The Chambers Legacy: Cisco and the Integrated Virtual Networking Stack
Arthur Cole examines the impact retiring Cisco CEO John Chambers has had on the networking world.
It isn’t overkill to call it the end of an era when a powerful CEO like John Chambers makes his exit. But before we start talking about the legacy he built during his time at Cisco, it is important to understand that legacies are not based on what happened in the past, but how that past plays out in the future.
Indeed, it might seem rather odd for someone like Chambers to step down just as the networking component of IT infrastructure is about to make the leap to virtualization and, by extension, the cloud. But since development of advanced platforms always precedes deployment – sometimes by several years – it may very well be that the technological aspects of the Chambers legacy are already in place.
But that still leaves the execution, and on that front it seems the Cisco strategy is far from certain. The lynchpin in the company’s software defined networking (SDN) strategy is the Application Centric Infrastructure, which offers an integrated hardware/software stack that purportedly provides a higher level of service than all-software approaches like VMware’s NSX. A key attribute of ACI is the ability to enable applications to provision and secure their own networking environment on the fly, and Chambers has long argued that this is most easily done through close coordination between software management systems and underlying hardware.
So far, there doesn’t seem to be any real complaints about the ACI platform, although to be fair it has seen only a limited number of deployments so far, with most of those in test/dev rather than actual production workloads. Still, initial feedback from key early adopters suggests that expectations, at least, are that the ACI platform and the Nexus 9000 switch on which it is based will not only provide a better switching environment, but one that is radically different from the static architectures of the past. Fortunately, those who have boarded the ACI train realize that change of this magnitude comes with a fairly steep learning curve and that this should be easier on an integrated platform than if a decoupled software stack needs to be deployed on disparate white box infrastructure.
And yet, in an example of Cisco hedging its bets somewhat, the company has made some overtures to the pure-software proponents of SDN. Earlier this year, Cisco announced that the Nexus 9000 would support BGP EVPN (Border Gateway Protocol Ethernet Virtual Private Networking), an open protocol for implementing virtual overlays that makes it possible for third-party SDN controllers to manage ACI environments. This is a big step for Cisco, because it essentially allows their proprietary flagship platform to fall under the auspices of another SDN vendor, possibly tarnishing the Cisco brand, if things don’t work out well.
No one vendor can control the cloud, however, and as data architectures become distributed both geographically and across multi-vendor platforms, those who keep to themselves will be at a distinct disadvantage in the emerging data ecosystem.
This could turn out to be Chambers biggest legacy of all: the realization that Cisco not only has an integral role to play in advancing enterprise network infrastructure into the cloud, but that once SDN is in place, it has to play nice with others in order to provide the most flexible, dynamic network architecture for its customers.
The future, of course, belongs to whoever provides the highest value at the lowest price point. There was a time when a company called Microsoft outmaneuvered a company called Apple by decoupling hardware from software on the PC. Cisco’s future depends very much on history not repeating itself in the cloud.
Photo courtesy of Shutterstock.
Arthur Cole covers networking and the data center for IT Business Edge. He has served as editor of numerous publications covering everything from audio/video production and distribution, multimedia and the Internet to video gaming.