Storing the meta data in DB

[atapas]

Currently, we are storing the meta data in a JSON file. How about persisting it into a DB?

Which DB? Let's discuss.

[atapas]

What are the DB Choices? Here are a few to list down,

• HarperDB
• PlanetScale
• Mongo Atlas
• Fauna
• Anything else?

[hasnainmakada-99]

Tapas Sir, I would like to add Firebase

[vasantisuthar]

Which schema structure should be implemented NoSQL or SQL?

[atapas]

Good Question. If we see from the requirements, we need,

• Store the play meta data
• Registered users
• Likes for plays
• Comments and it can be hierarchical.

I can thinks of these from the storage requirements. What do you think would be suitable? I think NoSQL.

[vasantisuthar]

Yes, NoSQL would be a better option. Also, like the comments and likes for play should be in real-time, so can we consider using RethinkDB, it is a database that provides real-time updates to the application.

[atapas]

I love RethinkDB. It is a great real-time database and used extensively in my past projects.

I wouldn't consider RethinkDB for the following reasons. Let me know if you agree.

• There is NO cloud offering of RethinkDB. We need to host and maintain it on Heroku or AWS like PAAS.
• The development of RethinkDB is slow over the last 4-5 years. Also, there was a notice: https://rethinkdb.com/blog/rethinkdb-shutdown/
• We do not have a great need for Real-Time use cases.

What say?

[vasantisuthar]

Surely the reasons are true, we cannot consider RethinkDB now.
And I didn't know much about HarperDB or PlanetScale but how about Firebase?
It provides a NoSQL cloud storage service

What do you think?

[atapas]

Surely worth looking.

[Abhishek-90]

I will be in favor of using Mongo Atlas or NoSQL database, for 2 reason, correct me if I am wrong.

1. For maintaining likes for Plays, using NoSQL schema will help us store details of users who have liked a particular play
2. JSON format storing of Data in MongoDB will also help us to maintain hierarchy

[atapas]

@Abhishek-90 Agree. Both satisfy the criteria we are looking for. What are your thoughts about the plan and price? Do you think, we will scale there too if we go for it?

[hasnainmakada-99]

I am going to put a table ASAP, which will show us the pros and cons of every database we discussed over here, so will be good for us to ultimately think which DB we should consider to integrate

[hasnainmakada-99]

ok sir

[hasnainmakada-99]

Tapas sir I've made tables which include 5 databases, I am gonna share it with you now

[hasnainmakada-99]

ReactPlay.io

List of Databases

Harper DB
Planet Scale
Mongo Atlas
Fauna
Firebase

Pros & Cons of every database

Harper DB

Pros	Cons	Cost	Harper DB Cloud
Easy to get started	Difficult to integrate with flask applications (As we are building on react it doesn’t matter to use)	The Cost of Harper DB Cloud...	100% Customizable & Easy Deployment
Good Experience as per testing from the free trial	If coming from a SQL background, will be difficult to adapt Quickly		1GB Free, After then 10GB at 2.5$/mo (190rs)
Customer/Tech Support is good compare to other companies	Less storage only 1gb is provided and only one instance allowed		Then 250GB at 62.50$/mo (4750rs)
			Then 1TB at 250$/mo (19000rs)

Planet Scale

Pros	Cons	Cost
Fully managed cloud native database built using Vitess and kubernetes, can access from anywhere using various cloud providers	Didn’t found out any flaws with Planet Scale	Planet Scale Developer (Free), limit one database per organization	Planet Scale Scalar	Planet Scale Enterprise
used by many big tech companies and small scale startups		10 Gb Storage/Mo	29$/mo (2200rs), perdatabase + usage cost	Custom pricing (to contact sales)
Has many alternatives such as amazon aurora, amazon RDS, google cloud SQL		10 million row writes/mo	25Gb storage & 100 bilion row reads/mo	unlimited database branches
Takes daily backup of databases		Daily backups & Community support	Auto Scaling	premium support

Mongo Atlas

Pros	Cons	Cost
Performance level is good compare to others, as it allows specialized reads where we can map our data	Does have responsive management, we can not open the database from mobile & doesn’t works well with search parameters	MongoDB (Serverless)	MongoDB (Dedicated)
Possible to access JSON from anywhere without any language specialization	If entries go more than 150, we have to use the mongo shell platform & visualize the data	For serverless applications with low and variable traffic	Built for production applications with sophisticated workload requirements
Provides accurate information to users & takes backup of databases time to time	when working with clusters, it is not possible to download a direct copy of the information	Upto 1Tb of Storage
		Starts from 0.30$/month (23rs) with millions reads	Starts from 57$/month (380rs), comes with 2 gb of ram

Fauna DB

Pros	Cons	Cost
Supports Both FQL (Fauna Query Language) & GraphQL	To write security rules we have to compulsory use GraphQL	Fauna DB Team	Fauna DB Business
FQL can be used to create databases, indexes, tables and many more in fauna	Consumed quite a time to get addapted to its domain specific FQL	135$/month (10,300rs)	450$/month (34,200rs)
Can be used as an external DB which we dont want to integrate in our app/website	Doesn’t allow to export the queries in JSON or CSV	Tech Support 24 business hours response time	same features as of team plan
Fast & Friendly Database		Unlimited Capacity & Multi factor authentication	Support provided within 8 business hours

Firebase by google

Pros	Cons	Cost (looking from the cloud side of firebase)
Provides all the database capabilities	Provide limited querying capabilities and as a result remain behind compare to other DB’s	Provides 100Gb of cloud storage & and 200 Gb of hi-res photos transferred at 23$/month (1800rs)
Provides a wide range of services and features	Provides limited data migration	If we want realtime database of 20GB then we have to pay 95$ extra (23+95= 118$) (9000rs (around))
The Basic plan is Free	Is platform dependent
Quick and easy to integrate and setup	Mainly android centered
Documentation is Concise and and provides every thing	And provides less support for iOS

Do Tell me if there are any mistakes in here

[atapas]

Thanks a lot, man, much appreciate it. Let me take a deep look at it and come back.

[hasnainmakada-99]

ok sir

[atapas]

I am inclined to select one between HarperDB and Firebase. More updates to follow.

[hasnainmakada-99]

Then sir you should go with harper db, as it is better than firebase

[atapas]

Tonight I'll be talking to people from HarperDB to understand the support part for our project. Let's see how that goes 😀

[atapas]

We are concluding the HarperDB as the database to go ahead with.

Advantages:

• Cloud-based instance
• Both SQL and NoSQL Supports
• Custom Functions
• Past experiences of using it.
• We will get some advantages with the plans, quota, and bandwidth(as per my discussion).

Thank You every one participating in the discussion of concluding it. I shall create relevant issues to,

• Configure the database instance.
• Get the existing play meta data into the DB
• Create a workflow to update the DB for new play meta data as part of the build process.