I had hosting issues in many of my companies and consulting projects, running from "WTF is this AWS bill $10k???" to "WTF did our 10gbit/s download just kick Heroku offline?"
But people can only appreciate the things I build for them if they are hosted well. So I took the obvious next step and purchased a small struggling hosting company. We then introduced nicer tooling that imitates the AWS and Heroku APIs if possible, to make using their servers as easy as possible for me. And now I can offer managed hosting to all of my consulting clients. For them, it's slightly cheaper prices and (in contrast with AWS and Heroku) they can just call me when things don't work. And for me, it's reliable high-margin revenue.
I'm going to save this so I can just link to it whenever the "but running your own servers is irresponsible!" thing comes up (as it seems to do weekly here on HN).
Cool idea!
I'm a pro SRE guy, with 20+ years of experience and working with US top500 companies. If anyone wants to partner up and replicate this, let me know! You can also reach me on matrix: @kim0:halogen.city
Great. If you don't mind, can you share your trade-offs and pain-points of managing own hosting? and How do you manage to get your service/apis compatible with forever(kinda) changing AWS/Heroku apis.
I never had the "hands off" hosting experience that cloud providers advertise in the first place. Before (with AWS and Heroku) if things went down, I was the person that had to investigate and escalate to the appropriate vendor's support. Now, I'm the person responsible for investigating and fixing things. For many small issues, it's easier to fix them than to get someone at AWS with the necessary clearance on the phone.
One big trade-off is that I only have one datacenter so in cloud lingo, I only offer one availability zone with offsite backups. But then again, when AWS us-east went offline, so went most companies, because the proudly advertised multi-AZ failover didn't work that well in practice.
As for the APIs, I just limit the subset of APIs that I offer. Since this is used only for my consulting clients, my cloud only needs to support those APIs that they require. In effect, that means CEPH+PostgreSQL do most of the heavy lifting.
I've seen Cloud APIs explode in funny ways when customers actually arrive with Petabyte-scale datasets. In that sense, my customers are pre-qualified because all of them have had horrible experiences with Cloud providers over-promising and under-delivering. So once they are on my platform, dissuading them from wanting more APIs is very easy. "Would you like to continue to use my cloud at the current price? Or would you like to pay 3x the monthly spend just to use the Amazon cloud and gain API call XY? And please remember the last time they crashed badly and you could not reach any Amazon support by phone. Didn't they even bill you for EC2 while it was unreachable?"
And if they insist, I can always point them towards the Open Source project that Amazon/Heroku use under the hood. But in my opinion, most Cloud APIs don't add business value anyway. "Keep it simple" always wins out in the long run.
> I never had the "hands off" hosting experience that cloud providers advertise in the first place.
We've had exactly the same experience. From a business perspective it does mean that you can shift the blame if something goes down but it still impacts your product all the same. And you're left without the ability to directly fix the underlying issue. Most of our infrastructure runs on our own hardware out of a single datacenter and we've been able to get significantly better performance/ value. I love your approach of commoditizing your complement and acquiring a hosting provider.
Point of clarification: us-east-1 is an entire AWS Region, which contains multiple availability zones. Multi-AZ setups in AWS are generally fairly straightforward, multi-region generally isn't.
...that said, one of the key techniques most people use for failing over between regions is changing DNS, which wouldn't have worked with Route 53 in this instance because its control plane is in us-east-1, so it's not like AWS looks spotless even in this analysis.
Judging from your username, you're affiliated with rsync.net. If that is correct, then no. That's a very interesting service for offsite backups, though. At the moment, I just rent colocation with lots of HDDs inside and do the RAID management myself.
> How can we work with you ?
That depends on what you need and how long you can wait. I've already committed to launching 2 AI-based products in 2022, so currently it looks like I'll be fully busy until end of the year.
But most of my customers used RDS and are now on MySql, MariaDb, or PostgreSql. Similarly, S3 -> CEPH. Nginx as front-end, Varnish as cache, Ha-Proxy for load balancing. Queues? RabbitMQ. Map-Reduce? Hadoop. Lots of data? ZFS.
There are A LOT of great open source projects to base your infrastructure on.
My customers want API compatibility with AWS and convenient deployments like on Heroku. So I needed a software layer to convert bare metal boxes into Cloud VMs and to imitate necessary APIs. And of course base images, buildpacks, etc. to convert a git link into a runnable docker image, like what Heroku does.
So basically I purchased them for all the scripts, tools, and stuff that happens between "Customer does a git push" and "Docker images are running on the correct bare-metal servers with the correct configuration"
Yes and no. Quite a lot of my customers have designed their systems around separation of concerns, so they need 100 small services running. By having my own layer in the middle, I can group those 100 service VMs onto 10 beefy physical servers, as opposed to actually needing 100 bare-metal servers to run their 100 services. But in the end, yes, I have roughly the same responsibility and much better margins.
Buy my startup @ https://PretzelBox.cc. Jokes aside, PretzelBox is a hosting company aka backend in a box and your comments are giving me a ton of hope.
