Hi everybody,
here is some summary of what we did so far: at SEMANTiCS 2015 we've
presented dld: dockerizing linked data [1]. The idea of this was to
provide a tool that basically creates docker-compose setups to easily
create an infrastructure for serving RDF datasets via a SPARQL endpoint
or other applications like OntoWiki. We've followed the principle of
having one container per service, so we didn't want to put the complete
stack including the data into a single container (Microservices [2],
Single Responsibility Principle [3]).
To achieve this we have identified three (or four) tasks in a setup:
(1a/b): load and back-up data, in case of DBpedia, there would only be
loading data
(2): storage of the data
(3): presentation, exploration and editing data
This idea is meant to be very generic to cover different types of setups
dealing with RDF data or data in general. For the case of DBpedia and
also to achieve a "scientific reproducibility", as Dimitris mentioned,
we might have to rethink this setup. Further I think also in the docker
community best practices have evolved. We should also keep an eye on
performance of services running in the containers vs. running them
"natively" on a system, we have experienced some impact here already,
but this might differ from setup to setup.
[1] https://dockerizing.github.io/;
http://www.bibsonomy.org/bibtex/2b1e393a0bfd62e83b99704a52c20c877/aksw
[2] https://en.wikipedia.org/wiki/Microservices
[3] https://en.wikipedia.org/wiki/Single_responsibility_principle
All the best,
Natanael
(Sorry for sending it twice, but I first had to subscribe)
--
Check out the vibrant tech community on one of the world's most
engaging tech sites, SlashDot.org! http://sdm.link/slashdot
___
DBpedia-developers mailing list
DBpedia-developers@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/dbpedia-developers