rhr has a point -- a lot of the commoditization is already underway. Although I have to consider that Cisco's switching margins bounced back after last year's dive. I'd thought of that dive as a permanent shift, and Cisco proved me wrong (or at least squirmed to safety for a few more quarters).

Anyway, chips_ahoy: The vendors will give you long dissertations about why OpenFlow isn't a threat to their hardware franchises because it makes the network flexible rather than commoditized, and because Layer 3 is still really hard to do. And in fact, some of the OpenFlow proponents say they don't expect the technology to commoditize switches and routers completely.

But yeah, it's hard not to see that possibility. "People won't want to program the network" is one argument that gets used, but I don't know... is learning CLI/IOS easier than (hypotheticalness alert) buying software that automates provisioning and comes with on-site customization? I'm thinking not.

> The service provider community should make Openflow mandatory in their requirements so that OEMs will take it seriously. It's revolutionary and cost effective

I'm not sure the success of OpenFlow is that assured... but given that practically every vendor except Cisco has at least pledged interoperability with OpenFlow, if not outright support (actually some people in Cisco *have* said they'll support OpenFlow in some places)... it wouldn't be a stretch.

I don't see how this is anything good for incumbent system vendors... if there is no hardware differentiation, the product is by definition a commodity.  Lower top line... much lower bottom line and massive off shoring of jobs... seems awful.

I would love to hear a detailed answer about why I'm wrong here.