And I thought to myself "So it's a well-marketed notion whitelabel reseller." You're probably good at marketing, but I'm not sure what Cloud back-end infrastructure IP you would have for sale...
No nonononono, I don’t have any idea how to make or sell Notion templates. This is an AWS backed storage and hosting solution through and through.
Your emails use SES for send/receive, files are stored in S3, auth is managed using cognito, the blog that users present to their site visitors are a Svelte SPA.
(The blog itself is a WIP so im using Notion to host my own content for now)
Think of it as a single replacement for Wordpress, dropbox/wetransfer, and gmail/gsuite.
The intellectual property is the working to provide a fully functioning business-y email inbox running on your custom domain, the code to host your own content at scale on your own domain, and code to secure access to private content.
So, your customers could use you for email management, use you to send and receive files, write blogs, and manage their team’s authentication and authorization.
I wish there were a simpler word to describe it but since it’s so many things, I went with “backend in a box”.
I’m probably using the term very loosely but PretzelBox can be used to host a website, store and share images and files (like Dropbox or Imgur or wetransfer) and integrates with AWS to give you an entire backend you can extend to run a whole business.
In my mind, I feel this could/should qualify as a hosting service
I'm going to amplify what others here have already mentioned.
I read your landing page, I have no clue what it is you're selling or what it costs. I'm busy, you were lucky few minutes of my time, in the few minutes I couldn't figure out why I should give you more time or money.
I would highly suggest you work on the landing page:
* What am I buying?
* What problem are you solving?
* What does it cost?
* How is your solution better than others?
My impressions on visiting PB:
Ah, a .cc domain, they're too underfunded to pay for a .com. I probably shouldn't use them for anything too important.
I do not want to "join your beta list". I want a service, now or never.
"PretzelBox" is my backend? Backend to what? You tell me not to worry about domains, email, storage... Does that mean PB provides those? Are you better than WordPress? Are you better than gsuite?
Then there are some random (and I'm guessing fake (rude, I'm sure, but why should I believe them?)) testimonials.
Then some arbitrary text: "See what people have twisted their PretzelBoxes into" with no links, description, or examples... So I'm assuming PB has been twisted into nothing?
Finally, after much scrolling and bullshit, a list of features. So I get some email, some storage, you say unlimited, but I doubt. Still no pricing info. I get some blogs, "No need to learn WordPress!" but... I have to learn PB? How hard is PB. WordPress at least has a ton of support and resources.
Usually this would be called a PaaS (Platform as a Service). It's another layer on top of a public cloud, AWS in this case.
Someone looking to save money on server costs by using a hosting company would effectively be going the other way in the stack: taking the cloud out of the equation and operating servers directly. The original poster in this thread acquired a hosting company to be able to provide this kind of service in their vertical.
Just to understand: did you buy any actual hardware?
Or "hosting company" means services on top of another hosting company that actually has hardware somewhere?
Long-term rental for colocation cages in an existing datacenter + long-term leasing contracts for the hardware + long-term support contract with a guy that lives close to the datacenter and keeps a stack of HDDs in his basement
So in effect I provide cloud services and cloud APIs on top of physical bare-metal servers that are maintained and housed by someone else.
I presume you mean site reliability engineer. Then yes, my clients purchase the whole package. Development, deployment, hosting, production maintenance. So if something goes wrong, really anything, I'm their first point of contact.
Before, we had many situations where they asked me to take a look because something could have been a software issue but then turned out to be an AWS issue instead. That's always nasty if I need to tell my client "sorry you're offline, try calling Amazon". Even if they understand that there's nothing I can do, emotionally they still feel disappointed in me. When I use my own infrastructure, I can avoid these situations because it's in my power to make the necessary hosting repairs, too.
I've always found it interesting how simply "Not Megacorporation X" can be a competitive advantage. It doesn't make a lot of economic sense, but people will go through a lot of work - and it can be lucrative to you on the business-side facilitating this - to avoid having to interface with a faceless conglomerate like Amazon.
For my customers, it does make A LOT of economic sense. The financial damage from being offline for a few hours and unable to reach any support person at your hosting company is very much real.
At their spending level, "Megacorporation X" would not pick up the phone for them. But I do. So in effect, my competitive advantage is not that I'm not a megacorp, but my competitive advantage is that I pick up the phone when they call.
I had hosting issues in many of my companies and consulting projects, running from "WTF is this AWS bill $10k???" to "WTF did our 10gbit/s download just kick Heroku offline?"
But people can only appreciate the things I build for them if they are hosted well. So I took the obvious next step and purchased a small struggling hosting company. We then introduced nicer tooling that imitates the AWS and Heroku APIs if possible, to make using their servers as easy as possible for me. And now I can offer managed hosting to all of my consulting clients. For them, it's slightly cheaper prices and (in contrast with AWS and Heroku) they can just call me when things don't work. And for me, it's reliable high-margin revenue.