I’m considering a business plan for people getting in to self-hosting. Essentially I sell you a Mikrotik router and a refurbished tiny x86 server. The idea is that the router plugs in to your home internet and the server into the router. Between the two they get the server able to handle incoming requests so that you can host services on the box and address them from the broader Internet.
The hypothesis is that $150 of equipment to avoid dozens of hours of software configuration is a worthwhile trade for some customers. I realize some people want to learn particular technologies and this is a bad fit for them. I think there are people out there that want the benefit of self-hosting, and may find it worth it to buy “self-hosting in a box”.
What do you think? Would this be a useful product for some people?
I assume “SFF PC” means “Small form-factor personal computer”.
The value add is not having to make a large number of technical decisions. IPv4 vs v6, which firewall rules to use, port-forwarding vs DMZ, flavor of Linux, partition scheme, filesystem type, application packaging system, and on and on. For many people they don’t care about these decisions, they want “to put something on the Internet” and do it safely. While safety isn’t a binary, and engineering is full of tradeoffs, an experienced practitioner can answer many of these questions reflexively and come out with good enough answers for some customers.
In the end the customer should be able to dig in and change whatever they want. But I want to see if flipping the decision dependency around will help. IE, start with stuff that works, then change things, rather than start with parts and make all the decisions before anything works.
Probably not much for people on a self hosting community, but those that want to get away from subscriptions and steal your data as a service cloud providers that might need some reassurance that they’ll have a working system.
How will you provide long term maintenance of their server for a one time payment of 150$?
My current thinking is the margin on the hardware would be intentionally low, essentially the cost of the hardware %+10 for configuring it a bit, installing NixOS, etc.
The business would survive on support and hosted services. Something like $20/month which gets you access to support to answer questions, help configure applications, troubleshoot issues, etc. Possibly rolling upgrades of your installed software on your behalf. Alerts on urgent security vulnerabilities. Could also handle tricky things like custom DNS (email servers, certificates) and off-site backups. I’m not totally sure what all would be included, but the goal is to make money while providing value, not build a garden or rent-seek.
So the problem with thin margins on the hardware side is what’s stopping a user from just installing their own OS once they figure out they can do the same thing you’re doing on the same hardware?
$20 per month would be enough to discourage me. It’s another relatively costly computer-related subscription and I already feel like I’m losing a battle to keep those minimal. There would have to be some very clear benefits for that price.
There would have to be some very clear benefits for that price.
Agreed, it would need to be very clear, and additionally we’d need to plan that a certain percentage of customers would grow out of a basic support offering, either by becoming experts or by growing their install size and complexity.
$20 per month would be enough to discourage me. It’s another relatively costly computer-related subscription and I already feel like I’m losing a battle to keep those minimal.
Understandable. Is there a price you think would be reasonable? What would you want for that price?
While true I feel like your comment misses the point. A raspberry pi is just a computer, not a magic solution box that’s kept maintained and updated by some guy. Their product isn’t a service, it’s just the device.
The fact that it’s an option that even remotely works is my point. They sell hardware. They don’t support software. The community does that. There is something to be gained from having a uniform platform for learning self hosting responsibly.
A Raspberry pi isn’t particularly great at any one thing. It’s greatest strength comes in bundling everything you need in a box at an affordable price. Once you know where your pain points are then you can build/design a system that overcomes those shortcomings.
Having a starter kit would be an easy way to get more people in the space. Would it cost $35 of course not. Level1Techs made their KVM to meet their own requirements and then the community benefits. To me, this project has that kind of energy. Or at least the potential for it.
The hypothesis is that $150 of equipment to avoid dozens of hours of software configuration
OK fair try, but you also need to sell me 20-25 TB of disk space on 5 spindles (plus a SSD for the bootdisk), 64 GB RAM (with a chance to go up to 128) and the CPU must have 16 threads or more.
What kind of workload do you run that makes you confident you need that much hardware? Do you think people just beginning could get buy on 4 cores and 8 GB RAM for a while? How long before you think most people need more?
This will be the spec for my next server. The current one is smaller, and several years old
I have several different requirements for my server, for example, my son does video editing and needs lots of storage. I want to experiment with more VM’s and containers, therefore RAM and threads.
Do you think people just beginning could get buy on 4 cores and 8 GB RAM for a while?
For most people I think they just want to have some NAS and a reliable machine. But please grant them 16 GB, otherwise they would ask why their laptop has so much more than their server :-)
Honestly, not really. I’m just kind of at a point in my life where it’s something I’m thinking about looking into, and an out-of-the-box option like this would be really handy.
I probably would. However it has become increasingly obvious that the flaws with solutions so far have been in the organisation. Not so much the particular hardware or software. If I’m going to buy something I’d like some hope that it’ll be there in 5 or 10 or 20 years. So please if you go serious with this, look into worker-owned organizations because I’m tired of dodging profit-maximizing traps and pretend-non-profit landmines. If the people building and supporting the thing aren’t the ones deciding what to do with the revenue and profit, you’re the only one doing it and you’re going to make mistakes that will hurt them and us. And then you become a landmine to dodge.
These are great points, and I fully agree. I’d be interested in knowing what kind of license or corporate structure or contract would give you confidence that the organization is worth investing in. I could put all the software out with a really strong Affero license so that you’ve got the source code, but I get the impression that you, like me, want more than that. Corporations like Mondragon are interesting to me, and I’m aware of a few different tech cooperative organizations. I’m not confident that a cooperative structure alone is enough. Yes, it helps avoid the company taking VC money, shooting for the moon, failing, and then selling everything that’s not clearly legally radioactive. But it doesn’t protect you against more insidious forces like the founders selling to private capital and adjusting the EULA every few months until they have the right to sell off your baby photos.
