexus portfolio is expected to include systems running NX-OS based on the P200 in the near future, extending flexibility and operational adaptability for AI-focused infrastructure.
Cisco Silicon One
Cisco Silicon One is a unified portfolio of networking devices across AI, hyperscaler, data centre, enterprise and service provider use cases. Introduced in 2019, Cisco Silicon One is playing critical roles in major networks around the world.
“The increasing scale of the cloud and AI requires faster networks with more buffering to absorb bursts,” says Dave Maltz, technical fellow and corporate VP for Azure Networking at Microsoft.
“We’re pleased to see the P200 providing innovation and more options in this space. Microsoft was an early adopter of Silicon One, and the common ASIC architecture has made it easier for us to expand from our initial use cases to multiple roles in DC, WAN, and AI/ML environments.”
Dennis Cai, Alibaba Cloud VP and head of network infrastructure, says: “As Alibaba continues to invest in and expand the cloud infrastructure, DCI is a critical pillar of our strategy. We are pleased to see the launch of Cisco Silicon One P200, the industry’s first 51.2T routing ASIC that delivers high bandwidth, lower power consumption, and full P4 programmability.
“This breakthrough chip aligns perfectly with the evolution of Alibaba’s eCore architecture. We plan to leverage the P200 to build a single chip platform, serving as a foundational building block for expanding our eCore deployment. Beyond supporting our Cisco Silicon One Q200 deployment scenarios, this new routing chip will enable us to extend into the Core network, replacing traditional chassis-based routers with a cluster of P200-powered devices.
“This transition will significantly enhance the stability, reliability, and scalability of our DCI network while keeping the simplicity. In addition, we are developing and exploring innovative disaggregated architectures using Cisco G200 for our high-performance datacenter network. The introduction of this advanced routing chip marks a pivotal step forward, empowering Alibaba to accelerate innovation and drive infrastructure expansion in the AI era.”
Dave Ward, Lumen CTO and product officer, says: “As a long-standing Cisco customer, Lumen is actively advancing our network infrastructure to support the AI-driven economy. Cisco’s 8000 Series, Cisco Silicon One and Cisco’s pluggable optic technology represent key innovations that align with our goals for scalable, efficient multi-cloud connectivity. As we look to the future, we are exploring how the new Cisco 8223 technology may fit into our plans to enhance network performance and roll out superior services to our customers.”
Patrick Moorhead, CEO and chief analyst for Moor Insights and Strategy, says: “As AI workloads rapidly outpace the capabilities of traditional data centers, the industry faces new challenges in bandwidth, reliability, and scale. The migration of data centers to remote locations for power access makes ultra-reliable, high-bandwidth interconnects essential. Cisco’s 8223, powered by Silicon One P200, marks a significant step forward, delivering the industry’s first 51.2-terabit fixed Ethernet router purpose-built for secure, power efficient scale-across networking.”