[LINK] Broadband powered by ... power

[rchirgwin at ozemail.com.au]

More on this ... I had time to look over the IST-Opera site.

Remembering that this is the 'home' of powerline boosting in Europe, 
it's an interesting read. Because the "reference specification" is more 
than 100 pages long, I just want to highlight a few items from this.

1) Most of Opera is not about powerlines, it's about building 
alternative carrier networks - by which I mean the fibre backbone and 
aggregation networks.

2) When we get to the access network, the reference specification 
tacitly admits that PLC isn't ready for prime time; it assumes that 
today, power utilities will have to use carriers' ULL copper to build 
out their access networks (along with radio).

3) In a triumph of stupidity, Opera is promoting single-vendor 
implementations in various parts of the spec. This says to me that the 
electricity utilities have fallen easy prey to the IT sales pitch. It is 
outstandingly dim.

4) The paper does however make a point which is often overlooked in the 
general public debate: the extremely hostile environment in power 
facilities. Telco switches are very 'hard' by comparison to (say) home 
networking stuff, but not hard enough for a substation, where 
temperatures are too high, there's too much dust, and there's very loud 
electromagnetic noise. The paper proposes hardened cabinets; very good, 
but it escalates the rollout cost.

5) The substation environment changes other things which are overlooked 
in the "powerline saves the world" view. For example, the reference spec 
makes it clear that engineers should never visit the telco racks 
unaccompanied, because power stations are a dangerous environment. 
Hence: wherever a carrier can use one person, a power utility needs two.

Opera refers to an IEEE Spectrum article on Powerline. I took a read; I 
am disappointed that even that august body falls prey to the most basic 
error of predicting outcomes for the "longline" PLC industry based on 
developments for in-home applications. Curing noise problems over the 
household 240V supply and distances of 5 to 10 metres does not mean a 
chip company has 'solved the noise problem'.

RC