Hi Redis community.
In the Pro version of Redis you provide two features that I find essential when we have a lot of data. The first one is Active-Active Replication based on CRDTs. The second is Redis on flash ROF. And eventually a third and last one which would be a Multiple Master feature to have more than one write node server.
I also saw that KeyDB (a fork of Redis) took the initial step of offering these features in OpenSource. However, I would like to stay on Redis and not switch to KeyDB. So my question is: Can you please integrate these really important features in a future OpenSource version of Redis ? If yes, when? If not, why ?
Thank you.
Comment From: madolson
There is no active plans to implement the features that you mentioned. Also worth noting that Redis Enterprise is maintained by Redis ltd, while this version is maintained by the Redis community.
Comment From: helple
Yes, this version is the open source version of the redis community. Anyway.
If you use redis and if you really look into the subject and go to the different discussions on forums and Discord chat. You will see that these features are in high demand. This is why I would like us to organize a community vote on whether or not these features should be developed and implemented.
Thank you.
Comment From: oranagra
These complex features are certainly out of the scope for the OSS Redis project for many reasons. One of them is that one of the main guidelines is to resist complexity. Maybe Redis Ltd will some day decide to change the license of their products or make them free, but I'm certainly not the person to discuss that.
Note that a community vote on the subject is pointless unless you have the time / budget to develop these. i.e. what you're asking for is a huge project.
Comment From: superhero
I just like to say, if anyone in the future is looking to evolve a similar solution to Redis on Flash for this community edition, then you can ask me to get involved.
I think the ability to separate hot vs cold data is extra important for us who use redis as a message queue, but I guess there is an endless reasoning to why the ability to archive key values to a cheaper media is of interest for the product.
I don't really see an issue with why we could not have a garbage collection alike functionality... but whatever... I don't mean to argue anything, just to show interest in helping if anyone else is willing to take the time to evolve the functionality.
...meanwhile I go pay Enterprise for this functionality