I'm considering rewriting all of the SparkShell backend for some reasons:

1. The code is spaghetti

2. Using JSON databases is not optimal. I don't really want to migrate 650+ lines of server code just for that reason.

3. SeaweedFS. Looks promising and could work great. Thanks @tnix

Note that most of the code will be reused, just reorganized, changed, and now using an AWS MongoDB cluster in Oregon (yep, its the closest one to me lol )