Palworld server costs near $500K per month as network engineer is ordered to 'never let the service go down no matter what'
Palworld server costs near $500K per month as network engineer is ordered to 'never let the service go down no matter what'
"Wait, maybe we'll go bankrupt from server fees?" joked the Pocketpair CEO.
Fast, cheap, reliable. You can have any two you want.
this is a server basterdization of "Good, Fast, Cheap" regarding producing just about anything I'm guessing, which tends to hold true in the real world quite well, yes?
As an engineer yeah, but honestly it’s usually pick one to prioritize, one to strive for, and one to ignore.
We can get it out fast, and it can be not bad but pretty expensive or it can be pretty cheap but not good. If we get it good we can try to do it cheaply and take our time, or we can try to do it quickly and it’ll be expensive.
That works for some contexts, but no amount of time can get you both total reliability and low costs, so in this case it's pick one.
In this context “fast” refers to speed of the system, not time to implement.
I'll take fast twice.
Double fast, yeah 😎
On spec, on time, on budget. Failure to meet those goals is a result of piss poor planning.
Those are all the same attributes, just the planned out version of it where the balance of speed, reliability and cost are decided upon ahead of time.
Well 19m players x $29 is $551,000,000 banked so far.
They could pocket a few dozen million and still run the servers for around 85 years.
Don't forget the cutshare
29 = (8.7 to Valve) (20.3 Pocket)
7m are on Xbox, so the count is:
Pocket = 243.6 m (on 12m copies sold)
Valve = 104.4 m ( on 12m copies sold)
It’s on Gamepass for Xbox, unless those numbers are non-gamepass copies.
You’re forgetting the fact that all stores take a fee, and many users are paying a pittance to play through game pass, which can cost as low as $1.
They still made quite a lot, but not $29 per user.
I was just making a loose estimate, there's a lot I haven't figured in but I'll bet you with the sales in the next month they'll surpass my numbers.
How much do gamepass copies pay?
I have no idea what the over-under is between the creators and MS. my estimate isn't exact, it's close but relies on increased sales moving forward.
I also haven't figured in production costs or debts.
If the game ever stops, people might realize they're playing Palworld.
*East India Pokemon Company
The thing is, I don't need to be online.
I bet most people are playing single player.
Apart from the people doing multiplayer 10-20%?) everyone else could just be offline.
This is for them.
But it also proves that if a company gives a shit, they can do it. This can be achieved with lower costs and experience, so in time the costs will come down.
Whereas Activision blizzard don't give a fuck and anytime there's a new DlC or game there's significant downtime despite being a multi billion dollar company. Why people continue to support them I'll never know
It's even weirder because I'd expect even those playing with friends to be doing so in their locally hosted servers with at most 4 friends I think? The people playing on the official servers are such a minority that I can't fathom this cost being worth it.
The people online don’t need it though, they just need a place where they can enter an IP
thats a fuckton of server space, i didnt think playing on random official servers with no admins or good anti cheats would be that popular
Or they're super inefficient.
Running a passworded Palworld server on Linux. Have about 7-10 active players on it and the server instance balloons up to ~33GB of RAM usage in less than 12 hours of uptime.
Supposedly disabling some features (like base raids) reduces resource utilization, but was curious what stock settings would do.
When it was restricted to 10GB on a container it would just crash every couple of hours, running out of resources.
It is, RAM usage is absolutely wild on it and it needs constant reboots.
$500k/mo isn't really even all that much in cloud costs. I did some work for a large company and just the PoC/development account for our project alone was $100k/mo.
Hey my boss tells me the same and I barely make six figures wtf
Imo they should:
- Ask money for a subscription to go online on official servers (after they ironed out a good anti cheat)
- Keep the self hosted / dedicated servers as a free alternative
500k a month to support 19 million person play base doesn't seem totally unreasonable. They've already made £400m+ in early access in the first month - so it's a drop on the ocean at the moment.
Costs will probably come down - at the moment they've been scrambling to keep up with demand which means expensive rapid deployment rather than long term server build out.
And presumably they plan to get the game out of early access so potentially get more players (although may not get many more players in this case as it's so popular) and more importantly start rolling out DLC content to make more money.
I doubt they need to go the subscription route plus may be too late as they launched without it.
Also, the player base will be a fraction of what it is today in a month. They're dealing with unprecedented demand that's gonna fall off into something more reasonable by throwing money at it.
It's the right thing for them to do. It would have been stupid to plan for this much demand. You'd delay the game by another year just building out a cloud native architecture. Letting the servers buckle would have killed momentum.
I bet they’re either already on or will migrate to AWS
But then they'd be paying $2M a month! /s
Tbh they probably already are, $500k/month is a lot of money. They would be able to get those costs down by hiring a few it engineers and renting a few racks at a CoLo. Geographic distribution is hard for a company of their size, though, and maybe it's not worth making that investment if the game's popularity isn't going to last.
Yeah the up front cost of the type of infrastructure necessary to handle the player volume they have is not only expensive, but requires a ton of expertise to be done correctly, AND requires lots individual geographically discrete locations to keep latency down.
The fear for them would be investing in all that infrastructure just for the game to fall off in popularity after a few years.
No one in their right mind would deploy own servers for this kind of load. It fluctuate way too much and in half a year you have unused servers that are junk. Initial purchase price would be millions, and setup would take months.
They are definately running in some cloud, and 500k/month is about what you would expect to host servers for a popular game like this in close to launch.
I don't particularly like the either or approach. You can certainly spin up some minimum on local hardware. You have some up front capex but something that doesn't have a fluctuating, expensive monthly opex bill.
You can then use cloud architecture to add capacity resources on demand and in different geographic locations. You can also utilize multiple cloud architectures to further add redundancy and cost optimization.
If you build out the scripts used to dynamically scale to also pull current pricing, you can have something that is both heavily redundant and somewhat cost effective. Sure it's not like azure, AWS, Google cloud, or any other public cloud option changes their pricing that frequently, but it would give a good way to compare specifically in different regions.
For a game like this, building capacity and the ability to scale early was clearly more important than optimizations in the server code base. 500k/mo isn't actually a lot to companies and it's likely to go down as optimizations are implemented and popularity stabilizes.
500k a month? Nope.