I’ve been batting around the idea of creating a compliment to the “end-user license agreement” - the “originating company license agreement”. Something like a poison pill that forces the company to pay out to customers in the event of a data breach, sale of customer data, or other events that a would-be acquirer may think is worth it for them.
I’m just not sure yet what kinds of controls would be strong enough to convince people who have been burned by this sort of thing in the past. What do you think?
Purely on the product side, if I decide to buy it, I wouldn’t buy it for myself. I’d buy it for friends and family who are not that tech literate. Either to make my life easier to give them self-hosted services, or ideally for themselves to be able to do so. I want this product to be a non-shitty, open source “Synology,” from a firm I can trist to support it for a very long time. Doesn’t have to have that form factor. And I’m totally fine with an ongoing subscription. I’d like to be able to say - hey friend, buy this from ACME Co-op and sign up for their support plan. Follow the wizard and you’ll have Immich, Nextcloud, etc. A support plan might include external cloud HTTP proxy with authentication and SSL that makes access trivial. Similar to how Home Assistant’s subscription (Nabu Casa) works. It could also include a cloud backup. Perhaps at a different subscription rate.
I don’t know enough to say what the structure should be but this should not be possible:
But it doesn’t protect you against more insidious forces like the founders selling to private capital
It implies that the founders have more voting power and ownership than the rest of the people in the org. In my mind, everyone should have an equal vote, which should prevent a sale on the whim of the founders or another minority group. If a sale is in the cards, a majority of the people in the org should have to approve for it to proceed. And this shouldn’t be advisory but a legal barrier to pass.
If I were to start a firm today, I’d be looking into this because not only this is the kind of firm I’d like to work in, but I think so would quite a few people in software. And those aren’t the dumb kids.
I can also say that as a customer, the few worker co-ops I’ve able to buy things from give me a much more trustworthy impression than the baseline. They just behave differently. Noticeably more ethically.
But it doesn’t protect you against more insidious forces like the founders selling to private capital
It implies that the founders have more voting power and ownership than the rest of the people in the org. In my mind, everyone should have an equal vote, which should prevent a sale on the whim of the founders or another minority group.
I’m not confident that simple democracy is enough. While I do expect that a one-worker-one-vote system would make it harder to sell out, it’s still possible. I do think that a cooperative has many benefits. I just want to make it fatal to the business to go down certain dark paths: selling user data, seller user compute, selling user attention, etc.
I wish there were more examples of functional high-tech cooperatives I could learn lessons from.
If I were to start a firm today, I’d be looking into this because not only this is the kind of firm I’d like to work in, but I think so would quite a few people in software. And those aren’t the dumb kids.
Coming from someone who started selfhosting on a pi 2B (similar-ish specs), you’d be surprised. If you don’t need anything fast or fancy, that 1GB will go a long way, and plenty of selfhosted apps require very little CPU. The only real problem I faced was that all HTTPS-related network tasks were limited at ~3MB/s, as that is how fast my pi could encrypt the data (presumably, I just saw my webserver utilising the entire CPU and figured this was the most likely explanation)
I’m currently hosting like 5 vms on a proxmox host (mostly ubuntu vms- pihole, nextcloud, home assistant, etc), which is an i5 4590 with 32 gb ram and I’m running up against the limits of how much ram I can provision and if 2 or more of my vms are doing something intensive at the same time I’m pinning the CPU. I don’t think my use-case is that crazy for someone doing a little self-hosting.
Luxury! My homeserver has an i5 3470 with 6GB or RAM (yes, it’s a cursed 4+2 setup)! </badMontyPythonReference>
Interesting, I also run Nextcloud and pihole, and vaultwarden, jellyfin, paperless-ngx, gitea, vscode-server and a minecraft server (every now and then).
You’re right that such a system really does show its age, but only when doing multiple intensive tasks at the same time. I try not to backup my photos to Nextcloud while running minecraft, for example, as the imagine identification task pins my CPU at 100%. So yes, I agree, you’re probably not doing anything out of the ordinary on your setup.
The point I was trying to make still stands though, as that pi 2B could run more than I would’ve expected beforehand. I believe it once even ran jellyfin, a simple file server, samba, and a webserver with a simple HTML website. Jellyfin worked just fine, as long as the pi didn’t have to transcode (never got hardware transcoding to work).
It is funny that you should run out of memory, seeing as everything fits (albeit, just barely) on my machine in 1/5 the memory. Would de overhead of running VM’s account for such a large difference?
I’m running the recognize app on nextcloud which I think requires at least 4-5 GB RAM, so I have 6 dedicated to that VM. I’m pretty sure the recommendation for Ubuntu in general was 2 GB RAM so I gave my pihole half that. Home assistant wanted 4 GB, but I gave it 2. I think my Jellyfin server has like 6 and I have another VM with like 4. So that’s a total of like 19gb RAM provisioned. Plus I have a 2 TB zfs pool for my nextcloud VM. When I go into proxmox it tells me I’m using like 29.5 GB.
I suspect if someone was using docker or some other sort of containerization one could expect better performance than what I am getting with VMs.
It was ok at the time, and if it isn’t ok now, that means you want to run something that is too bloated for its own good.
