[seek-dev] ecogrid and ogsa-dai

[Matt Jones]

Bing,

Thanks for the update.  I did catch your earlier notes, but hadn't 
realized that OGSA-DAI was going to be a bad fit with SEEK's needs. I 
had a suspicion, but wasn't totally sure.  Your email clarified that so 
thanks.

Maybe you could tell us more about the XQuery implementation for SRB. 
We hadn't started doing that for Metacat because I htought we had 
decided in our last call that it introduced too many new feature 
requirements on our target data services?  Was it hard to do for SRB? 
Any problems that you can forsee for using it generally for Metacat, 
DiGIR, Xanthoria, etc?  I had thought we had dropped back to our 
previous architectural plan to use the query api that Dave specified in 
the WSDL interface, with a few changes?  Do you still think that's a 
good idea?

Raja, Jing, Dave -- what do you think?

As we have a couple of weeks left before the EcoGrid meeting in Seattle 
maybe we should have a conference call to see if there's anything that 
we should be reviewing and preparing for the meeting.

Matt

Bing Zhu wrote:
> Matt,
> 
> You are right. For OGSA-DAI 2.5, I had some unexpected difficulty to install
> OGSA-DAI in the
> machine, dev.nceas.ucsb.edu. I guess this was because OGSA-DAI is a
> combination
> of too may software. (I am amazed by the way the OGSA-DAI team can put the
> puzzles
> together.) Meanwhile, I sent some help and technical questions to the
> OGSA-DAI
> team.
> 
> Then I moved to my laptop and installed the newly released OGSA-DAI 3.0. For
> testing
> the real OGSA-DAI connection with a database, I had to install the following
> software.
>    Tomcat
>    OGSA (3.0 core distribution)
>    MySql  (Ver 12.21 Distrib 4.0.14 for Windows)
>    OGSA-DAI (3.0)
> The installation went well. To test database access, I also spent one day to
> get familiar
> with MySql commands and created some test data in the MySql database running
> in my laptop.
> Then I was able to access the database via OGSA-DAI interface.
> 
> For both installation efforts, I spent a lot of time in reading OGSA-DAI's
> technical papers
> to understand the structures of DAI Service Group Registry (DAISGR), Grid
> Data Service
> Factory (GDSF) and Grid Data Service(GDS) and their configurations for
> different data
> sources and OGSA-DAI users, which has some change from OGSA-DAI 2.0 to 3.0.
> 
> One of the question I sent to OGSA-DAI team was how to integrate other data
> source(s) into
> OGSA-DAI framework and use OGSA-DAI interface to access that/those data
> source(s). The answer from
> OGSA-DAI team is that currently the OGSA-DAI access a data source via the
> JDBC
> driver for that data source. I forwarded this information to our seek-dev
> mailing
> list on July 18, 2003. This requirement restricts the way for other data
> sources such
> as SRB to use OGSA-DAI implementation unless SRB team builds a JDBC driver
> which
> I don't think worth trying. I am not sure if the Metacat has a JDBC driver.
> If not, the conclusion
> is clear that the current OGSA-DAI is not fit to serve as a common
> implementation layer
> for the heterogeneous data sources within SEEK project.
> 
> The OGSA-DAI implementation also forces users to use GSI for authentication.
> Otherwise,
> it is impossible for users to access data sources with different user name
> and password
> pairs. I will forward a mail from OGSA-DAI support team for my question
> regarding this
> issue to seek-dev today. (I noticed another mail from you this morning that
> our SEEK team
> is going to have our own CA's for using OGSA in SEEK.)
> 
> For the past months, I have also helped with Arun at SRB to partially
> implement XQuery
> with SRB through our own web service approach, which is called Matrix
> project in SRB.
> The SRB Matrix will allow us to have an web service based implementation and
> therefore can be
> easily integrated with other systems. I can do a demo with LTER data stored
> inside SRB
> in our coming SEEK meeting in Sept.
> 
> And we have been also considering building a SRB implementation for SEEK
> QueryInterface.
> For this, We (Metacat and SRB) can discuss it in our Sept meeting and work
> together.
> 
> Sincerely,
> Bing
> 
> 
> 
> 
> 
> 
> 
> -----Original Message-----
> From: seek-dev-admin at ecoinformatics.org
> [mailto:seek-dev-admin at ecoinformatics.org]On Behalf Of Matt Jones
> Sent: Sunday, August 24, 2003 4:02 PM
> To: Bing Zhu
> Cc: seek-dev
> Subject: [seek-dev] ecogrid and ogsa-dai
> 
> 
> Bing,
> 
> I was in Santa Barbara this week and worked with Jing on the EcoGrid
> stuff.  We are coming down to the wire.  Our next EcoGrid meeting is in
> about 3 weeks, and I want to be sure all of the assigned tasks are done
> by then.  It sounds like you've had some difficulty with the OGSA-DAI
> install -- could you send a general update on its status, and an
> assessment of whether it would actually be useful in the context of
> SEEK?  My impression is that its not ready yet for our needs, but I'd
> like to hear your thoughts on it.  I'm assuming that it is pretty
> immature as you've been working on it for several months now, which is
> about 3 times longer than I had expected it to take given that you are
> 100% time on SEEK.
> 
> Jing has GT3.0.1 installed on dev.nceas.ucsb.edu, and the CA is working
> there.  We worked together on getting the Metacat client to run as an
> OSGA service, and Jing and I will be getting the EcoGrid Query Interface
> implemented as a grid service before our September meeting.  I think the
> grid registry stuff will work out well for us, but I am a little worried
> about the gt3 maturity level (I've had some trouble with it under redhat
> 8).  What's your experience there?
> 
> Thanks for the update,
> 
> Matt
> 
> 
> _______________________________________________
> seek-dev mailing list
> seek-dev at ecoinformatics.org
> http://www.ecoinformatics.org/mailman/listinfo/seek-dev