Technology

Cisco Nexus Hyperfabric delivers speed and simplicity in Cisco's latest AI infrastructure build

As customer zero for Cisco Nexus Hyperfabric, Cisco IT gained the ability to build AI infrastructure with radical speed and simplicity. 

The ask from the business:

Cisco IT set out to bolster employee productivity with its internal AI Assistant, but it needed more in-house infrastructure, and quickly.

Challenge

Cisco's existing AI infrastructure took as long as six months to design, order, and deploy. It was vital that we be able to do it more quickly, overcoming:

  • Complex ecosystem: AI infrastructure forms a connected and complex landscape, which made it extremely difficult to design, build, and scale.
  • Delays that led to waste: Every delay in installation led to wasted hardware lifespan, time, and money. The acceleration of AI innovation only exacerbated this waste.
  • Limitations of network management solution: No one solution existed that could provide comprehensive data visibility through networking and compute, so engineers spent countless hours compiling data, manually planning, and investigating issues.
  • Talent and skill gap: Building AI networks requires deep knowledge of advanced networking technologies. Finding many engineers with this level of expertise was challenging.

Solution

The team became the first real-world user of Cisco Nexus Hyperfabric, using it to quickly order and deploy a full stack of Cisco Data Center Networking solutions, including:

  • Cisco 6000 Series Switches: Cloud-managed secure switches that deliver simplicity and high performance, built for Cisco Nexus Hyperfabric.
  • Cisco UCS Servers: Bring together compute, networking, and storage, all in a single system to power our AI applications. 
  • Cisco Silicon One: This architecture for our Cisco switches delivers unmatched routing and switching capabilities for AI and networking needs. 
  • Cisco Optics: Fast, reliable optical transceivers deliver resilience across our complex infrastructure.

Outcomes

Dramatically increased speed to deploy

Standing up the AI fabric took just six days with Nexus Hyperfabric, whereas previous builds took as long as six months. This dramatic acceleration maximized time-to-value on our infrastructure and allowed us to power our applications with urgency.

Simplicity in deployment and management

AI integration and full stack data visibility simplify the design, ordering, deployment, and management processes. This makes it easier for engineers of any expertise level to build and manage AI networks.

More for you

Technology has changed how we live and work

Leading organizations are innovating with Cisco solutions to connect, secure, and transform.

Organizations like yours rely on Cisco

Each industry has its own challenges that require tailored solutions.

You are a changemaker, innovator, and discoverer

We want to help you share your story. Learn more about how you can build your organization's profile—and your own—as you expand your network.


Let’s power an inclusive future for all

Demo Nexus Hyperfabric

Explore Nexus Hyperfabric to see how quick and easy it can be to design your own AI infrastructure.  

Find a partner

Solve your business challenges by finding a Cisco partner authorized to design, sell, and support custom solutions.