.63W: N100 server with 2 HDD & 2 SSD. Cable modem. 5-port POE switch. WAP. Ooma VOIP device.

The POE & WAP are like 10W between them, but I had to add them to get strong enough wifi to one particular client. Authentik somehow consumes 3W. Immich also has a high idle load, so I leave it down most of the time.

Yeah, I started with a PCI (no e) card, but had to switch to USB when it got hard to find cheap motherboards with PCI slots. It’s an old setup :) Honestly amazed that they can fit the whole thing into a thumb-sized USB dongle, although I suppose it’s easier without the analog side.

I use a USB tuner like https://hauppauge.com/pages/products/data_dualhd.html and https://mythtv.org/ Has to be plugged into an external antenna, and it really helps for that antenna to be on the 2nd floor, in a window, with clear view unobstructed by aluminum siding.

tvheadend.org or HDhomerun are probably more general solutions.

You might think that ordering cases of canned tomatoes, or a 10-year supply of rubber gloves are poor management decisions, but that’s because this AI is playing seven dimensional chess against your tic-tac-toe. Just wait until it’s cornered the tomato market, and then you’ll see.

I added homeassistant and some power monitors to my stack, and the IT rack comes in around 1.5 kWh/day - one of the biggest power budgets in the house, even with a low-power CPU, after adding in a few HDDs, a couple switches, and the cable modem. I’m also in a cheap power state, so it’s not a financial pressure, just surprising how quickly 10W here, 10W there…add up. At $0.50/kWh, I’d think solar would be a no-brainer.

Keep power in mind. For most home-use services, you don’t really need much computing power, and you might be able to do all you want with a single box. Even 30W, 24/7 is $25 (@10¢/kWh)-125(@50¢)/year of electricity. That said, it’s a small price to learn how to do clustering or swarms.

I’d guess that your biggest load would be transcoding in Jellyfin, for which Intel Gen 6 added h265 to quicksync. The Gen 3/4 CPUs in M73 would be extra slow with most modern codecs.

My setup is a pile of kludges built on top of each other over the last two decades.

I started with ULAs distributed through DHCP, connected to named, which allows hosts do declare their own name and let me access local services as though I had a real domain.

My ISP eventually started supporting IPV6, but only assigned /128, so the ULAs got NAT-6ed out to the real world.

I eventually learned how to request prefix delegation from the ISP and set up SLAAC.

So now, my PIv6 clients have a) their link-local address, b) the ULA, c) a “privacy” SLAAC, and d) a unique SLAAC. All my internal services still refer to the ULAs.

I don’t think I’d recommend this system for someone setting up from scratch. The easiest thing would be to go with SLAAC, if you can get prefix delegation, and set your DNS/pihole to send the unique-SLAAC address of any servers you run.

Maybe they do commercial customers different, but I’m about 30 miles north of the site in question, and my water use is reported in real time. I can even get a daily report from their web site. It’s hard to believe they’d be less interested in the usage of their 1e6-gallon-per-year commercial customers than their 1e4-gallon-per-year residential customers.

Water company can measure the water that leaves their pumping station(s) - just put a flow meter on the one big pipe. If that doesn’t match the sum of all their customer meters, then water is going somewhere else - broken pipe, illegal connection, meter fraud, whatever.

I would guess that most jurisdictions already have that one big flow meter, because they have to comply with water rights agreements, have to know how much chlorine & fluoride to inject, etc.

Kind of fascinating that they don’t do any kind of reconciliation of water delivered against water billed. You’d think that would be an easy thing to do and a good way to discover leaks (or theft). I mean, there would definitely be ‘missing’ water due to leaks, fire department, etc, but one imagines that would have some kind of normal/tolerable range, and that 30 million missing gallons would trigger some kind of investigation prior to customer complaints.