3 ms·
A water treatment plant would need about 2 people to a shift (and 4 sets of people) to have 24/7 monitoring (one to watch the control screens, and one to handle
by jcranmer 2y ago
A water treatment plant would need about 2 people to a shift (and 4 sets of people) to have 24/7 monitoring (one to watch the control screens, and one to handle tasks like running tests on water, handling deliveries, etc., that takes you away from the screens), and that basically doesn't change if you're a small facility making 10KGD of water or a large facility making 100MGD of water. There is serious economy of scale going on here.
If you're a small facility servicing a few thousand people, you can't afford to have that kind of monitoring, and so you have to economize in various ways. One of the popular ways is pooling together with other small facilities so that you have one person doing that monitoring for several sites at once, which requires some form of remote operation.
Furthermore, when I worked at a large water company, all of our network, even the telemetry to the various pumping stations dotted around the service area, was on a private network airgapped from the internet. But there's also economy of scale here; a large company servicing 1.5 million people in a large metropolitan area can afford to do custom fiber backhaul in a way that even a bunch of small companies in the rural Midwest cannot, and so the control systems end up being Internet-accessible because it's too expensive for them not to be.
- joe_the_user 2y agoYeah, Small cities are quite willing to economize even if it means X risk some attacker will muck with the system. But the thing I'd take issue is the "can't afford it" part. Of course these cities could afford onsite worker. As gp pointed out, these districts operated long before the Internet and they could provide water then. But the neoliberal paradigm has appeared and suddenly the constant claim is no organization "can afford" not to do any given automation measure, no matter how illogical or dangerous. And so a key thing calling organizations on this baloney.
- elmerfud 2y agoI understand that monitoring and status is reasonable to optimize for cost and this comes with some tradeoffs. Loss of monitoring should not equal a loss if service not unless that monitoring is off line for an extended period. The actual functional control for ensuring a critical service like water is working should remain as an analog computer with something like pneumatics, or other such technology. These are robust and can continue to operate even when electronic circuits have failed. Loss of visibility, and loss of service should be separated. This should be the same even for power stations as well.
- applied_heat 2y agoI only work on power stations, but about 30 of 33 I have worked on are accessible by VPN. They aren’t staffed 24/7 and one crew will be responsible for several facilities in a geographic area. Bad weather causes problems and makes travel difficult or dangerous, so being able to control equipment remotely saves time, money or lost revenue, and is safer by avoiding dangerous travel. The ones not on VPN were owned by big enough utility to have their own private fibre networks. I am ripping out analog electronics from the 70s that nobody alive wants to support and installing PLCs with Ethernet. The attention paid to cybersecurity goes up with the size of the facility as downtime is more expensive. Failsafes preventing damage or loss of life are seperate processors from the plc and independent often redundant systems, but still digital.