Awesome! Little servers in every meter; pushing out bits at regular intervals; aggregated at the distribution centers and tied to delivery systems so they can allocate output where it is most needed, or so they can regulate it where it is being abused; and linked to producers and market indexes (although the dream of public-sector resources makes this moot) to inform pricing and quota management ... all open source, open protocol and open mind. Of course, the "wisdom of the crowds" insists that Jerry Springer be in charge of all of that, but what the heck. We gotta start somewhere!
"The smart grid is essentially a nascent energy Internet."
I'd much rather that it were not. Think of all the maladies the Internet has: traffic storms, malware, worms, and the like. Now imagine that in your electric grid controls.
The Internet may seem reliable to most people, but the energy grid needs to be even more reliable. Outages aren't just a nuisance; they can make cities uninhabitable in a matter of hours.
We need to be extra careful which technologies we use. Just because it has worked in the IT realm for five, ten, or even 15 years is no guarantee of suitability for a public utility.
This is high voltage stuff folks (literally). Tread with care.
I agree completely. The internet protocols have plenty enough redundancy to provide drastic improvements in the way we operate the grid. Even if we experience some failures, each and every command/control success will be one that we do not enjoy today. For the most part, our electric grid is completely controlled by a consortium of cheap thermostats located in walls and water heaters and buried within each appliance in millions of homes. It is high time we wrestled control of our power grid away from these stupid devices and use IP to shape our demand to match our resources. I call this "infotricity." You can read more about this concept at http://rbg.glasgow-ky.com/
I don't advocate using the public Internet for the smart grid but the idea of standardizing protocols is to allow many appliances and service providers to interwork - allowing for a lot of competitive offerings. The Internet protocols may prove useful for the grid controls but even if they are, I would expect this energy management network to be segregated. There is an very tricky problem associated with appliances that are both on the power grid and on the public Internet (e.g. so you can program your entertainment systems from the public Internet). We need strong firewalls to prevent crossover from the public Internet into the power grid control systems.
I think protocols like TCP will be almost useless for the energy grid anyway. I'd figure datagrams will be more useful, both unicast and multicast. Some will need to be reliable, some are discardable.
As far as addressing goes, I think something like IPv6 might be ok, but I'm not so sure. I think mostly we are going to need a nice hierarchy of multicast addresses. Maybe bits that indicate the approx. power consumption of the device, the type of device (heating vs entertainment vs cooling vs lighting)... things that can be subscribed to or triggered by various monitoring programs that are interested in specific types of events or situations. E.g., when you are coming back home, start up the cooling and heating devices, but not lighting until you actually get home.
Would that crowds were as wise as Vint Cerf. NIST would be well-served to remember the openness with which Internet standards were and continue to be developed as they approach smart grid standardization.
Awesome! Little servers in every meter; pushing out bits at regular intervals; aggregated at the distribution centers and tied to delivery systems so they can allocate output where it is most needed, or so they can regulate it where it is being abused; and linked to producers and market indexes (although the dream of public-sector resources makes this moot) to inform pricing and quota management ... all open source, open protocol and open mind. Of course, the "wisdom of the crowds" insists that Jerry Springer be in charge of all of that, but what the heck. We gotta start somewhere!
ReplyDelete"The smart grid is essentially a nascent energy Internet."
ReplyDeleteI'd much rather that it were not. Think of all the maladies the Internet has: traffic storms, malware, worms, and the like. Now imagine that in your electric grid controls.
The Internet may seem reliable to most people, but the energy grid needs to be even more reliable. Outages aren't just a nuisance; they can make cities uninhabitable in a matter of hours.
We need to be extra careful which technologies we use. Just because it has worked in the IT realm for five, ten, or even 15 years is no guarantee of suitability for a public utility.
This is high voltage stuff folks (literally). Tread with care.
Jake Brodsky
I agree completely. The internet protocols have plenty enough redundancy to provide drastic improvements in the way we operate the grid. Even if we experience some failures, each and every command/control success will be one that we do not enjoy today. For the most part, our electric grid is completely controlled by a consortium of cheap thermostats located in walls and water heaters and buried within each appliance in millions of homes. It is high time we wrestled control of our power grid away from these stupid devices and use IP to shape our demand to match our resources. I call this "infotricity." You can read more about this concept at http://rbg.glasgow-ky.com/
ReplyDeleteJake,
ReplyDeleteI don't advocate using the public Internet for the smart grid but the idea of standardizing protocols is to allow many appliances and service providers to interwork - allowing for a lot of competitive offerings. The Internet protocols may prove useful for the grid controls but even if they are, I would expect this energy management network to be segregated. There is an very tricky problem associated with appliances that are both on the power grid and on the public Internet (e.g. so you can program your entertainment systems from the public Internet). We need strong firewalls to prevent crossover from the public Internet into the power grid control systems.
vint cerf
I think protocols like TCP will be almost useless for the energy grid anyway. I'd figure datagrams will be more useful, both unicast and multicast. Some will need to be reliable, some are discardable.
ReplyDeleteAs far as addressing goes, I think something like IPv6 might be ok, but I'm not so sure. I think mostly we are going to need a nice hierarchy of multicast addresses. Maybe bits that indicate the approx. power consumption of the device, the type of device (heating vs entertainment vs cooling vs lighting)... things that can be subscribed to or triggered by various monitoring programs that are interested in specific types of events or situations. E.g., when you are coming back home, start up the cooling and heating devices, but not lighting until you actually get home.
Would that crowds were as wise as Vint Cerf. NIST would be well-served to remember the openness with which Internet standards were and continue to be developed as they approach smart grid standardization.
ReplyDelete