Consult the design concept behind ZestDB

[bofu]

Hi Databox Team,

I played with ZestDB and it’s really easy-to-use and has quick response time, thank you for the great works!

To understand the design concept behind (e.g. what are the problems ZestDB can do better than PostgreSQL and Mongo), I visited ZestDB website and GitHub doc/wiki, and Tosh also encouraged me to initiate a discussion here.

If you can share more design/technical documents or concepts, it will be very helpful for a novice like me to understand ZestDB more.

Thank you!

[Tosh]

@jptmoore @mort are probably the best team members to address this question.

There is an overview if zest in this paper https://arxiv.org/abs/1902.07009 which may be of value.

Cheers

Tosh

[mort]

The paper @Tosh references is a good place to start.
I think the key motivations were:

• good support for timeseries data
• low memory footprint, reasonably efficient (to run on an RPi)
• support for the data access control mechanisms in Databox (ie., macaroons covering paths)
• amenable to being integrated with the other IO control mechanisms in Databox (ie., network connectivity)
• support for provenance data collection for audit purposes (data just about now available though we haven’t produced any audit mechanisms yet, and there’s still a bit of work to do to finish integrating the latest logging mechanisms into Databox)

If anyone else wants to speak up to add to that (or to correct me!) feel free
Hope that helps!