Merging Mint into ReDBox - thoughts?

[Duncan Dickinson]

Hi all,

As part of some design work we're doing atm the question about bringing Mint into ReDBox has come up. Essentially this would mean that ReDBox would hold the information traditionally held in Mint and means that you'd no longer need to run a separate Mint.

No timeframes around this - I'm primarily wanting to know if anyone has a grave concern.

--

Cheers,


Duncan


Duncan Dickinson
QCIF Project Manager 
Central Queensland University

Contact me:

ph: 07 3138 2084
m: 0432 402 511
skype: de.dickinson

website | calendar | LinkedIn

[Christopher McAvaney]

G'day Duncan,

I presume that all the data and functionality Mint currently maintains would be replicated within the new ReDBox?
Is the aggregation due to performance of the loosely coupled Mint/RedBox separation?

Regards,
Christopher

[Jay van Schyndel]

Hi All,

An upgrade path would definitely be required. 

A process of merging existing Mint and ReDBox installations into a single storage system would be required.

Thanks,

            jay.

[Duncan Dickinson]

Hi all,

My main thinking around this is that ReDBox and Mint are basically the same platform with some custom functionality. Essentially a "merged" ReDBox-Mint would provide the functionality of both systems and start us heading towards a unified storage that then helps us consider a "ReDBox Core" that better deals with the storage and management of our objects. First step though would be to load Mint data into ReDBox and "trick" ReDBox into calling itself instead of Mint.

The performance shouldn't change greatly with this as it's still lookups from ReDBox forms into the data but other work we're doing around curation and harvesting (I'll send an emailabout this in the coming week) are part of a program to improve performance across the platform.

One key area that would be worth considering is the move of vocabularies (such as the FoR, SEO and Language codes) into their own, pre-canned Solr core. Essentially you wouldn't need to load these vocabs into the system, you'd get them ready to go as a Solr core. This is similar to what we did with GeoNames and really works on the idea that these vocabs change infrequently (if ever) and are really just used for lookups. The Mint side of things would then only relate to objects such as People, Groups, Service, Activities and these are objects that could be managed alongside Datasets & Documents within the combined ReDBox-Mint system.

I agree with Jay - we'd need a migration path around this. Noting, of course, that this thinking doesn't represent a concrete plan.

Lastly, why did Mint and ReDBox end up as 2 separate systems?

1. There was an early hope to setup Mint as a name cleaning/rationalising system that could tidy up the messy name management in institutional repositories
1. This didn't take off and we have some better models (not implementations) for this that mean we don't need to do this in a production Mint install
2. Data security - Mint could be locked down behind a firewall as it contains some private information (e.g. Staff IDs)
1. Ultimately we should make sure that this is secure via the platform and the separation and firewall currently makes it hard for things like handles etc to link to anything
3. Mint could be used as an institutional name/vocab lookup for use by other apps.
1. I reckon this is still possible under a combined platform 

Cheers,

Duncan

--
-- Website: http://www.redboxresearchdata.com.au
 
You received this message because you are subscribed to the Google Groups ReDBox group. To post to this group, send email to redbo...@googlegroups.com. To unsubscribe from this group, send email to redbox-repo...@googlegroups.com. For more options, visit this group at https://groups.google.com/d/forum/redbox-repo?hl=en
---
You received this message because you are subscribed to the Google Groups "ReDBox" group.
To unsubscribe from this group and stop receiving emails from it, send an email to redbox-repo...@googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

[Peter Sefton]

Hi all,

We're using Mint as a name lookup for other systems at UWS, and this is growing fairly rapidly so I think it is important to make sure this use case is preserved, I guess that wouldn't be a problem with a merged system, but if you just wanted a Mint service would you have a lot of extra stuff to think about that might cause problems?

There is another area where Mint usage might grow, too. Since a Mint installation may already have a great deal of information about research and its context, and it is easy to ingest more (eg by slurping up publications from an Institutional Publications Repository) it could be used as the basis of a research hub system, like the ones emerging at Griffith and JCU, etc. I'm not sure how much interest there is in this use case in the community but we are certainly looking at acquiring a hub at UWS soon.

I imagine that even if the builds are merged you could still install two or more separate instances if you wanted to, Duncan?

Peter

[Duncan Dickinson]

Hi Peter,

Under the ReDBox 1.x version you'd notice little difference if you installed ReDBox & Mint as 2 systems or as a single system. From an "outside" view the systems would respond the same. So, yup, my goal is that this wouldn't break any of those plans you've outlined.

Under the ReDBox 2 version I'm really interested in creating a "ReDBox Core" consisting of storage+security+workflow+transformation behind a RESTful API. This would help us better encapsulate the types of things being stored and allow for more flexibility in terms of other systems accessing the data. This would also allow for "name authority" sources that you want to relay from an authoritative source such as a research system but also provide web forms etc for items such as data management plans and publication descriptions.

The current work we're doing under v1.8 will see curation and harvesting move to their own "micro-services" and this moves us towards the v2 ReDBox Core. However, there's a big chunk of work around getting that core set up and I'm a little unsure how we'll resource that - hence the work around harvest and curation seeking to work in a modular fashion.

I'm glad that my query is getting people responding - it helps make sure we're not cutting options out. Perhaps it'd be worth having a technical teleconference?

Cheers,

Duncan