I dislike those black and white takes a lot. It's absolutely true that most startups that just run an EC2 instance will save a lot of cash going to Hetzner, Linode, Digital Ocean or whatever. I do host at Hetzner myself and so do a lot of my clients.
That being said, the cloud does have a lot of advantages:
- You're getting a lot of services readily available. Need offsite backups? A few clicks. Managed database? A few clicks. Multiple AZs? Available in seconds.
- You're not paying up-front costs (vs. investing hundreds of dollars for buying server hardware) and everything is available right now [0]
- Peak-heavy loads can be a lot cheaper. Mostly irrelevant for you average compute load, but things are quite different if you need to train an LLM
- Many services are already certified according to all kinds of standards, which can be very useful depending on your customers
Also, engineering time and time in general can be expensive. If you are a solo entrepreneur or a slow growth company, you have a lot of engineering time for basically free. But in a quick growth or prototyping phase, not to speak of venture funding, things can be quite different. Buying engineering time for >150€/hour can quickly offset a lot of saving [1].
Does this apply to most companies? No. Obviously not. But the cloud is not too expensive - you're paying for stuff you don't need. That's an entirely different kind of error.
[0] Compared to the rack hosting setup described in the post. Hetzner, Linode, etc. do provide multiple AZs with dedicated servers.
[1] Just to be fair, debugging cloud errors can be time consuming, too, and experienced AWS engineers will not be cheaper. But an RDS instance with solid backups-equivalent will usually not amortize quickly, if you need to pay someone to set it up.
> But the cloud is not too expensive - you're paying for stuff you don't need. That's an entirely different kind of error.
Agreed.
These sort of takedowns usually point to a gap in the author's experience. Which is totally fine! Missing knowledge is an opportunity. But it's not a good look when the opportunity is used for ragebait, hustlr.
I'm fully aware this is pedantic, but you can't save 10x. You can pay 1/10. You can save 90%. Your previous costs could have been 10x your current costs. But 10x is more by definition, not less. You can't save it.
I'd be more interested to understand (from folk who were there) what the conditions were that made AWS et al such a runaway hit. What did folks gain, and have those conditions meaningfully changed in some way that makes it less of a slam dunk?
My recollection from working at a tech company in the early 2010s is that renting rack space and building servers was expensive and time consuming, estimating what the right hardware configuration would be for your business was tricky, and scaling different services independently was impossible. also having multi regional redundancy was rare (remember when squarespace was manually carrying buckets of petrol for generators up many flights of stairs to keeps servers online post sandy?[1]).
AWS fixed much of that. But maybe things have changed in ways that meaningfully changes the calculus?
You're falling into the false dichotomy that always comes up with these topics: as if the choice is between the cloud and renting rack space while applying your own thermal paste on the CPUs.
In reality, for most people, renting dedicated servers is the goldilocks solution (not colocation with your own hardware).
You get an incredible amount of power for a very reasonable price, but you don't need to drive to a datacenter to swap out a faulty PSU, the on site engineers take care of that for you.
I ordered an extra server today from Hetzner. It was available 90 seconds afterwards. Using their installer I had Ubuntu 24.04 LTS up and running, and with some Ansible playbooks to finish configuration, all in all from the moment of ordering to fully operational was about 10 minutes tops. If I no longer need the server I just cancel it, the billing is per hour these days.
Bang for the buck is unmatched, and none of the endless layers of cloud abstraction getting in the way. A fixed price, predictable, unlimited bandwidth, blazing fast performance. Just you and the server, as it's meant to be.
I find it a blissful way to work.
AWS also made huge inroads in big companies because engineering teams could run their own account off of their budget and didn’t have to go through to IT to requisition servers, which was often red tape hell. In my experience it was just as much about internal politics as the technical benefits.
> although actually many people on here are American so I guess for you aws is legally a person...
Corporate legal personhood is actually older than Christianity, and it being applied to businesses (which were late to the game of being allowed to be corporations) is still significantly older than the US (starting with the British East India Company), not a unique quirk of American law.
"To them, it’s way too convenient to be on AWS: not only it solves their problem, but it’s also a shiny object. It’s technically complex, it makes them look smart in front of other devs, "
why? why be so obnoxious to other people who you claim are being obnoxious to you. no need to read your blog post now/
The author touches on it briefly, but I'd argue that the cloud is immensely helpful for building (and tearing down) an MVP or proving an early market for a new company using startup credits or free tiers offered by all vendors. Once a business model has been proven, individual components and the underlying infrastructure can be moved out of the cloud as soon as cost becomes a concern.
This means that teams must make an up-front architectural decision to develop apps in a server-agnostic manner, and developers must stay disciplined to keep components portable from day one, but you can get a lot of mileage out of free credits without burning dollars on any infrastructure. The biggest challenge becomes finding the time to perform these migrations among other competing priorities, such as new feature development, especially if you're growing fast.
