Cisco Blogs

Nexus- Defined

January 28, 2008 - 2 Comments

Nexus: The Core, or Center. After spending 12% of my life working with the finest engineers, marketing professionals, and product management today we are announcing the Cisco Nexus Family. Led by the flagship Cisco Nexus 7000 Series– the first Data Center Class Switching Platform.What do we mean by data center class? Start with carrier class, then build up. Make a system that is designed to operate inline with a customer’s operational best practices. Make a system that works the way you do.Why Nexus? A nexus is a connecting point, or the core depending which line you read in the above URL. To us at Cisco Nexus is the foundation for building virtualized data centers, the foundation for unifying the data center fabric, the convergence point of SAN, LAN, and High Performance Computing. I’ll probably be writing a lot about the Cisco Nexus 7000 over the next several weeks – we’ll discuss the new OS – NX-OS and how it enables data center virtualization, we’ll chat about unified fabrics and what makes them the single most important innovation in Ethernet since switching was pioneered 15 years ago, and we’ll delve into Cisco TrustSec: scalable security that simplifies operations. For now click through to and root around a little. dg

In an effort to keep conversations fresh, Cisco Blogs closes comments after 60 days. Please visit the Cisco Blogs hub page for the latest content.


  1. JT- interesting idea. While we are not ruling out any direction like this one area that will challenge us is the sheer performance of the system.i.e. with several hundred watts per slot of cooling capacity it is hard to put enough memory and CPU and custom silicon in there to process the full workload of an application at a speed powerful enough to take advantage of enough of the 230Gb per slot in the Nexus to make it worthwhile. Same argument holds true on firewalls, App Delivery, etc. As silicon technologies mature , denisities increase, voltages lower, certainly it may be possible to get 40Gb+ performance for one of these applications per slot. Until then it is more cost effective to use either a Catalyst 6500 with services modules or to use external ASA and ACE appliances.I do think it may be possible in the future to split some of the control plane and data plane functionality between them though. Let the switch do the data plane and the appliance the control plane. dg

  2. Since the Data Center 3.0 ‘press release’ at Networkers last summer, we have presumed something like this was in the works. On the surface this looks like a new chassis to meet the future ’10Gbps and beyond’ Ethernet requirements. With the virtualization and modular software aspects, though, there are obviously a lot of possibilities here. A vendor or two in the marketplace today has built chassis that incorporate the actual application into the box – taking the actual ‘virtualization’ and ‘the network as the platform’ ideas one step further. Can you comment on your thoughts regarding this? With a VMware relationship, could we expect to see actual application modules in this box soon?