Really though, special hardware for this doesn’t make too much sense. A raspberry pi with two ethernet interfaces would be great, but if you can live with ethernet plus wifi, the current rpi’s will do it. Otherwise there are lots of similar boards that really do have two ethernet.
I have not really felt much use for self hosted server hardware at home. I use VPS’s for that and it’s less hassle. Maybe it doesn’t count as completely self hosted, but conceptually it’s a miniature colo box.
I admire the thought of lowering the barrier to entry to start self-hosting for “normies”. Not sure where you are located, but where I am, this price point is not realistic even for used equipment, not including RAM or storage. I’m not really sure what value add you are bringing to the table that one wouldn’t get from just buying used hardware from an office surplus and if one is very inexperienced in self-hostong, looking into something like LTT is partnered with like Hexos.
this price point is not realistic even for used equipment, not including RAM or storage
I’m doing experiments currently on a refurbished Intel i5-6500 with 8Gb DDR4 and a 0.5Tb SSD. It’s tiny, quiet (~45 decibels) and so far runs ~8 watts idle, 25 watts normal usage. I haven’t stress-tested the power draw. The router I’m testing with is a Mikrotik hEX lite 5. That’s around ~$150, though clearly if you are accustomed to more “rack-mount” style homelab these will seem very modest.
What I’m testing for now is getting representative loads on the devices to see how they perform.
I’m not really sure what value add you are bringing to the table that one wouldn’t get from just buying used hardware from an office surplus and if one is very inexperienced in self-hostong, looking into something like LTT is partnered with like Hexos.
Oh, I totally agree, my value add just isn’t there if you are experienced at hosting. The value add is to help people get started, and to keep them running at a modest level. Not everyone wants to experiment with Kubernetes at home or train LLMs. Some folks just want a password manager, a shared calendar, something to organize their tax documents, a pihole, and a Minecraft server for their kids.
I don’t follow LTT, I was under the impression it was more hardware reviews for the experienced than tutorials to help people get started.
I’ve read a bit about Hexos, I’m thinking of some similar things, and it would make sense to work with them. I’m excited for their coming beta.
I recently upgraded my homelab/self-hosting server from an old Dell T410 with dual X5650’s (2 - 6 core/12 thread CPU) and 24 GB ram to an old Dell Optiplex (7020 I think) with an i5-4590 (4 core/4 thread) and 32 GB ram. Its barely enough for a proxmox host with 5 VMs; but way faster than the old T410.
If you are offering some sort of self-hosting box, would it be bundled with some sort of software for someone to easily spin up whatever services they want?
Are you going to be able to make money at the $150 mark with all this hardware and configuration? If you are targeting people who are new to self-hosting, it will need to be a complete package (will need to have ram and storage installed).
The idea is that the router plugs in to your home internet and the server into the router. Between the two they get the server able to handle incoming requests so that you can host services on the box and address them from the broader Internet.
Why would I need a separate router for that? I’d need to configure the main router anyway.
I would absolutely want the extra router because most people have one from their service provider. For self hosting, you want an additional router with your own software.
Market to tax funded institutions. If you can market “self hosted” as cheaper and easier than mother solutions you’ll have guaranteed clients for a long time.
That’s an interesting idea I hadn’t thought much about. I’ve been more focused on individuals than organizations. Do you have experience with tax-funded institutions? I assumed they generally have strict procurement rules and long support contracts with large established players by policy.
It’s not like a package thing you can sell if you’re not supporting it. Then you’re just selling hardware at an inflated price. It’s not even self-hosting at that point. Why wouldn’t you just pay a regular company for a product?
What is the aim? People who want to get into it, but does not know how, or experts?
Think half of the attraction of selfhosting is the diy aspect.
What extra would this bring if people can just buy the parts cheaper?
And for those who only want the out of the box experience why would this be better than, let’s say a beestation? (Yeah price, I know, but you obviously would not have the same support level.)
What is the aim? People who want to get into it, but does not know how, or experts? Think half of the attraction of selfhosting is the diy aspect.
I don’t disagree, and I would imagine what I’m offering would only be useful to people who are very early and don’t yet know they enjoy the DIY aspect.
The aim, though, is this: I’ve enjoyed self-hosting. It’s given me some powers that most people don’t get to have who aren’t also technical professionals. I’m also deeply frustrated by the environment created by the various major tech companies. If I can, I’d like to lower the barrier for people to get some of those powers without having to become experts and to make it more feasible for them to do what they want to do, rather than just what they are permitted to do.
What extra would this bring if people can just buy the parts cheaper?
Much shorter time going from “how can I control some of my own data” to "I’m running NextCloud, and its kinda like iCloud/Google Drive/Whatever Microsoft does and it’s running right here under my control! Not everyone knows the path from buying parts online to having a working reverse-proxy and reasonable firewall rules. Also, standardization makes it much easier to support people, which is really what the business would be doing.
why would this be better than, let’s say a beestation?
I knew about Synology, but as a NAS product, which assumes a certain familiarity with backup schemes, etc. Kind of a prosumer-only thing. The Beestation is new to me, thanks for the tip. Quite possible what I’m proposing would have some overlap and compete with it, I’ll have to read up on it.
No. People who want the benefit of self housing without worrying about hardware will rent a vps or something simpler. The hard part of hardware isn’t the purchase, it’s the maintenance.
People who want the benefit of self housing without worrying about hardware will rent a vps or something simpler.