Our startup is mostly built on Google Cloud, but I don't think our sales rep is very happy with how little we spend or that we're unwilling to "commit" to spending. The ability to move off of the cloud, or even just to another cloud, provides a lot of leverage in the negotiating seat.
Cloud vendors can also lead to an easier risk/SLA conversation for downstream customers. Depending on your business, enterprise users like to see SLAs and data privacy laws respected around the globe, and cloud providers make it easy to say "not my problem" if things are structured correctly.
Seems like nowadays people seem less concerned with vendor lockin than they were 15 years ago. One of the reason to want to avoid lockin is to be able to move when the price gouging gets just a little bit too greedy that the move is worth the cost. One of the drawbacks of all these built in services at AWS is the expense of trying to recreate the architecture elsewhere.
Its not actually that hard to get your own server racked up in a data centre, I have done it. Since it was only one box that I built and installed at home I just shipped it and they racked it in the shared area and plugged the power and network in and gave me the IP address. It was cheaper than renting from someone like hetzner, was about £15 a month at the time for 1A and 5TB a month of traffic at 1gbps. Also had a one off install fee of £75.
At the time I did this no one had good gaming CPUs in the cloud, they are still a bit rare really especially in VPS offerings and I was hosting a gaming community and server. So I made a pair of machines in a 1U with dual machines in there and had a public and private server with raid 1 drives on both and redundant power. Ran that for a gaming server for many years until it was obsolete. It wasn't difficult and I think the machine was about £1200 in all, which for 2 computers running game servers wasn't too terrible.
I didn't do this because it was necessarily cheaper, I did it because I couldn't find a cloud server to rent with a high clockspeed CPU in it. I tested numerous cloud providers, sent emails asking for specs and after months of chasing it down I didn't feel like I had much choice. Turned out to be quite easy and over the years it saved a fortune.
Cheap shot maybe, but the fact that the page takes 10 seconds to load when it hits the HN front page is a great, inadvertant illustration of why you might want to use the cloud sometimes.
The failure mode of self-hosting is that your site gets hugged to death, the failure mode of the cloud is that you lose a ton of money. For a blog that doesn't earn you anything, the choice is clear.
Besides, you can just put it behind cloudflare for free.
That happens a lot to blogs deployed on the cloud too. They just need to put a small cache in front and they'll be able to serve one or two orders of magnitude more requests per second.
The first couple of paragraphs of price comparisons are useful. Then there are many paragraphs of sheer waffle. The author doesn't even seem able to define what "the cloud" is:
> The whole debate of “is this still the cloud or not” is nonsense to me. You’re just getting lost in naming conventions. VPS, bare metal, on-prem, colo, who cares what you call it. You need to put your servers somewhere. Sure, have a computer running in your mom’s basement if that makes you feel like you’re exiting the cloud more, I’ll have mine in a datacenter and both will be happy.
I read the whole thing and I didn't see any waffle. Sure, undeniably some excess word count, some emotion in responding to critics. But no waffle.
The "is this cloud or not" debate in the piece makes perfect sense. Who cares whether Hetzner is defined as "the cloud" or not? The point is, he left AWS without going to Azure or some other obvious cloud vendor. He took a step towards more hands on management. And he saved a ton of money.
The cheap hosting service they switched to is arguably "cloud".
If you can't drive to the location where your stuff is running, and then enter the building blindfolded, yet put your hands on the correct machine, then it's cloud.
I would really be interested in an actual comparison, where e.g. someone compares the full TCO of a mysql server with backup, hot standby in another data center and admin costs.
On AWS an Aurora RDS is not cheap. But I don't have to spend time or money on an admin.
Is the cost justified? Because that's what cloud is. Not even talking about the level of compliance I get from having every layer encrypted when my hosted box is just a screwdriver away from data getting out the old school way.
When I'm small enough or big enough, self managed makes sense and probably is cheaper. But when getting the right people with enough redundancy and knowledge is getting the expensive part...
But actually - I've never seen this in any if these arguments so far. Probably because actual time required to manage a db server is really unpredictable.
> Probably because actual time required to manage a db server is really unpredictable.
This, and also startups are quite heterogeneous. If you have an engineer on your team with experience in hosting their own servers (or at least a homelab-person), setting up that service with sufficient resiliency for your average startup will be done within one relaxed afternoon. If your team consists of designers and engineers who hardly ever used a command line, setting up a shaky version of the same thing will cost you days - and so will any issue that comes up.
