Re: It's Monday morning, let's hope BT are already analysing the code on Git-Hub.
Adam:
I am not the person you're replying to, but as a former Cisco employee: I do know how much those switches cost - and I can tell you that you're still incorrect in your thinking.
Aside from the fact that telcos won't be using Catalyst switches for their FTTP network, the cost of telco gear is a rounding error in the overall budget for a nationwide network - and it's an unavoidable expense. As someone else stated, the real cost is in paying the burly blokes to stuff fibre into the ground. You can't virtualise or automate that.
You still need a piece of hardware to actually interface with the (shared) fibre and talk to the ONTs (FTTP equivalent of a cable or DSL modem). That piece of hardware is called an OLT (equivalent to a CMTS or DSLAM), and is already pretty dumb as it is, and is already configured through software in a sense. The actual broadband termination, IP routing and heavy lifting is done elsewhere in the network (which already exists, as it already serves today's FTTH customers along with everyone else on VDSL, and would be a separate task to consider virtualising).
Of course in the home you still need that ONT (or the ONT function integrated into another device), so that's an unavoidable cost too.
The only difference is that in this virtualised utopia (which is more likely to come out of various projects that big telcos and ETSI are working on, not this Facebook thing) you'll replace vendor specific element managers (like Cisco Prime) with something more open and flexible. Much of the same hardware stays in place - you still need physical interfaces, and ASICs are still better than software at terabit scale routing. Netconf and YANG might replace SNMP and SSH, but it's not a huge step change. I fully expect these new whizzy virtual routers to run crusty old BGP and the IGP of choice too.
Telcos are risk averse beasts, after all.