In Depth Analysis of Network Energy Efficiency, but what about ‘Always-On’ ?

I have been neglectful my topic of the intersection between Information and Communication Technologies (ICTs) and sustainability. Let me try and redress that balance a little with a couple of posts that I have been meaning to write.

I was recently (thanks Lynn !) sent a link to an in depth review by Tim Hills in Light Reading Europe Recent Moves in Green Telecom. We often focus on data centers because of the concentrated nature of their energy consumption, while overlooking networks because they are so much more dispersed in their energy consumption patterns. This paper gives a pretty comprehensive review of the various industry bodies and their related initiatives on improving the energy efficiency of the network and of the customer premise equipment (CPE) that hangs off it. Most of the initiatives reviewed focus on improving the efficiency of the network architecture we now have and it is great to see how much is going on in this space. Worth a look for people involved in networks and CPE.

The only thing I thought missing was explicitly drawing out the ‘always-on’ issue we currently face. That is, as I can best describe it, that the devices we attach to the network tend to be ‘always-on’ because the IP network was designed to work that way. The network doesn’t like to have to do too many of the signaling handshakes required when we power up a piece of equipment. And we (the user) don’t like the delay when we power something up and wait for it to connect, so we leave things ‘always-on’. But we don’t need ‘always on’, we need ‘always available’ (just like old telephones used to be – no current drawn unless the bell was ringing or the handset lifted). Future generations of networks and associated standards need to try and address this and be designed in a way that no current need be drawn by the CPE device when it is idle.