That’s certainly an option. I think of dedicated hardware as working for several different people, some of which care a great deal about not using a VPS provider because they don’t trust them with their data, or don’t trust them to be around for a long time, or don’t trust them not to raise the prices.
The hard part of hardware isn’t the purchase, it’s the maintenance.
I’m inclined to agree, but I’ve been doing hardware for a long time as a hobbyist and I sometimes forget how far I’ve come. It sounds like you might be somewhat like me in that regard. I’m often surprised when people see assembling system parts and flashing an OS as a complex, inscrutable task.
What do you see as the hard part of maintenance? Scheduling time to do it? Unexpected errors or failures?
I agree with this. Self-hosting requires the user to understand their network, their software, how it all interacts.
If you provide a hardware product and call it a solution, people are going to expect a turn-key solution like a plug-and-play router.
You’re going to end up supporting a bunch of newbies who, by no fault of their own, can’t tell you an error code in the console let alone whatever UI you give them.
I think a better solution would be a course that walks newbies through self hosting.
I think this needs to exist, but as a community supported system, not as a commercial product.
Pick a set of open technologies - but not the best, lightest weight, just pick something open.
Come up with a security architecture that’s reasonably safe and only adds a moderate amount of extra annoyance, and build out a really generic “self-hosted web hosting and VM company-like thingy” system people can rally around.
Biggest threat to this, I think, is that this isn’t the 90s and early 2000s any longer, and for a big project like this, most of the oxygen has been sucked out already by free commercial offerings like Facebook. The technical family friend offering to self-host email or forums or chat no longer gets gratitude and love, they get “why not Facebook?”
So… small group effort, resistant to bad actors joining the project to kill it, producing a good design with reasonably safe security architecture, that people can install step by step, and have fun using while they build and learn it.
I think a possibility is a series of open source anvil or nixos scripts that you can run on most hardware with minimal changes, in an extendable architecture of some kind to add or remove functionality and they perhaps get maintained by the community or some structure of the kind of Linux distributions.
This could enable people with minimal skills set up and maintain a reasonably useful but secure environment just by changing a few variables.
Nixos is an os that’s defined by its config stored in .nix files. Everything is defined here all the software and configurations. Two people with the same script will have the exact same os.
Any changes you make that aren’t in the scripts won’t be present when you reboot.
You could maintain a very custom linux distribution (kinda) by just maintaining these config scripts.
So a user wouldn’t need to install all required software and dependencies. They could get a nixos and the self-host config and adjust some settings and have a working system straight after install.
I think this needs to exist, but as a community supported system, not as a commercial product.
…
The technical family friend offering to self-host email or forums or chat no longer gets gratitude and love, they get “why not Facebook?”
I think this is a great point, it doesn’t help much to create a business that ends up with the same incentives and the same end-game as the existing systems.
So… small group effort, resistant to bad actors joining the project to kill it, producing a good design with reasonably safe security architecture, that people can install step by step, and have fun using while they build and learn it.
That is precisely what I’m looking to build. I don’t want to get rich, I want people without 10 years of industry experience to get some of the benefits we have all been able to build for ourselves.
We already have that, the first problem is we have like a dozen of them, a few are even well supported. The second problem is that usually the technical knowledge required to set up the systems are still lower than the technical knowledge required to keep it running.
I’ve been struggling to wrap my head around a good security architecture for my mspencer.net replacement crap. Could I bug you for links?
I figured out a while ago to keep VM host management on a management VLAN, and I put each service VM on its own VLAN with heavy, service-specific firewalling and a private OS update repo mirror - but after hearing about ESXi jackpotting vulns and Broadcom shenanigans, I’ve gotten really disheartened. I’d love some safe defaults.
It sounds like you’re getting into the keeping it running phase.
First, going back to your previous comment, self-hosting email is difficult. It’s not hard for a small provider to end up blacklisted and you’re probably kind of just done at that point and it will feel very unfair. I get that it’s a fun set of technical challenges, but you couldn’t pay me enough to help someone self-host email.
Second, guessing, but it sounds like you may be trying to expose your services directly and doing a lot to make that work which goes against what most would recommend for hosting your own services. Big companies don’t expose their intranet like that, follow their example. Almost every guide or system is going to warn against that. If you’re going to host more than one thing, highly recommend focusing on minimizing entry points and looking into a VPN-like solution for accessing most if not all of your services. Still spend time on securing your intranet, but most of your risk is going to come from how hard it is for people to get past the front door (or doors).
Thank you for your reply, but to be clear, I’m not looking for individual details to be spelled out in comments. What you said is absolutely correct, thoughtful, and very helpful. But emotions are running a little high and I’m worried I’ll accidentally lash out at someone for helping. Apologies in advance.
But do you have any links? Beyond just the general subjects of security architecture, secure design, threat modeling, and attack surface identification, I’d love to see this hypothetical “generic VM and web application housing provider in a box” come with a reasonably secure default architecture. Not what you’re running, but how you’re running it.
Like, imagine decades in the future, internet historians uncover documentation and backups from a successful generic hosting company. They don’t necessarily care what their customers are hosting, their job is to make sure a breach in one customer’s stuff doesn’t impact any other customer. The documentation describes what policies and practices they used for networking, storage, compute, etc. They paid some expensive employees to come up with this and maintain it, it was their competitive advantage, so they guarded it jealously.
I’d want to see that, but (a) a public, community project and (b) now, while it’s still useful and relevant to emulate it in one’s own homelab.
