Back in December of 2019, we made a bold announcement about how we’ll forever change the economics of the internet and drive innovation at a speed like no one had ever seen before. Since then, we’ve been very busy. In October of 2020, we announced six new devices, and in March of 2021, we announced the addition of three new devices, including our first 25.6 Tbps web scale switching device.
Today, less than seven months from our last product rollout, I’m thrilled to introduce our latest member of the Cisco Silicon One family, the P100, which brings us to a total of 11 devices in less than two years, proving once again that our pace of innovation is unmatched in the industry. Having a fully unified and scalable architecture allows us to continue building new silicon at an unmatched pace, covering routing and web scale switching with purpose-built devices.
Cisco Silicon One Family
Cisco Silicon One P100
With our Q200 device, we already had the leading routing silicon in the market, but with the introduction of the Cisco Silicon One P100 we increase the lead significantly with the industry’s first 112G Serializer/Deserializer (SerDes) high bandwidth routing silicon, achieving an unheard of full-duplex 19.2 Tbps.
Leveraging our second-generation P4 run-to-completion engine, coupled with large table scale and deep buffers, the Cisco Silicon One P100 is the ultimate routing silicon for high bandwidth applications. Like the G100, our P100 device enables a true single flow 1.6 Tbps port allowing for optimal link utilization for high bandwidth point-to-point connections.
Until now, modular line cards were limited by optical faceplate bandwidth, fabric card connector density constraints, and low bandwidth silicon. The Cisco Silicon One P100 breaks through all these barriers by enabling the industry’s first 28.8 Tbps, 36 port QSFP-DD800 line card, an impressive two times increase over any other solution in the market.
With Cisco Silicon One’s unique sliced-based architecture, every port can be a true ethernet port. This can be used to create a single rack unit 24 port QSFP-DD800 system or a two-rack unit 48 port QSFP-DD56 system. Using other routing silicon, it requires at least six pieces of silicon which radically changes the cost, power, latency, and size of the system. And, more importantly, we’re able to achieve all this in 7nm, once again showing the efficiency of the Cisco Silicon One architecture.
The Cisco Silicon One P100 is sampling to our customers now.
Expansion of G100
When we launched the G100 in March of 2021 we focused on its capability as a standard ethernet switch. Today we’re pleased to announce that with new P4 software, the G100 can now operate as the industry’s highest bandwidth, fully scheduled, non-blocking fabric element at 25.6 Tbps.
Using the G100 as the fabric element and the P100 as the line-card element, it’s possible to build the highest bandwidth single chassis router, achieving more than two times the bandwidth of any other solution.
The P100 and G100 can also be used to build large, disaggregated chassis with G100 in the spine and P100 as the leaf, thereby achieving a multi-petabit router. Combing the highest bandwidth fabric element and the highest bandwidth routing silicon means customers can create an astonishing 2 Pbps router in a single-stage Clos fabric, offering more than three times higher bandwidth than any other solution on the market.
Today Cisco Silicon One continues to increase its lead in the market with the industry’s highest bandwidth routing silicon. The bandwidth, programmability, scale, and features of the P100 are a testament to the efficiency of the Cisco Silicon One architecture.
But most excitingly, we continue to see the effects of the journey we started seven years ago. The pace of innovation is speeding up, not slowing down, because we’ve got the right investment, the right technology, and the right team in place to make it happen.
Stay tuned for new developments!
Learn more about
architecture, devices, and benefits.