HazelBase

This is going to be a polyglot backend service similar to Firebase/Parse/Auth0/Okta/AWS Cognito...

Pitch

Problem

You need to get started with an app (web/mobile/...); but first you need to lay the groundwork for "users" before implementing anything app specific.

So the dreaded "user" table in your database and all the basic/boring functionality around that.

Why is it a problem

Consider what that "basic" functionality is:

• Register users
  • send verification sms's
  • send welcome emails
    • now you need "workers" (as you can't do these in request/response cycle)
      • your "workers" solution will tell you to run a "message broker" (hello Redis/RabbitMQ)
• Authenticate users
  • Lots of tight secure code. You can't screw up here.
• Lots of other trivia
  • Authorization (permissions) for user types
  • "Forgot my password" and friends
  • UI's for administration of users
  • Social logins. "Login via Instagram/Facebook". Good luck with OAuth.
  • Importing user's from an older system
  • ... I can go on and on ...
• All of those in a performant way, (reads "cached").

What you can do:

1) Do all those yourself: costs months, even up to 1 man-year.

• LOL no.

2) Use one of your popular frameworks of your ecosystem.

• Ready made user models and most of the functionality above.

• Weeks to configure.

• Exactly why RoR, Django, Node/Express... saw explosive growth in late 2000's - early 2010's.

• Still not fast enough for this day and age.

3) Buy ready made app templates from the likes of CodeCanyon.

• funny but true.

• very active ecosystem on app templates all around the web.

4) Use a hosted solution (Firebase/Parse/Auth0/Okta/AWS Cognito)

• Cost hours/days (if your language is supported)
• Younger generation's favorite AFAICS.

• Expensive after a threshold
• Platform lock-in to death. once you're in you're in.
• All closed source; no way to self host.
• Not much room for customization.

• Huge market; A few months ago Okta bought Auth0 for 6.5 billion usd.
• Firebase still haven't been abondoned by Google.

5) aaaaaaand: HazelBase?

• Hazelcast Platform already provides most of the moving/crucial parts of a typical Backend app:
  • caching
  • it's a message broker
  • great IPC
  • clients for popular languages.
  • all sorts of integrations (S3, other databases, etc..)
  • all this in a simple to deploy package (nothing beats "java -jar" (well, except a Go binary))) and performant.
  • HAZELCAST IS THE BACKBONE FOR YOUR WHOLE BACKEND.

Sitting on top all that; why don't we use it:

1. Slap a webserver in front of Hazelcast Platform.
2. Implement func. above;
3. Let the developer use his preferred language/framework

• communicate with Hazelcast via the clients or HTTP.

4. ... call this HazelBase?
5. profit?

• This IMO is the killer use case for Hazelcast Platform.
• It could help a lot for brand awareness.
  • (Sales team would appreciate that a lot; RedisLabs sales team must have a much easier life.)
  • Especially if it reaches to the younger generation. (xxx emphasis on that later)
• Hosted offering under Hazelcast Cloud umbrella.

Emphasis on younger generation.

xxx

xxxxxxxxxxxxxxxxx

Intended audience

• Serial freelancers/small teams who copy/paste from their old Django/Rails/Node projects to hit the ground running on their new projects.
• Developers who want to get rid of depending on Firebase, Parse, Auth0, AWS Cognito, Okta etc...
• Younger generation who want to spend as little time as possible on backend work.

Technical Details

• The core will be a Java/Clojure web server.
• Hazelcast will provide the caching and task execution and message queue layer.
• Persistence layer will be Postresql or SQLite.