It's interesting see people still working in this niche, I think it used to be called "backend as a service"? There were several contenders at one point, but Firebase seemed to "win" several years ago, and has been the last one standing for some time.
Just curious what the market size is for these types of services? It doesn't seem like anyone really uses them; for tech people it's just as easy to spin up an instance in some cloud somewhere. For non technical people, well these types of services still seem to be too complex.
I think the one thing Firebase has going for it, that is a differentiator is it's mobile push service that pushes to both iOS and Android. There aren't many options there.
Yea Nhost is a Backend as a Service (BaaS) but I'm not sure Firebase has "won".
Whant we hear form our users is that they like Nhost because of:
- Open Source
- SQL (Postgres)
- GraphQL (Hasura)
- GitHub Integratoin (like Vercel & Netlify)
> for tech people it's just as easy to spin up an instance in some cloud somewhere.
I think this is the biggest missconseption.
Just look at Vercel and Netlify. They host html/css/js as a core business and they have million of users combined. Why? Because they make it extreamly easy for developers to be *productive* and they hook into developers' workflow with Git.
Most of our users at Nhost can spin up servers and infrastructure on their own, but are they using their time optimally by doing so? Or should they instead spend time building what's important for their business and users?
>Yea Nhost is a Backend as a Service (BaaS) but I'm not sure Firebase has "won".
Firebase is a nightmare since they folded it into GCP. As a standalone product it was complete, easy to grok, and simple to secure. Now I feel like I literally need to hire a CISSP to make sure my whole company wont be hosed by a single security rule misconfiguration buried somewhere in GCP.
Perhaps Forebase was the first one to do it right, the biggest features were Real-Time and as you've mentioned, Push notifications. Everyone else has been attempting to compete. The big win for us is that many of these solutions are open source. Shout-out to Supabase and shout-out to Postgresql which allows for this to be possible.
Firebase at it's core is just Google Cloud, same with AWS Amplify just being various AWS services neatly delivered. I imagine some people are using the exact same services just through Google Cloud and AWS. Cloud Messaging == SNS, Lambda == Cloud Functions, Firestore == DynamoDB etc.
I think you might be underestimating their reach though, they're very popular in app development shops, since an app developer doesn't need to know how to build a serverside application in order to get an app that needs remote services working.
I’m primarily a frontend developer. The task i hate most is dev-ops like stuff and dealing with Aws in any capacity. If i was to create a start up I would 100% use something like this or Supabase.
The mobile push service is indeed a great feature from Firebase, which we will also implement when possible. We already provide a realtime API through GraphQL subscriptions.
There is a lot of movement indeed in this space. Here are a few things that I think we at Nhost are doing well:
- Open Source: Nhost is open source. Some competitors are, too, and some are not.
- Infrastructure: Our infrastructure is set up to handle anything from side-projects to companies with hyper-scale. Each service, except the Postgres database, is stateless and can scale horizontally.
- UI: The Nhost Dashboard, although under development, will give us an edge in the next six months. Most database and backend dashboards look like they were built by engineers (without designers) in 1997. There is so much room for innovation.
- Tech: We're using well-used and open source technologies with little to no abstractions in our stack, from SQL (Postgres) to GraphQL (Hasura), to Node.js for Serverless Functions. Here I see competitors using either old technologies or building abstractions, forcing them to re-invent already solved problems, and forcing users to re-learn already battle-tested solutions.
- Workflow: We're not just an "infrastructure company." The goal with Nhost is to make developers as productive as possible. We have a CLI for local development and a GitHub integration that automatically deploys your backend (just like Netlify & Vercel). This way, our customers can focus on their business and users instead of figuring out development workflows. Here I see competitors either building custom in-house branching solutions outside Git or nothing.
In the end, I'm glad to see competition in this space. The winner of healthy competition is developers who can pick the services that best suit their needs. And that's the ultimate goal. To empower developers to build.
If I had to guess, they are not adverse to "Docker" the company but actually to immutable container images that are shipped from the developer and often assume that they can run as root.
Hasura GraphQL Engine is one of the open source tools we're using for the Nhost stack. But it's "only" one of the open source technologies we're using.
Just paying Hasura gets you just Hasura (at least before Neon it used to)
NHost gets you Hasura, the underlying postgres db, pre-configured auth with JWTs, blob storage, functions and a client-side library to get all of that working together with nearly zero configuration.
edit: also, Hasura's cheapest offering ($100 / month) is still prohibitively expensive for indie projects (That's just Hasura - you would have to add database costs, auth costs and blob storage costs to that $100).
Just to add to the comment: you also have backups, automatic logs on all services, raw access to postgres (you can do whatever you want with it), custom integration with stripe, plus specific dashboard for users, database, and storage.
Was it tested, by chance, to run with Cockroachdb or Yugabytedb? We are migrating off non distributed solutions and the Postgres native solutions we tried just weren't easy to set up and use, while the two newcomers are. Any advise in that direction? We are using Yugabyte mostly now for testing if everything works and if it's robust; so far it works very well -> we were very happy with Cockroach, but not with the license.
Hasura CEO here. If you're trying or evaluating using hasura (or nhost) along with Yugabyte / Cockroach feel free to reach out (email in bio). Happy to have our team help with any architecture or benchmarking things you might run into!
Storage is using S3. We use AWS S3 at Nhost, but you can use Minio (open source) if you self-host.
But the good thing with Nhost is that it's open source, so if you want to implement something other than S3 as the core storage layer, you can do so: https://github.com/nhost/hasura-storage
Most of the users I've seen using Nhost (and Hasura) treat data/the db as the source. Given that you have raw access to postgres, you are also the full owner of your data and you can swap interfaces as you see fit (I've seen users use Hasura internally and not expose it to the internet)
Notice, also, that Hasura has support for REST endpoints.
I _just_ started using Firebase and I'm already worried about cost. Is this a drop in replacement for it ? All my code will work against nhost , just change the config files ? Or what does the transition look like at the moment ?
Better to migrate to something open-source without vendor lock-in while small. If you need any help to understand what would be needed, feel free to email me at nuno@nhost.io.
Just curious what the market size is for these types of services? It doesn't seem like anyone really uses them; for tech people it's just as easy to spin up an instance in some cloud somewhere. For non technical people, well these types of services still seem to be too complex.
I think the one thing Firebase has going for it, that is a differentiator is it's mobile push service that pushes to both iOS and Android. There aren't many options there.