Its a skillset that is out of favour at the moment as well but having someone who has done serverops and devops and can develop as well is a bit of a money saver generally because they open up possibilities that don't exist otherwise. I think its a skillset that no one really hired for past about 2010 when cloud was mostly taking off and got replaced with cloud engineers or pure devops or ops people but there used to be people with this mixed skillset in most teams.
If you’re going to write a post about why self-hosting is better than cloud*, then it’s probably a good idea to make sure your site loads in under a minute.
* at least I assume what this post is; I’m still waiting for it to load.
I've been at too many startups with a devops team that would rather provision 15 machines with 4GB RAM THAN ONE WITH 64GB.
I once got into an argument with a lead architect about it and it's really easy to twist the conversation into "don't you think we'll reach that scale?" To justify complexity.
The bottom line is for better or worse, the cloud and micro services are keeping a lot of jobs relevant and there's no benefit in convincing people otherwise
I have a VPS. It costs me £1.34 per month. It's way over-powered for what I need it for.
However, one situation where I think the cloud might be useful is for archive storage. I did a comparison between AWS Glacier Deep Storage and local many-hard-drive boxes, for storing PB-scale backups, and AWS just squeaked in as slightly cheaper, but only because you only pay for the amount you use, whereas if you buy a box then you have to pay for the unused space. And it's off-site, which is a resilience advantage. And the defrosting/downloading charge was acceptable at effectively 2.5 months worth of storage. However, at smaller scales you would probably win with a small NAS, and at larger scales you'd be able to set up a tape library and fairly comprehensively beat AWS for price.
It's an informative post but I really dislike the language and style that are becoming common in this kind of posts, e.g.:
> Look, first of all, you’re as unique as the other 1000 peanut gallery enjoyers that have made the same astute observation before you. Congratulations. But you’re absolutely missing the point.
> Most people complaining about what I did happen to have “devops”, “cloud engineer”, “serverless guy”, “AWS certified”, or something similar in their bio.
“It is difficult to get a man to understand something, when his salary depends on his not understanding it.”
Jeez, this was a painful read. I actually stopped after a few paragraphs and asked AI to make it more technically focused and remove the ranting so I could stomach it.
Strawman arguments, ad hominem attacks and Spongebob mocking memes, and the casual venturing into conspiracy theories and malicious intentions...
> Why do all these people care if I save more money or not? ... If they’re wrong, and if I and more people like me manage to convince enough people that they’re wrong, they may be out of a job soon.
I have a feeling AWS is doing fine without him. Cloud is one of the fastest growing areas in tech because their product solves a need for certain people. There is no larger conspiracy to keep cloud in business by silencing dissent on Twitter.
> You will hear a bunch of crap from people that have literally never tried the alternative. People with no real hands-on experience managing servers for their own projects for any sustained period of time.
This is more of a rant than a thoughtful technical article. I don't know what I was expecting, because I clicked on the title knowing it was clickbait, so shame on me, I guess...
>This happens with Hetzner all the time because they have no VLANs and all customers are on a single LAN and IPFS tries to discover other nodes in the same LAN by default.
I think a lot of teams using cloud are using SaaS rather than IaaS. They want a redis and a postgres and a S3 and a ... You can set all that up on a server, but it's not very fun if you've never done it before.
Sorry but my $3 AWS instance is still cheaper than all of those options.
If you need a lot of, well, anything, be it compute, memory, storage, bandwidth etc., of course cloud stuff is going to be more expensive... but if you don't need that, then IMO $3/mo on-demand pricing really can't be beat when I don't have to maintain any equipment myself. Oracle also offers perpetually free VM instances if you don't mind the glow.
With a quick LLM-assisted search, looks like the cheapest EC2 instance is t4g.micro, which comes in at $2.04/mo. It has 2 vCPUs and and only 512MiB of RAM. (I assume that doesn't include disk; EBS will be extra.)
I can certainly see a use for that small amount of compute & RAM, but it's not clear that your level of needs is common. I've been paying for a $16/mo VPS (not on AWS) for about 15 years. It started out at $9/mo, but I've upgraded it since then as my needs have grown. It's not super beefy with 2 vCPUs, 5GiB of RAM, and 60GiB of disk space (with free data ingress/egress), but it does the job, even if I could probably find it cheaper elsewhere.
But not at Amazon. Closest match is probably a t3.medium, with 2 vCPUs and 4GiB RAM. Add a 60GiB gp2 EBS volume, and it costs around $35/mo, and that's not including data transfer.
The point that you're missing is we're not looking for the cheapest thing ever, we're looking for the cheapest thing that meets requirements. For many (most?) applications, you're going to overpay (sometimes by orders of magnitude) for AWS.
You say "if you need a lot", but "lot" is doing a bit of work there. My needs are super modest, certainly not "a lot", and AWS is by far not the cheapest option.
