Cisco looks to ‘supercharge’ the internet for the future with improved networking economics

Looking to support the rise of IoT, AI and ML, 28.8T 800G line card is said to be up to six times more space-efficient compared with current offerings

Cisco has upgraded its line of 800G routers, in an aim to transform the economics and sustainability of the Internet for the Future.

One of the consequences in the explosion of the internet of things (IoT), where devices have grown from billions to trillions, has been demand for bandwidth increasing not only from connecting devices with 5G and Wi-Fi, but also from the artificial intelligence/machine learning (AI/ML) workloads required to drive insights from connected devices.

Cisco said that this new age of connectivity has seen applications such as generative AI, search, language processing and recommendation engines drive rapid growth of AI/ML clusters in datacentre environments that require more bandwidth over traditional workloads, adding that AI/ML fabrics need to scale with denser spines that are critical to support the massive number of processors with low latency.

While bandwidth growth seems unlimited, Cisco stressed that space and power are limited, meaning dense and power-efficient platforms are required. As an offering, Cisco is doubling the capacities of communication service provider and Webscale customer backbones, metro core and datacentre networks compared with its 400G/100G modular services.

The 28.8Tbps/36 x 800G line card for Cisco 8000 Series Routers is powered by the Cisco Silicon One processor, and is said to be able to lower operational costs while protecting investments as networks evolve from 100G, 400G and 800G capacities. Customers are also said to be able to benefit from carbon savings by using less hardware to scale and equipment reuse.

“Based on our extensive market research and traffic analysis, we are forecasting continued growth in data traffic with fixed and mobile services, including for 5G, broadband, IoT and cloud,” said Analysys Mason research director Simon Sherrington.

“These trends are putting networks under increasing pressure, which is why scaling to 800G throughput in the future with solutions such as Cisco 8000 will be in demand, while helping service providers and cloud providers improve operational efficiency, sustainability and user experience.”

Read more about network infrastructures

Kevin Wollenweber, senior vice-president and general manager of Cisco networking, datacentre and provider connectivity, said: “We continue to expand 800G to more use cases, from AI/ML fabrics to the core, to help our customers meet their performance and sustainability goals. With our dense core and spine solutions using new double-density line cards with Cisco Silicon One, we have accelerated the transition to 800G anywhere.”

The Cisco 8000 Series Router Systems are also attributed with being able to create up to 83% space savings, allowing users denser networks with much of the same infrastructure to support use cases such as 5G, broadband, IoT and AI/ML. By doubling the capacity in the same chassis footprint, the Cisco 8000 Series Router platform has up to twice the space efficiency of 400G single chassis systems.

Outlining a typical use, Chris Griffin, chief network architect of Florida LambdaRail – a high-speed US computer network owned and operated by the US research and education community – said equipment space was a primary concern as well as computer power as it exhausted its existing footprint in many of its sites.

“With the implementation of Florida LambdaRail’s new FLRnet4 400G backbone, the combination of unbelievable forwarding capacity, operational efficiency and the dependable IOS XR network operating system made the Cisco 8000 Series the obvious choice for our new network,” he said. “We couldn’t be more pleased with our choice. We not only have a state-of-the-art network, but the Cisco 8000 series solution ensures there is enough opex savings to scale our network for years to come.”

Read more on Network hardware

Data Center
Data Management