A Server for Every Hot Water Tank (TFN)
Theoretical Future News
CES Live Updates
Jan 5 2022
After months of speculation about what Google and Microsoft’s joint announcement would be we are about to get an answer.
…
“Blue Lagoon represents a true win for users, homeowners, and the environment, the cloud is coming home for everyone. For the gamer, who wants the lowest ping while playing Halo Mountain Dew Extreme Edition, she will have thousands of new servers to connect to. Homeowners who choose to install a Blue Lagoon system will save on hot water and free internet for however long they have Blue Lagoon in their home. Best of all Blue Lagoon is the best shade of green, “super green”, the entire system exclusively uses open design standards and connection types and will keep hardware useful for decades to come. For those wondering about power….”
Now back to the present
The basic idea of using servers to heat homes has been around for a while, from crypto miners in Siberia, to Danish home heating companies. The reason for the interest in warming homes with what is often referred to as “waste heat” stems from the cost of keeping server farms cool. On average for every 100 watts of electricity you use to power your servers an additional 40-60 watts of electricity will need to be used to keep everything at the optimal temperature. This desire for cost effective cooling is enough that Microsoft has deployed a test server that is designed to operate on the seafloor, to take advantage of relatively low ocean temperatures.
In 2014 US server farms consumed about 70 billion kilowatt hours (roughly what the state of Kentucky uses). Globally server farms currently represent roughly 2% of electricity demand, with some models indicating that over the next decade 8% of electricity use could go to powering servers. By moving servers into home basement’s we could help to stem the growth in energy demand.
The reason I suggested that Blue Lagoon would be used as a water pre-heater has two major reasons, usefulness and geographic flexibility. Most servers have a target temperature that they want to operate around, much hotter than that target temperature and the server might not last as long, this temperature is lower than what most of us would want coming out of the shower. Most people will set their hot water heater to between 120 and 130°F, but for computers they would prefer to be around 80°F, by using Blue Lagoon as a preheater, you reduce how much additional energy you are using while keeping the electronics happy. Preheating water vs heating the air was a geographic consideration, while there aren’t too many places on Earth where you always want to be warming your home’s air, most people like hot water daily.
Saying the Blue Lagoon would be used for game streaming was inspired by news stories talking about services like Google Stadia, where consumers no longer own their gaming rig, they simply pay to stream the games when they want. I think for lots of people this concept makes sense, but there is a concern with latency, the further you live from a server farm the less enjoyable your gaming experience. By putting a collection of servers in denser population centers you should be able to reduce how long the data spends in transit. While this solution wouldn’t eliminate all forms of input lag, I think it could be a solid win win.
My argument for making Blue Lagoon an open standard piece of hardware is more of my personality rather than an easy economic case right now, but broad strokes, by making it easy to connect additional servers, you would be able to get buy in from other cloud providers. While my limited life experiences only make it feel like game and video streams would make the most sense to be hosted semi-locally, I don’t doubt others with more experience could make a strong case for distributed processing loads being moved onto specialized gear. You might even imagine a marketplace of the future where many buildings have one or more Blue Lagoon base stations installed with various stakeholders buying capacity as needed. The other reason would be for people like me, where I don’t necessarily want to give up on owning my computer (or as much as I am able in this age), but I also don’t need to physically touch the tower nearly as frequently, I would have the option of hiding equipment away as needed, while maintaining access.
The biggest risk to Blue Lagoon would be reliability. Leaky water and power outages are no fun and now imagine that some random person’s flooded basement means that your game saves are lost forever. Obviously that scenario is a bit extreme, but when designing something like Blue Lagoon, the technologies and installation policies used would need to account for these eventualities. Power losses would need to account for handing off data to systems still fully operational. For something like Blue Lagoon to operate a lot of stakeholders would need their concerns accounted for.
Update Nov 22 2019
With all of the chaos associated with Google Stadia’s launch, I think a strong case could be made for moving game streaming server farms into urban centers. Another lower cost alternative to something like Blue Lagoon would be to take a page out of Comcast’s playbook and hide a computer in another product (I mean for Comcast its more that the modem that they rent you also works as an Xfinity brand wireless node). It wouldn’t be that crazy to imagine a swarm of computers that are incentivized to always be left online might also be used to provide more local horsepower. The reason I’m not as big a fan is simply that consumers are less consistently getting a functional benefit from the waste heat.
I hope you found this post interesting, questions, feedback, ideas are always welcome.