Lichess is incredibly well optimized [0] (and an amazing public service). I'm sure that this is very cost effective for TTT, so a win-win.
[0] https://lichess.org/@/revoof/blog/optimizing-the-tablebase-s...
Lichess is incredibly well optimized [0] (and an amazing public service). I'm sure that this is very cost effective for TTT, so a win-win.
[0] https://lichess.org/@/revoof/blog/optimizing-the-tablebase-s...
Lichess is written in Scala and is hosted on dedicated OVH for a very significantly small amount of money (I think just a few thousand dollars per month) and hosts so many millions of players and games.
It's an understatement how well optimized they are right down to the optimization techniques that they use and the infra providers that they use. The same thing even in something like AWS could cause significantly more amount of money.
It also shows that you don't need AWS/GCP/Azure for basically just about everything, to be honest.
Lichess is a beacon of hope and congrats to the lichess team for this cooperation with TTT.
> It also shows that you don't need AWS/GCP/Azure for basically just about everything, to be honest.
That's where they won, people think AWS/GCP/Azure has to be the default while in reality, the number of platforms that actually need to be able to scale up/down fast are probably below 1% of all platforms out there. Most platforms would save money and run better with proper dedicated hardware rather than going for clouds by default.
Flashback to a moment in my life where a team pushed (successfully) for building a distributed architecture for an app that we didn't even knew if it had product market fit yet. Fast forward 3 years to today and the app is no longer online, but while it had 5 users they were using really reliable infrastructure, I guess that's cool.
It's the same principle that makes so many people unreasonably sympathetic to the concerns of those much wealthier than themselves: The statistically implausible assumption that they, too, will soon be part of the 1% somehow. Better start acting the part early!
Cloud is more cost effective the less of it you have because it doesn’t cost 3x more to maintain a kubernetes cluster with thrice the nodes, but it does cost 3x more to rent one. This is even more true for serverless.
I can imagine a lot of small apps buy into serverless at a time where it’s legitimately the most cost-effective solution and then they’re stuck because serverless platforms are easy to lock yourself into.
Depending on the development cost of being in the cloud, it was probably the right choice. Optimizing for cost per user when you don't have product market fit is probably the worse early optimization.
> That's where they won, people think AWS/GCP/Azure has to be the default while in reality, the number of platforms that actually need to be able to scale up/down fast are probably below 1% of all platforms out there. Most platforms would save money and run better with proper dedicated hardware rather than going for clouds by default.
This. I kind of wish if more people knew about it. Also even 1% can be too big. I mean Lichess is literally having millions of people if not more, It's definitely within the 0.001% group.
I kind of wish to do something in this space in the future, I do feel like its just that people don't know about it. I have been thinking to approach some companies and just tell them how much they can save if they migrate and use open source solutions and these dedi servers and setting things on top of these dedi's/vps's.
I have been thinking of (within future), to contact a few companies and to actually have them save net money from migration while charging them a few hundred bucks a month and I can just have a very handful selection of companies (say 15-20) to have enough money so that I can eat french fries and manage their servers!
It feels a win-win-win situation for everyone except AWS/GCP/Azure who wish to suggest that scalability is hard etc. and this false premise for most if not essentially* all businesses.
Personally, I am also saying things like slack for example, I don't understand why people might want slack when things like matrix exist and can be self-hosted securely with proper 3-2-1 backups and for most intents and purposes is actually good if not better than slack.
To me, a bit of concern though with this and I am not sure if it is well-founded is what if I set these servers for them, now I will wish to set them up so that they have as little errors as possible but what if the companies start to think that I am doing nothing and then they stop paying my contracts after I have set them up on these dedicated. I guess I hope that they believe in the value of human support and I guess I am also a bit unsure of where do I find such businesses are.
My brother does some freelancing on the side and I ask him these things and he mentions that mostly he has to use AWS, I mention why not dedi and he says that he does what he is asked to do and that company wants him to use AWS so he uses AWS, so I guess within this context, I need companies who are atleast interested in being a bit more open about thinking about dedicated servers.
I am sure that there would be companies interested in all of this and I am interested in doing things for them but I am not sure about the middle part of connecting the two. I would be genuinely interested to hear your thoughts on all of this and have a nice day emsh!
It also seems to have significantly better availability than chess.com, where I regularly had games end abruptly and be completely removed from my game history due to what I can only assume is a server restart.