If I can get some of that sweet, sweet dopamine from others liking the idea and wishing for my success, maybe I can build my own first version of it, publish my flawed version, and it can get feedback.
You are not logged in. However you can subscribe from another Fediverse account, for example Lemmy or Mastodon. To do this, paste the following into the search field of your instance: !selfhosted@lemmy.world
A place to share alternatives to popular online services that can be self-hosted without giving up privacy or locking you into a service you don’t control.
Rules:
Be civil: we’re here to support and learn from one another. Insults won’t be tolerated. Flame wars are frowned upon.
No spam posting.
Posts have to be centered around self-hosting. There are other communities for discussing hardware or home computing. If it’s not obvious why your post topic revolves around selfhosting, please include details to make it clear.
Don’t duplicate the full text of your blog or github here. Just post the link for folks to click.
Submission headline should match the article title (don’t cherry-pick information from the title to fit your agenda).
No trolling.
Resources:
selfh.st Newsletter and index of selfhosted software and apps
If I wanted that I would just buy Synology/QNAP/Zima, etc.
What’s the value-add over just buying a SFF PC?
I assume “SFF PC” means “Small form-factor personal computer”.
The value add is not having to make a large number of technical decisions. IPv4 vs v6, which firewall rules to use, port-forwarding vs DMZ, flavor of Linux, partition scheme, filesystem type, application packaging system, and on and on. For many people they don’t care about these decisions, they want “to put something on the Internet” and do it safely. While safety isn’t a binary, and engineering is full of tradeoffs, an experienced practitioner can answer many of these questions reflexively and come out with good enough answers for some customers.
In the end the customer should be able to dig in and change whatever they want. But I want to see if flipping the decision dependency around will help. IE, start with stuff that works, then change things, rather than start with parts and make all the decisions before anything works.
Probably not much for people on a self hosting community, but those that want to get away from subscriptions and steal your data as a service cloud providers that might need some reassurance that they’ll have a working system.
How will you provide long term maintenance of their server for a one time payment of 150$?
My current thinking is the margin on the hardware would be intentionally low, essentially the cost of the hardware %+10 for configuring it a bit, installing NixOS, etc.
The business would survive on support and hosted services. Something like $20/month which gets you access to support to answer questions, help configure applications, troubleshoot issues, etc. Possibly rolling upgrades of your installed software on your behalf. Alerts on urgent security vulnerabilities. Could also handle tricky things like custom DNS (email servers, certificates) and off-site backups. I’m not totally sure what all would be included, but the goal is to make money while providing value, not build a garden or rent-seek.
So the problem with thin margins on the hardware side is what’s stopping a user from just installing their own OS once they figure out they can do the same thing you’re doing on the same hardware?
Nothing stops them, but that’d be fine. If they buy the hardware they should be able to do what they want with it.
$20 per month would be enough to discourage me. It’s another relatively costly computer-related subscription and I already feel like I’m losing a battle to keep those minimal. There would have to be some very clear benefits for that price.
Agreed, it would need to be very clear, and additionally we’d need to plan that a certain percentage of customers would grow out of a basic support offering, either by becoming experts or by growing their install size and complexity.
Understandable. Is there a price you think would be reasonable? What would you want for that price?
Raspberry pi was able to do it with $35.
Raspberry Pi is not a server. That people use it as one does not mean it’s fit for purpose.
While true I feel like your comment misses the point. A raspberry pi is just a computer, not a magic solution box that’s kept maintained and updated by some guy. Their product isn’t a service, it’s just the device.
The fact that it’s an option that even remotely works is my point. They sell hardware. They don’t support software. The community does that. There is something to be gained from having a uniform platform for learning self hosting responsibly.
A Raspberry pi isn’t particularly great at any one thing. It’s greatest strength comes in bundling everything you need in a box at an affordable price. Once you know where your pain points are then you can build/design a system that overcomes those shortcomings.
Having a starter kit would be an easy way to get more people in the space. Would it cost $35 of course not. Level1Techs made their KVM to meet their own requirements and then the community benefits. To me, this project has that kind of energy. Or at least the potential for it.
OK fair try, but you also need to sell me 20-25 TB of disk space on 5 spindles (plus a SSD for the bootdisk), 64 GB RAM (with a chance to go up to 128) and the CPU must have 16 threads or more.
What kind of workload do you run that makes you confident you need that much hardware? Do you think people just beginning could get buy on 4 cores and 8 GB RAM for a while? How long before you think most people need more?
This will be the spec for my next server. The current one is smaller, and several years old
I have several different requirements for my server, for example, my son does video editing and needs lots of storage. I want to experiment with more VM’s and containers, therefore RAM and threads.
For most people I think they just want to have some NAS and a reliable machine. But please grant them 16 GB, otherwise they would ask why their laptop has so much more than their server :-)
I’d probably buy it.
Do you already have an idea of what kinds of things you’d want to run on it?
Honestly, not really. I’m just kind of at a point in my life where it’s something I’m thinking about looking into, and an out-of-the-box option like this would be really handy.
I probably would. However it has become increasingly obvious that the flaws with solutions so far have been in the organisation. Not so much the particular hardware or software. If I’m going to buy something I’d like some hope that it’ll be there in 5 or 10 or 20 years. So please if you go serious with this, look into worker-owned organizations because I’m tired of dodging profit-maximizing traps and pretend-non-profit landmines. If the people building and supporting the thing aren’t the ones deciding what to do with the revenue and profit, you’re the only one doing it and you’re going to make mistakes that will hurt them and us. And then you become a landmine to dodge.