Just get a raspberry pi and run it from your own home internet. You should already be paying for a VPN service and your regular internet service, so you should be able to trivially work out a self-hosted solution. You'll recover your costs inside of two years and come out the other end better off for it.
Don't give the big cloud companies an inch if you don't absolutely have to. The internet needs and deserves the participation of independent people putting up their own services and systems.
Amazon really doesn't care if your $10,000 bed folds up on you like a sandwich and cooks you when AWS us-east-1 goes down, or stops your smart toilet from flushing, or sets bucket defaults that allow trivial public access to information you assume to be secure, because nobody in their right mind would just leave things wide open.
Each and every instance of someone doing something independently takes money and control away from big corporations that don't deserve it, and it makes your life better. You could run pihole and a slew of other useful utilities on your self-hosted server that benefit anyone connected to your network.
AI can trivially walk you through building your own self-hosted setups (or even set things up for you if you entrust it with an automation MCP.)
Oracle and AWS and Alphabet and the rest shouldn't profit from eating the internet - the whole world becomes a better place every time you deny them your participation in the endless enshittification of everything.
yet another obsessive take on "cloud is bad and expensive" eh? I think they vastly forget the value of some SaaS offerings in terms of time saving for small companies. running and managing numerous DBs, k8s clusters, ci/cd pipelines and stateless container systems is simply impossible with a team of 1-2 people. sure if the setup is simple and only requires a few classic components, this is way cheaper and for a 99.9% SLA will work fine. otherwise it only makes sense if you had very large cloud bills and can dedicate multiple engineers to the newly created tasks.
I think we've gone a little nuts defining "production system" these days. I've worked for companies with zero-downtime deployments and quite a lot of redundancy for high availability, and for some applications it's definitely worthwhile.
But I think for many (most?) businesses, one nine is just fine. That's perfectly doable by one person, even if you want, say, >=96% uptime, which allows for 350 hours of downtime per year. Even two nines allows for ~88 hours of downtime per year, and one person could manage that without much trouble.
Most businesses aren't global. Downtime outside regular business hours for your timezone (and perhaps one or two zones to the west and east of you) is usually not much of a problem, especially if you're running a small B2B service.
For a small business that runs on 1-3 servers (probably very common!), keeping a hot spare for each server (or perhaps a single server that runs all services in a lower-supported-traffic mode) can be a simple way to keep your uptime high without having to spend too much time or money. And people don't have to completely opt out of the cloud; there are affordable options for e.g. managed RDBMS hosting that can make maintenance and incident response significantly easier and might be a good choice, depending on your needs.
(Source: I'm building a small one-person business that is going to work this way, and I've been doing my research and gaming it out.)
The post is about that 99% of companies that will never go large scale. Its point is that they don't need cloud, buying a server or two is all they need.
I dislike those black and white takes a lot. It's absolutely true that most startups that just run an EC2 instance will save a lot of cash going to Hetzner, Linode, Digital Ocean or whatever. I do host at Hetzner myself and so do a lot of my clients.
That being said, the cloud does have a lot of advantages:
- You're getting a lot of services readily available. Need offsite backups? A few clicks. Managed database? A few clicks. Multiple AZs? Available in seconds.
- You're not paying up-front costs (vs. investing hundreds of dollars for buying server hardware) and everything is available right now [0]
- Peak-heavy loads can be a lot cheaper. Mostly irrelevant for you average compute load, but things are quite different if you need to train an LLM
- Many services are already certified according to all kinds of standards, which can be very useful depending on your customers
Also, engineering time and time in general can be expensive. If you are a solo entrepreneur or a slow growth company, you have a lot of engineering time for basically free. But in a quick growth or prototyping phase, not to speak of venture funding, things can be quite different. Buying engineering time for >150€/hour can quickly offset a lot of saving [1].
Does this apply to most companies? No. Obviously not. But the cloud is not too expensive - you're paying for stuff you don't need. That's an entirely different kind of error.
[0] Compared to the rack hosting setup described in the post. Hetzner, Linode, etc. do provide multiple AZs with dedicated servers.
[1] Just to be fair, debugging cloud errors can be time consuming, too, and experienced AWS engineers will not be cheaper. But an RDS instance with solid backups-equivalent will usually not amortize quickly, if you need to pay someone to set it up.
> But the cloud is not too expensive - you're paying for stuff you don't need. That's an entirely different kind of error.
Agreed. These sort of takedowns usually point to a gap in the author's experience. Which is totally fine! Missing knowledge is an opportunity. But it's not a good look when the opportunity is used for ragebait, hustlr.
linode was better and had cheaper pricing before being bought by akamai
Whoa, an acquisition made things worse for everyone but the people who cashed out? Crazy, who could have seen that coming
I'm fully aware this is pedantic, but you can't save 10x. You can pay 1/10. You can save 90%. Your previous costs could have been 10x your current costs. But 10x is more by definition, not less. You can't save it.
