It seems to me that this is the beginning of the Spin Cycle around folding Insieme back into the mothership.

Be interesting to see if anyone calls out the Emporer for having no clothes.

Whatever happened to "Simple Cisco"? Wasn't Cisco about to get serious in securing their networking gear market position?

Commodity hardware managed from server software. That's a bit like 1st gen computer networks sort of worked; use generic computers with multiple network ports for handling the networking functions. The markets decided that specialized hardware was more effective for the wirespeed network protocol processing.

It wouldnt seem too difficult for Cisco to make networking hardware that commodity servers cannot compete with performance wise. They do however need to add critical capabilities that commodity hardware, whateverway software defined, cannot support.

Otoh, nothing prevents creating 'commodity hardware' that has the critical capabilities.

And btw, it is not so much SDN, but the pure layer 2 MPLS-TP that is leading to commoditization of service provider switching/routing technologies: one can do MPLS core networking based on 'white box' L2 MPLS switches (LSRs) where the labels are managed from NMS and not via the L3 IP based signaling protocols that are de-facto proprietary to the dominating router vendors; L3 routers (LERs) (from the controlling router vendors) will be needed only at network administrative boundaries. To defend profit margins against this (welcome) wave of commoditization, one needs differentiating capabilities reaching down to the layers below L2, incl of course how the layers function together.

agree with you. SDN is now fashion but let be clear, is there any real and clear $$$ behind that? Till now it is a good idea but at the end of the day an ASIC exist because you need performance and specific actions. SDN it is good and I would be more scare if I am in MPLS modeling SW companies because if I create my own applications to plan the workload of my network elements why I need  routing or MPLS planning application?

SDN's potential really is that heavy.  Actually, my attitude's pretty tame compared with what some of the early SDN predictions were.

I'll admit there are a lot of "ifs" in there, but given the number of ex-Cisco and ex-Juniper people populating SDN startups, it's hard not to think something's going on there.

SDN could redefine what we mean by "networking." Or, it could end up being one small protocol buried in the stack somewhere. I actually think the former is more likely than the latter.

SDN is the new FCoE.  Every vendor wants to talk about it and no one can find a good combined operational/business reason to implement it in it's current format.

SDN is just one topic (hot one !!) of the networking industry.. isnt the article just gives too much weight to SDN on deciding on a large company's future ??