These are great points, and I fully agree. I’d be interested in knowing what kind of license or corporate structure or contract would give you confidence that the organization is worth investing in. I could put all the software out with a really strong Affero license so that you’ve got the source code, but I get the impression that you, like me, want more than that. Corporations like Mondragon are interesting to me, and I’m aware of a few different tech cooperative organizations. I’m not confident that a cooperative structure alone is enough. Yes, it helps avoid the company taking VC money, shooting for the moon, failing, and then selling everything that’s not clearly legally radioactive. But it doesn’t protect you against more insidious forces like the founders selling to private capital and adjusting the EULA every few months until they have the right to sell off your baby photos.
I’ve been batting around the idea of creating a compliment to the “end-user license agreement” - the “originating company license agreement”. Something like a poison pill that forces the company to pay out to customers in the event of a data breach, sale of customer data, or other events that a would-be acquirer may think is worth it for them.
I’m just not sure yet what kinds of controls would be strong enough to convince people who have been burned by this sort of thing in the past. What do you think?
Purely on the product side, if I decide to buy it, I wouldn’t buy it for myself. I’d buy it for friends and family who are not that tech literate. Either to make my life easier to give them self-hosted services, or ideally for themselves to be able to do so. I want this product to be a non-shitty, open source “Synology,” from a firm I can trist to support it for a very long time. Doesn’t have to have that form factor. And I’m totally fine with an ongoing subscription. I’d like to be able to say - hey friend, buy this from ACME Co-op and sign up for their support plan. Follow the wizard and you’ll have Immich, Nextcloud, etc. A support plan might include external cloud HTTP proxy with authentication and SSL that makes access trivial. Similar to how Home Assistant’s subscription (Nabu Casa) works. It could also include a cloud backup. Perhaps at a different subscription rate.
I don’t know enough to say what the structure should be but this should not be possible:
It implies that the founders have more voting power and ownership than the rest of the people in the org. In my mind, everyone should have an equal vote, which should prevent a sale on the whim of the founders or another minority group. If a sale is in the cards, a majority of the people in the org should have to approve for it to proceed. And this shouldn’t be advisory but a legal barrier to pass.
If I were to start a firm today, I’d be looking into this because not only this is the kind of firm I’d like to work in, but I think so would quite a few people in software. And those aren’t the dumb kids.
I can also say that as a customer, the few worker co-ops I’ve able to buy things from give me a much more trustworthy impression than the baseline. They just behave differently. Noticeably more ethically.
I’m not confident that simple democracy is enough. While I do expect that a one-worker-one-vote system would make it harder to sell out, it’s still possible. I do think that a cooperative has many benefits. I just want to make it fatal to the business to go down certain dark paths: selling user data, seller user compute, selling user attention, etc.
I wish there were more examples of functional high-tech cooperatives I could learn lessons from.
I strongly agree with this sentiment.
https://www.freedombox.org/ ?
Doesn’t seem like you could self-host a whole lot with that…
Coming from someone who started selfhosting on a pi 2B (similar-ish specs), you’d be surprised. If you don’t need anything fast or fancy, that 1GB will go a long way, and plenty of selfhosted apps require very little CPU. The only real problem I faced was that all HTTPS-related network tasks were limited at ~3MB/s, as that is how fast my pi could encrypt the data (presumably, I just saw my webserver utilising the entire CPU and figured this was the most likely explanation)
I’m currently hosting like 5 vms on a proxmox host (mostly ubuntu vms- pihole, nextcloud, home assistant, etc), which is an i5 4590 with 32 gb ram and I’m running up against the limits of how much ram I can provision and if 2 or more of my vms are doing something intensive at the same time I’m pinning the CPU. I don’t think my use-case is that crazy for someone doing a little self-hosting.
Luxury! My homeserver has an i5 3470 with 6GB or RAM (yes, it’s a cursed 4+2 setup)! </badMontyPythonReference>
Interesting, I also run Nextcloud and pihole, and vaultwarden, jellyfin, paperless-ngx, gitea, vscode-server and a minecraft server (every now and then).
You’re right that such a system really does show its age, but only when doing multiple intensive tasks at the same time. I try not to backup my photos to Nextcloud while running minecraft, for example, as the imagine identification task pins my CPU at 100%. So yes, I agree, you’re probably not doing anything out of the ordinary on your setup.
The point I was trying to make still stands though, as that pi 2B could run more than I would’ve expected beforehand. I believe it once even ran jellyfin, a simple file server, samba, and a webserver with a simple HTML website. Jellyfin worked just fine, as long as the pi didn’t have to transcode (never got hardware transcoding to work).
It is funny that you should run out of memory, seeing as everything fits (albeit, just barely) on my machine in 1/5 the memory. Would de overhead of running VM’s account for such a large difference?
I’m running the recognize app on nextcloud which I think requires at least 4-5 GB RAM, so I have 6 dedicated to that VM. I’m pretty sure the recommendation for Ubuntu in general was 2 GB RAM so I gave my pihole half that. Home assistant wanted 4 GB, but I gave it 2. I think my Jellyfin server has like 6 and I have another VM with like 4. So that’s a total of like 19gb RAM provisioned. Plus I have a 2 TB zfs pool for my nextcloud VM. When I go into proxmox it tells me I’m using like 29.5 GB.