+1 words matter
Clarity of expression is a superpower
I don’t feel it’s pedantic at all.
I'd be more interested to understand (from folk who were there) what the conditions were that made AWS et al such a runaway hit. What did folks gain, and have those conditions meaningfully changed in some way that makes it less of a slam dunk?
My recollection from working at a tech company in the early 2010s is that renting rack space and building servers was expensive and time consuming, estimating what the right hardware configuration would be for your business was tricky, and scaling different services independently was impossible. also having multi regional redundancy was rare (remember when squarespace was manually carrying buckets of petrol for generators up many flights of stairs to keeps servers online post sandy?[1]).
AWS fixed much of that. But maybe things have changed in ways that meaningfully changes the calculus?
[1] https://www.squarespace.com/press-coverage/2012-11-1-after-s...
You're falling into the false dichotomy that always comes up with these topics: as if the choice is between the cloud and renting rack space while applying your own thermal paste on the CPUs. In reality, for most people, renting dedicated servers is the goldilocks solution (not colocation with your own hardware). You get an incredible amount of power for a very reasonable price, but you don't need to drive to a datacenter to swap out a faulty PSU, the on site engineers take care of that for you. I ordered an extra server today from Hetzner. It was available 90 seconds afterwards. Using their installer I had Ubuntu 24.04 LTS up and running, and with some Ansible playbooks to finish configuration, all in all from the moment of ordering to fully operational was about 10 minutes tops. If I no longer need the server I just cancel it, the billing is per hour these days.
Bang for the buck is unmatched, and none of the endless layers of cloud abstraction getting in the way. A fixed price, predictable, unlimited bandwidth, blazing fast performance. Just you and the server, as it's meant to be. I find it a blissful way to work.
AWS also made huge inroads in big companies because engineering teams could run their own account off of their budget and didn’t have to go through to IT to requisition servers, which was often red tape hell. In my experience it was just as much about internal politics as the technical benefits.
One factor was huge amounts of free credit for the first year or more for any startup that appeared above-board and bothered to ask properly.
Second, egress data being very expensive with ingress being free has contributed to making them sticky gravity holes.
Et al is for people, Et cetera is for things.
Edit: although actually many people on here are American so I guess for you aws is legally a person...
> although actually many people on here are American so I guess for you aws is legally a person...
Corporate legal personhood is actually older than Christianity, and it being applied to businesses (which were late to the game of being allowed to be corporations) is still significantly older than the US (starting with the British East India Company), not a unique quirk of American law.
As an American who studied Latin:
Et al. = et alii, "and other things", "among other things".
Etc. = et cetera, "and so on".
Either may or may not apply to people depending on context.
"To them, it’s way too convenient to be on AWS: not only it solves their problem, but it’s also a shiny object. It’s technically complex, it makes them look smart in front of other devs, "
why? why be so obnoxious to other people who you claim are being obnoxious to you. no need to read your blog post now/
The author touches on it briefly, but I'd argue that the cloud is immensely helpful for building (and tearing down) an MVP or proving an early market for a new company using startup credits or free tiers offered by all vendors. Once a business model has been proven, individual components and the underlying infrastructure can be moved out of the cloud as soon as cost becomes a concern.
This means that teams must make an up-front architectural decision to develop apps in a server-agnostic manner, and developers must stay disciplined to keep components portable from day one, but you can get a lot of mileage out of free credits without burning dollars on any infrastructure. The biggest challenge becomes finding the time to perform these migrations among other competing priorities, such as new feature development, especially if you're growing fast.
Our startup is mostly built on Google Cloud, but I don't think our sales rep is very happy with how little we spend or that we're unwilling to "commit" to spending. The ability to move off of the cloud, or even just to another cloud, provides a lot of leverage in the negotiating seat.
Cloud vendors can also lead to an easier risk/SLA conversation for downstream customers. Depending on your business, enterprise users like to see SLAs and data privacy laws respected around the globe, and cloud providers make it easy to say "not my problem" if things are structured correctly.
Seems like nowadays people seem less concerned with vendor lockin than they were 15 years ago. One of the reason to want to avoid lockin is to be able to move when the price gouging gets just a little bit too greedy that the move is worth the cost. One of the drawbacks of all these built in services at AWS is the expense of trying to recreate the architecture elsewhere.
Its not actually that hard to get your own server racked up in a data centre, I have done it. Since it was only one box that I built and installed at home I just shipped it and they racked it in the shared area and plugged the power and network in and gave me the IP address. It was cheaper than renting from someone like hetzner, was about £15 a month at the time for 1A and 5TB a month of traffic at 1gbps. Also had a one off install fee of £75.
