Cisco’s "Architectures": A Seismic Shift?
It's the first attempt by a big networking vendor to frame a new vision of the network and to engage new benefits to drive spending.
Cisco has long been known as a sales-driven company, one whose marketing activities were more focused on buffing the brand than on educating the customer or even developing the market. At their "Live" event this month, Cisco gave a completely different kind of pitch, one that emphasized "architecture". Of course, this might be just a PR play, but there are some deeper indications that it might be real, and the "why" is pretty interesting to industry watchers.
Markets generally cycle between the adoption of new stuff that creates new value propositions, and the orderly replacement of aging gear associated with a mature technology. Often at the end of the replacement cycle, buyers are eager to adopt the new because they're eager to explore new technology options. Sellers sometimes lag, fearing that the new paradigm won't favor their products or value their current incumbency. Sellers also often fear the task of market education--which is needed to drive a new wave of adoption--even though it might eventually make them more money. They have to get comfortable.
That's what I think is behind Cisco's "architecture" move. If you don't change how technology is used, you don't change the benefit case. If you don't change the benefit case, spending is static and sellers just compete each other into lower-margin deals. Cisco sees a chance to break out, a chance created by some aggressive moves it's made since it had its famous bad quarter and promised to structure itself for the future.
An "architecture" is a framework in which technology is deployed and cooperates to create benefits. It defines the scope of business impact and so defines the total benefits available to justify the cost. It defines how products integrate to create cooperative technical ecosystems and so it lays out the projects for buyers and identifies the skills they'll need.
Arguably, the market has been groping for an architecture, talking about "the cloud" or "SDN" but never really getting a vision of what either would mean. Cisco seems to think that's the case because they talked very specifically about both at their event, and in fact linked them loosely together.
If you distill the sense of the Cisco talks, it's that the cloud defines the future of IT and so cloud networking defines the future of networking. Cisco's vision of cloud networking is that it's software-defined because the cloud is software-defined. It's orchestrated, in that the management of IT resources demands coordinated management of resource connectivity, and so does the task of providing users access to applications that might be hosted anywhere, on anything.
VPNs are a kind of foundation for the Cisco vision, because current-generation VPNs can be mapped to "virtual private clouds" for public, private, and hybrid applications, and because next-generation software-defined VPNs are Cisco's vision of Software Defined Networking. To cement VPNs into the cloud, Cisco is offering a cloud-hostable software router, which lets you host a VPN component in the cloud, and "Cloud Connectors," which provide vertical APIs for application control (i.e., they "connect" the application to the cloud). All of this serves to build cloud/application awareness into the network. The VPN linkage means that Cisco's visions map to current WAN technologies and are compatible with currently installed devices. This is an architecture, but an evolutionary one rather than a revolutionary one.
That point is the crux of the debate over Cisco's position. Proponents of a revolutionary vision of the cloud and SDN will argue that not only has Cisco dumbed the whole process down by linking it to current protocols and products, they've walked away from the emerging standard OpenFlow, poster child for SDN implementation so far.
Supporters of Cisco will say that first-off, nobody is going to revolutionize networks that still retain a trillion bucks or so of residual value. Second, Cisco and supporters note that OpenFlow is an experimental protocol, one that allows for applications to control networks but doesn't define how the applications know what services they want or how to induce the devices under control to provide them.
So which view will prevail? Flip a coin, at least until more market experience with Cisco's vision accumulates.
Another question that may be answered later is how this whole Cisco architecture thing will relate to UC/UCC, telepresence, etc. It seems pretty clear that to the extent that UC/UCC applications are cloud-hosted, they fit into the vision like any other cloud application, meaning that UC/UCC could control network behavior via the "Cloud Connector" API. For voice calls that's of little value, but for videoconferencing it could be big. If so, then the Cisco move could potentially move other players to look at software-defining their UC networks.
The competitive response to Cisco's moves is probably important overall and not just for UC. Nobody has really pushed cloud networking as a meaningful architecture, nor has anyone really looked at the whole top-to-bottom software-defined networking picture. Will competitors snipe at Cisco's seemingly self-serving evolutionary position? If they do, and if Cisco is right that an architecture is needed to help buyers take the next big step in IT evolution, then competitors may be sniping their way out of the market.
For the industry we're all a part of, Cisco's is an important move because it's the first attempt by a big networking vendor to frame a new vision of the network and to engage new benefits to drive spending. My model says that over the next eight years, the industry could gain over $180 billion in new sales by building a better benefit case for buyers. Even a little of that could go a long way in tough times!