I suspect if someone was using docker or some other sort of containerization one could expect better performance than what I am getting with VMs.
It was ok at the time, and if it isn’t ok now, that means you want to run something that is too bloated for its own good.
Really though, special hardware for this doesn’t make too much sense. A raspberry pi with two ethernet interfaces would be great, but if you can live with ethernet plus wifi, the current rpi’s will do it. Otherwise there are lots of similar boards that really do have two ethernet.
I have not really felt much use for self hosted server hardware at home. I use VPS’s for that and it’s less hassle. Maybe it doesn’t count as completely self hosted, but conceptually it’s a miniature colo box.
Was my first impulse too, but looking at their app selection now, it seems kind of … inutile? Unsexy? Old?
I admire the thought of lowering the barrier to entry to start self-hosting for “normies”. Not sure where you are located, but where I am, this price point is not realistic even for used equipment, not including RAM or storage. I’m not really sure what value add you are bringing to the table that one wouldn’t get from just buying used hardware from an office surplus and if one is very inexperienced in self-hostong, looking into something like LTT is partnered with like Hexos.
I’m doing experiments currently on a refurbished Intel i5-6500 with 8Gb DDR4 and a 0.5Tb SSD. It’s tiny, quiet (~45 decibels) and so far runs ~8 watts idle, 25 watts normal usage. I haven’t stress-tested the power draw. The router I’m testing with is a Mikrotik hEX lite 5. That’s around ~$150, though clearly if you are accustomed to more “rack-mount” style homelab these will seem very modest.
What I’m testing for now is getting representative loads on the devices to see how they perform.
Oh, I totally agree, my value add just isn’t there if you are experienced at hosting. The value add is to help people get started, and to keep them running at a modest level. Not everyone wants to experiment with Kubernetes at home or train LLMs. Some folks just want a password manager, a shared calendar, something to organize their tax documents, a pihole, and a Minecraft server for their kids.
I don’t follow LTT, I was under the impression it was more hardware reviews for the experienced than tutorials to help people get started.
I’ve read a bit about Hexos, I’m thinking of some similar things, and it would make sense to work with them. I’m excited for their coming beta.
I recently upgraded my homelab/self-hosting server from an old Dell T410 with dual X5650’s (2 - 6 core/12 thread CPU) and 24 GB ram to an old Dell Optiplex (7020 I think) with an i5-4590 (4 core/4 thread) and 32 GB ram. Its barely enough for a proxmox host with 5 VMs; but way faster than the old T410.
If you are offering some sort of self-hosting box, would it be bundled with some sort of software for someone to easily spin up whatever services they want?
Are you going to be able to make money at the $150 mark with all this hardware and configuration? If you are targeting people who are new to self-hosting, it will need to be a complete package (will need to have ram and storage installed).
A small home media server running off a raspberry pi could be that cheap.
Why would I need a separate router for that? I’d need to configure the main router anyway.
I would absolutely want the extra router because most people have one from their service provider. For self hosting, you want an additional router with your own software.
Market to tax funded institutions. If you can market “self hosted” as cheaper and easier than mother solutions you’ll have guaranteed clients for a long time.
That’s an interesting idea I hadn’t thought much about. I’ve been more focused on individuals than organizations. Do you have experience with tax-funded institutions? I assumed they generally have strict procurement rules and long support contracts with large established players by policy.
Their procurement policy is basically “has it been recommended? Is anyone else using it? Is it cheap?”
I work in public sector.
Oooohhhhh boy. Another one of these 🤣
It’s not like a package thing you can sell if you’re not supporting it. Then you’re just selling hardware at an inflated price. It’s not even self-hosting at that point. Why wouldn’t you just pay a regular company for a product?
Good point, I should have mentioned the plan is to sell support.
Then I’m not sure what the product you’re selling is though. Tech Support? That’s going to be a hard sell.
What is the aim? People who want to get into it, but does not know how, or experts? Think half of the attraction of selfhosting is the diy aspect.
What extra would this bring if people can just buy the parts cheaper?
And for those who only want the out of the box experience why would this be better than, let’s say a beestation? (Yeah price, I know, but you obviously would not have the same support level.)
I don’t disagree, and I would imagine what I’m offering would only be useful to people who are very early and don’t yet know they enjoy the DIY aspect.
The aim, though, is this: I’ve enjoyed self-hosting. It’s given me some powers that most people don’t get to have who aren’t also technical professionals. I’m also deeply frustrated by the environment created by the various major tech companies. If I can, I’d like to lower the barrier for people to get some of those powers without having to become experts and to make it more feasible for them to do what they want to do, rather than just what they are permitted to do.
Much shorter time going from “how can I control some of my own data” to "I’m running NextCloud, and its kinda like iCloud/Google Drive/Whatever Microsoft does and it’s running right here under my control! Not everyone knows the path from buying parts online to having a working reverse-proxy and reasonable firewall rules. Also, standardization makes it much easier to support people, which is really what the business would be doing.
I knew about Synology, but as a NAS product, which assumes a certain familiarity with backup schemes, etc. Kind of a prosumer-only thing. The Beestation is new to me, thanks for the tip. Quite possible what I’m proposing would have some overlap and compete with it, I’ll have to read up on it.
Needs serious market research to not flop out of the box.
No. People who want the benefit of self housing without worrying about hardware will rent a vps or something simpler. The hard part of hardware isn’t the purchase, it’s the maintenance.
Also, why the separate router?