At the time I did this no one had good gaming CPUs in the cloud, they are still a bit rare really especially in VPS offerings and I was hosting a gaming community and server. So I made a pair of machines in a 1U with dual machines in there and had a public and private server with raid 1 drives on both and redundant power. Ran that for a gaming server for many years until it was obsolete. It wasn't difficult and I think the machine was about £1200 in all, which for 2 computers running game servers wasn't too terrible.
I didn't do this because it was necessarily cheaper, I did it because I couldn't find a cloud server to rent with a high clockspeed CPU in it. I tested numerous cloud providers, sent emails asking for specs and after months of chasing it down I didn't feel like I had much choice. Turned out to be quite easy and over the years it saved a fortune.
Cheap shot maybe, but the fact that the page takes 10 seconds to load when it hits the HN front page is a great, inadvertant illustration of why you might want to use the cloud sometimes.
The failure mode of self-hosting is that your site gets hugged to death, the failure mode of the cloud is that you lose a ton of money. For a blog that doesn't earn you anything, the choice is clear.
Besides, you can just put it behind cloudflare for free.
That happens a lot to blogs deployed on the cloud too. They just need to put a small cache in front and they'll be able to serve one or two orders of magnitude more requests per second.
Only took 1 second for me.
Fittingly, his website was hugged to death
It's almost like Clouds are really good at scaling and some rented server isn't! Perfect, almost poetic.
I use Cloudflare in front of my personal stuff. Then it's just a quick DNS switch to go direct if I need to.
The cloud is a good idea. It becomes a bad idea when it is the only thing you know or, most likely, is the only cloud you know.
The first couple of paragraphs of price comparisons are useful. Then there are many paragraphs of sheer waffle. The author doesn't even seem able to define what "the cloud" is:
> The whole debate of “is this still the cloud or not” is nonsense to me. You’re just getting lost in naming conventions. VPS, bare metal, on-prem, colo, who cares what you call it. You need to put your servers somewhere. Sure, have a computer running in your mom’s basement if that makes you feel like you’re exiting the cloud more, I’ll have mine in a datacenter and both will be happy.
I read the whole thing and I didn't see any waffle. Sure, undeniably some excess word count, some emotion in responding to critics. But no waffle.
The "is this cloud or not" debate in the piece makes perfect sense. Who cares whether Hetzner is defined as "the cloud" or not? The point is, he left AWS without going to Azure or some other obvious cloud vendor. He took a step towards more hands on management. And he saved a ton of money.
The cheap hosting service they switched to is arguably "cloud".
If you can't drive to the location where your stuff is running, and then enter the building blindfolded, yet put your hands on the correct machine, then it's cloud.
I would really be interested in an actual comparison, where e.g. someone compares the full TCO of a mysql server with backup, hot standby in another data center and admin costs.
On AWS an Aurora RDS is not cheap. But I don't have to spend time or money on an admin.
Is the cost justified? Because that's what cloud is. Not even talking about the level of compliance I get from having every layer encrypted when my hosted box is just a screwdriver away from data getting out the old school way.
When I'm small enough or big enough, self managed makes sense and probably is cheaper. But when getting the right people with enough redundancy and knowledge is getting the expensive part...
But actually - I've never seen this in any if these arguments so far. Probably because actual time required to manage a db server is really unpredictable.
> Probably because actual time required to manage a db server is really unpredictable.
This, and also startups are quite heterogeneous. If you have an engineer on your team with experience in hosting their own servers (or at least a homelab-person), setting up that service with sufficient resiliency for your average startup will be done within one relaxed afternoon. If your team consists of designers and engineers who hardly ever used a command line, setting up a shaky version of the same thing will cost you days - and so will any issue that comes up.
Its a skillset that is out of favour at the moment as well but having someone who has done serverops and devops and can develop as well is a bit of a money saver generally because they open up possibilities that don't exist otherwise. I think its a skillset that no one really hired for past about 2010 when cloud was mostly taking off and got replaced with cloud engineers or pure devops or ops people but there used to be people with this mixed skillset in most teams.
every box is a screwdriver away
If you’re going to write a post about why self-hosting is better than cloud*, then it’s probably a good idea to make sure your site loads in under a minute.
* at least I assume what this post is; I’m still waiting for it to load.
a simple valid point wrapped in an enormous amount of garbage arguments from both sides. watching idiots argue is exhausting
I've been at too many startups with a devops team that would rather provision 15 machines with 4GB RAM THAN ONE WITH 64GB.
I once got into an argument with a lead architect about it and it's really easy to twist the conversation into "don't you think we'll reach that scale?" To justify complexity.
