Re: Ongoing architecture work

[Matthew Turk <matthewturk@xxxxxxxxx>]

On Wed, Oct 8, 2014 at 11:24 AM, Ian Foster <foster@xxxxxxx> wrote:
> Dear Matt:
>
> As NDS is supposed to be a collaborative project, I don't think it is appropriate to call this the "preliminary architecture for the NDS." Call it an NDS pilot, perhaps?

Sure!  I agree with that, and I have changed all the locations in the
document where I've used the singular "the" to "a" or "an" -- I
thought I emphasized this in my original email, but see now that I
could have done a much better job of saying so.

>
> We are implementing (in fact, have operational) another NDS pilot based around Globus services--the Materials Data Facility. Having other parallel efforts exploring other technologies is fine, but calling it the NDS architecture disenfranchises us at Chicago and I believe is premature.

I've changed the name to be "co-pilot" as well.  From what I can tell,
the MDF is progressing really well, and I (the primary author of this
document) have attempted at all times to keep interoperation with what
I know of for MDF technologies as a top priority.

>
> Ian.
>
>
> On Oct 8, 2014, at 11:13 AM, Matthew Turk <matthewturk@xxxxxxxxx> wrote:
>
>> Hi all,
>>
>> A couple of us have been working on deploying at NCSA a preliminary
>> architecture for the NDS, based around commonly available software and
>> with some homegrown development as well.  We've been taking notes
>> here:
>>
>> https://docs.google.com/document/d/1SBRatAebX5-eQXB4qTFb5UGV0Cwy8Ppux_2RPs2ZUZY/edit#heading=h.b1qfdq9dlvv
>>
>> (Please excuse the first person, that's all me...)
>>
>> There are some action items in the document, and it's also able to be
>> commented on by anyone with the link, so feel free to do so.  Right
>> now, we're trying to stand up:
>>
>> * NDS Dashboard (IanT, do you want to chime in about this?)
>> * Analysis launching (Kacper, feel free to share about this)
>> * Authentication using first LDAP here at NCSA and soon InCommon
>> * Storage backed by disk (frontend by iRODS), cloud storage (NCSA and
>> SDSC, fronted by SciDrive), and access to other providers (fronted by
>> ownCloud)
>> * EZID DOI generation -- IanT has been exploring this, and we're
>> hoping to sync up with some folks from the UIUC library to understand
>> better the process and for some direct collaboration
>>
>> There's some more going on as well, but I'm writing primarily to open
>> up what has been a somewhat insular discussion to the broader
>> community and to seek feedback, solicit engagement and to see where we
>> can take this together.  We're still in the process of getting code
>> online, but it will be hosted at either the bitbucket or github sites,
>> and we're keen to get people to explore, contribute feedback or
>> changes, and help out with conceptualizing.  Please don't take the
>> list above as "all inclusive"!
>>
>> One principal goal is that whatever we develop is able to have new
>> components brought in with a minimum of fuss, components swapped out
>> with a similarly minimum of fuss, and that we aim to make something
>> that's (easily) deployable anywhere, not just at NCSA.
>>
>> We're keen to get some feedback, either on the document, or here over email!
>>
>> Thanks,
>>
>> Matt
>>
>