That’s certainly an option. I think of dedicated hardware as working for several different people, some of which care a great deal about not using a VPS provider because they don’t trust them with their data, or don’t trust them to be around for a long time, or don’t trust them not to raise the prices.
I’m inclined to agree, but I’ve been doing hardware for a long time as a hobbyist and I sometimes forget how far I’ve come. It sounds like you might be somewhat like me in that regard. I’m often surprised when people see assembling system parts and flashing an OS as a complex, inscrutable task.
What do you see as the hard part of maintenance? Scheduling time to do it? Unexpected errors or failures?
I agree with this. Self-hosting requires the user to understand their network, their software, how it all interacts.
If you provide a hardware product and call it a solution, people are going to expect a turn-key solution like a plug-and-play router.
You’re going to end up supporting a bunch of newbies who, by no fault of their own, can’t tell you an error code in the console let alone whatever UI you give them.
I think a better solution would be a course that walks newbies through self hosting.
I think this needs to exist, but as a community supported system, not as a commercial product.
Pick a set of open technologies - but not the best, lightest weight, just pick something open.
Come up with a security architecture that’s reasonably safe and only adds a moderate amount of extra annoyance, and build out a really generic “self-hosted web hosting and VM company-like thingy” system people can rally around.
Biggest threat to this, I think, is that this isn’t the 90s and early 2000s any longer, and for a big project like this, most of the oxygen has been sucked out already by free commercial offerings like Facebook. The technical family friend offering to self-host email or forums or chat no longer gets gratitude and love, they get “why not Facebook?”
So… small group effort, resistant to bad actors joining the project to kill it, producing a good design with reasonably safe security architecture, that people can install step by step, and have fun using while they build and learn it.
I think a possibility is a series of open source anvil or nixos scripts that you can run on most hardware with minimal changes, in an extendable architecture of some kind to add or remove functionality and they perhaps get maintained by the community or some structure of the kind of Linux distributions.
This could enable people with minimal skills set up and maintain a reasonably useful but secure environment just by changing a few variables.
What’s a nixos script?
Nixos is an os that’s defined by its config stored in .nix files. Everything is defined here all the software and configurations. Two people with the same script will have the exact same os.
Any changes you make that aren’t in the scripts won’t be present when you reboot.
You could maintain a very custom linux distribution (kinda) by just maintaining these config scripts.
So a user wouldn’t need to install all required software and dependencies. They could get a nixos and the self-host config and adjust some settings and have a working system straight after install.
A viable alternative is Guix, which uses Scheme for its scripts and could also use the Hurd kernel instead of Linux, but works the same.
I think this is a great point, it doesn’t help much to create a business that ends up with the same incentives and the same end-game as the existing systems.
That is precisely what I’m looking to build. I don’t want to get rich, I want people without 10 years of industry experience to get some of the benefits we have all been able to build for ourselves.
We already have that, the first problem is we have like a dozen of them, a few are even well supported. The second problem is that usually the technical knowledge required to set up the systems are still lower than the technical knowledge required to keep it running.
I’ve been struggling to wrap my head around a good security architecture for my mspencer.net replacement crap. Could I bug you for links?
I figured out a while ago to keep VM host management on a management VLAN, and I put each service VM on its own VLAN with heavy, service-specific firewalling and a private OS update repo mirror - but after hearing about ESXi jackpotting vulns and Broadcom shenanigans, I’ve gotten really disheartened. I’d love some safe defaults.
It sounds like you’re getting into the keeping it running phase.
First, going back to your previous comment, self-hosting email is difficult. It’s not hard for a small provider to end up blacklisted and you’re probably kind of just done at that point and it will feel very unfair. I get that it’s a fun set of technical challenges, but you couldn’t pay me enough to help someone self-host email.
Second, guessing, but it sounds like you may be trying to expose your services directly and doing a lot to make that work which goes against what most would recommend for hosting your own services. Big companies don’t expose their intranet like that, follow their example. Almost every guide or system is going to warn against that. If you’re going to host more than one thing, highly recommend focusing on minimizing entry points and looking into a VPN-like solution for accessing most if not all of your services. Still spend time on securing your intranet, but most of your risk is going to come from how hard it is for people to get past the front door (or doors).
Thank you for your reply, but to be clear, I’m not looking for individual details to be spelled out in comments. What you said is absolutely correct, thoughtful, and very helpful. But emotions are running a little high and I’m worried I’ll accidentally lash out at someone for helping. Apologies in advance.
But do you have any links? Beyond just the general subjects of security architecture, secure design, threat modeling, and attack surface identification, I’d love to see this hypothetical “generic VM and web application housing provider in a box” come with a reasonably secure default architecture. Not what you’re running, but how you’re running it.
Like, imagine decades in the future, internet historians uncover documentation and backups from a successful generic hosting company. They don’t necessarily care what their customers are hosting, their job is to make sure a breach in one customer’s stuff doesn’t impact any other customer. The documentation describes what policies and practices they used for networking, storage, compute, etc. They paid some expensive employees to come up with this and maintain it, it was their competitive advantage, so they guarded it jealously.
I’d want to see that, but (a) a public, community project and (b) now, while it’s still useful and relevant to emulate it in one’s own homelab.
If I can get some of that sweet, sweet dopamine from others liking the idea and wishing for my success, maybe I can build my own first version of it, publish my flawed version, and it can get feedback.
The tech savvy will just buy a Raspberry Pi and install yunohost on it.