The bottom line is for better or worse, the cloud and micro services are keeping a lot of jobs relevant and there's no benefit in convincing people otherwise
I have a VPS. It costs me £1.34 per month. It's way over-powered for what I need it for.
However, one situation where I think the cloud might be useful is for archive storage. I did a comparison between AWS Glacier Deep Storage and local many-hard-drive boxes, for storing PB-scale backups, and AWS just squeaked in as slightly cheaper, but only because you only pay for the amount you use, whereas if you buy a box then you have to pay for the unused space. And it's off-site, which is a resilience advantage. And the defrosting/downloading charge was acceptable at effectively 2.5 months worth of storage. However, at smaller scales you would probably win with a small NAS, and at larger scales you'd be able to set up a tape library and fairly comprehensively beat AWS for price.
Yeah, but in 800 months you'd come out ahead with a dedicated server in your closet.
I run a tiny local dedicated server 24/7 that consumes around 10W on average, which is about $2/mo in electricity costs where I live.
I meant the upfront cost of the machine.
Not trying to be dismissive of the article but, the way it's written, it reads like a lot of whining.
He could have summed up with "AWS is expensive, host your own server instead".
What I always say when given a false choice: ¿porque no las dos?
vcpu, iops, transfer fees, storage -- they are all resources going into a pool .
If Hetzner is giving you 10TB for $100 , then host your static files/images there and save $800.
Apps are very modular. You have services, asyncs, LBs, static files . Just put the compute where it is most cost effective.
You don't have to close your AWS account to stick it to the man. Like any utility, just move your resources to where they are most affordable.
Help convince me I should be confident taking responsibility for:
* off-site db backups
* a guaranteed db restore process
* auditable access to servers
* log persistence and integrity
* timely security patching
* intrusion detection
so that my employer can save money.
How many URLs of Google et al failing to provide per instance security (leaking your files etc) to other users would you like to see?
It's an informative post but I really dislike the language and style that are becoming common in this kind of posts, e.g.:
> Look, first of all, you’re as unique as the other 1000 peanut gallery enjoyers that have made the same astute observation before you. Congratulations. But you’re absolutely missing the point.
The article is hugged to death. Maybe it wasn't hosted in the cloud?
> Most people complaining about what I did happen to have “devops”, “cloud engineer”, “serverless guy”, “AWS certified”, or something similar in their bio.
“It is difficult to get a man to understand something, when his salary depends on his not understanding it.”
― Upton Sinclair
Jeez, this was a painful read. I actually stopped after a few paragraphs and asked AI to make it more technically focused and remove the ranting so I could stomach it.
Strawman arguments, ad hominem attacks and Spongebob mocking memes, and the casual venturing into conspiracy theories and malicious intentions...
> Why do all these people care if I save more money or not? ... If they’re wrong, and if I and more people like me manage to convince enough people that they’re wrong, they may be out of a job soon.
I have a feeling AWS is doing fine without him. Cloud is one of the fastest growing areas in tech because their product solves a need for certain people. There is no larger conspiracy to keep cloud in business by silencing dissent on Twitter.
> You will hear a bunch of crap from people that have literally never tried the alternative. People with no real hands-on experience managing servers for their own projects for any sustained period of time.
This is more of a rant than a thoughtful technical article. I don't know what I was expecting, because I clicked on the title knowing it was clickbait, so shame on me, I guess...
Is this what I'm missing by not having Twitter?
As a note hetzner has a lot of auction servers and I believe they lack the setup fee
They have also threatened to cancel my account more than once because I typed "ipfs daemon".
https://github.com/ipfs/kubo/issues/10327
https://discuss.ipfs.tech/t/moved-ipfs-node-result-netscan-d...
>This happens with Hetzner all the time because they have no VLANs and all customers are on a single LAN and IPFS tries to discover other nodes in the same LAN by default.
Hetzner is also sinkholed by lots of EDR products because they host a ton of malicious garbage. They are a bad actor.
Why is it their job to be the arbiters of what customers are allowed to do on their platform?
Vercel is my favorite.. They charge you to pay for AWS.
I think a lot of teams using cloud are using SaaS rather than IaaS. They want a redis and a postgres and a S3 and a ... You can set all that up on a server, but it's not very fun if you've never done it before.
"I FINALLY got everything off the cloud"
...
...
"P.S. follow me on Twitter"
So uh, not everything
Sorry but my $3 AWS instance is still cheaper than all of those options.
If you need a lot of, well, anything, be it compute, memory, storage, bandwidth etc., of course cloud stuff is going to be more expensive... but if you don't need that, then IMO $3/mo on-demand pricing really can't be beat when I don't have to maintain any equipment myself. Oracle also offers perpetually free VM instances if you don't mind the glow.
With a quick LLM-assisted search, looks like the cheapest EC2 instance is t4g.micro, which comes in at $2.04/mo. It has 2 vCPUs and and only 512MiB of RAM. (I assume that doesn't include disk; EBS will be extra.)
I can certainly see a use for that small amount of compute & RAM, but it's not clear that your level of needs is common. I've been paying for a $16/mo VPS (not on AWS) for about 15 years. It started out at $9/mo, but I've upgraded it since then as my needs have grown. It's not super beefy with 2 vCPUs, 5GiB of RAM, and 60GiB of disk space (with free data ingress/egress), but it does the job, even if I could probably find it cheaper elsewhere.
But not at Amazon. Closest match is probably a t3.medium, with 2 vCPUs and 4GiB RAM. Add a 60GiB gp2 EBS volume, and it costs around $35/mo, and that's not including data transfer.
The point that you're missing is we're not looking for the cheapest thing ever, we're looking for the cheapest thing that meets requirements. For many (most?) applications, you're going to overpay (sometimes by orders of magnitude) for AWS.
You say "if you need a lot", but "lot" is doing a bit of work there. My needs are super modest, certainly not "a lot", and AWS is by far not the cheapest option.
Just get a raspberry pi and run it from your own home internet. You should already be paying for a VPN service and your regular internet service, so you should be able to trivially work out a self-hosted solution. You'll recover your costs inside of two years and come out the other end better off for it.
Don't give the big cloud companies an inch if you don't absolutely have to. The internet needs and deserves the participation of independent people putting up their own services and systems.
Amazon really doesn't care if your $10,000 bed folds up on you like a sandwich and cooks you when AWS us-east-1 goes down, or stops your smart toilet from flushing, or sets bucket defaults that allow trivial public access to information you assume to be secure, because nobody in their right mind would just leave things wide open.
Each and every instance of someone doing something independently takes money and control away from big corporations that don't deserve it, and it makes your life better. You could run pihole and a slew of other useful utilities on your self-hosted server that benefit anyone connected to your network.
AI can trivially walk you through building your own self-hosted setups (or even set things up for you if you entrust it with an automation MCP.)
Oracle and AWS and Alphabet and the rest shouldn't profit from eating the internet - the whole world becomes a better place every time you deny them your participation in the endless enshittification of everything.
yet another obsessive take on "cloud is bad and expensive" eh? I think they vastly forget the value of some SaaS offerings in terms of time saving for small companies. running and managing numerous DBs, k8s clusters, ci/cd pipelines and stateless container systems is simply impossible with a team of 1-2 people. sure if the setup is simple and only requires a few classic components, this is way cheaper and for a 99.9% SLA will work fine. otherwise it only makes sense if you had very large cloud bills and can dedicate multiple engineers to the newly created tasks.
Not agreeing/disagreeing with your core point, but this doesn't seem right:
> running and managing numerous DBs, k8s clusters, ci/cd pipelines and stateless container systems is simply impossible with a team of 1-2 people.
That's a medium to large homelab worth of stuff, which means it can be run by a single nerd in their spare time.
Homelab =/= Production systems
The gulf between these two insofar as what approach, technologies, and due-diligences are necessary is vast.
I think we've gone a little nuts defining "production system" these days. I've worked for companies with zero-downtime deployments and quite a lot of redundancy for high availability, and for some applications it's definitely worthwhile.
But I think for many (most?) businesses, one nine is just fine. That's perfectly doable by one person, even if you want, say, >=96% uptime, which allows for 350 hours of downtime per year. Even two nines allows for ~88 hours of downtime per year, and one person could manage that without much trouble.
Most businesses aren't global. Downtime outside regular business hours for your timezone (and perhaps one or two zones to the west and east of you) is usually not much of a problem, especially if you're running a small B2B service.
For a small business that runs on 1-3 servers (probably very common!), keeping a hot spare for each server (or perhaps a single server that runs all services in a lower-supported-traffic mode) can be a simple way to keep your uptime high without having to spend too much time or money. And people don't have to completely opt out of the cloud; there are affordable options for e.g. managed RDBMS hosting that can make maintenance and incident response significantly easier and might be a good choice, depending on your needs.
(Source: I'm building a small one-person business that is going to work this way, and I've been doing my research and gaming it out.)
Idiotic piece - the purpose of 'the cloud' is to scale large demand applications. Rental hardware can't really do that.
The post is about that 99% of companies that will never go large scale. Its point is that they don't need cloud, buying a server or two is all they need.
The vast majority of businesses are not "large demand applications".
> Idiotic piece
That's unnecessary; please don't do that here. Weird that you created an account just to post an unsubstantive comment.
An argument which begins by reducing an entire industry down to a single "purpose" is not convincing.