[ogsa-dmi-wg] [Fwd: Re: [GSM-WG] Data Transfer Chapter from OGSA Data Architecture Document]

[Ravi Madduri]

Alex Sim's Comments

-------- Original Message --------
Subject: Re: [GSM-WG] Data Transfer Chapter from OGSA Data Architecture 
Document
Date: Tue, 5 Dec 2006 10:27:51 -0800
From: Alex Sim <asim at lbl.gov>
Reply-To: asim at lbl.gov
Organization: Lawrence Berkeley National Laboratory
To: 'Dave Berry' <daveb at nesc.ac.uk>, <shoshani at lbl.gov>, 
<peter.kunszt at cern.ch>, <makeit at eyou.com>, <a00ymw00 at nchc.org.tw>, 
<kju at kisti.re.kr>
CC: gsm-wg at ogf.org, "'Allen Luniewski'" <luniew at us.ibm.com>
References: <2AC085FCF3E737448BE6E2D11A2709F64F1A37 at crosby.nesc.ed.ac.uk>

A few comments:
1. it's not clear how the client finds out the source and target "WS-Name".
2. it's not clear how the client monitors status and adjusts performance and
terminates the transfer. Termination part can be assumed to be there from
the higher arch (is it resource properties?), but can clients assume the
status and performance adjustment of the data transfer to be done as well
through the same? If so, operations services such as pause/resume/stop can
be done through the same higher (or separate) mechanism? Same for
GetSupportedProtocols?
4. Who makes sure the data transfer does not fail because of the lack of
storage space in the "sink"?  Is it DMI service or the client or the actual
underlying transfer service such as RFT or SRM ? In the requirements of the
data transfer mechanism, should that be added as a statement about that?

--Alex


> -----Original Message-----
> From: Dave Berry [mailto:daveb at nesc.ac.uk]
> Sent: Friday, December 01, 2006 9:46 AM
> To: Dave Berry; shoshani at lbl.gov; peter.kunszt at cern.ch; makeit at eyou.com;
> a00ymw00 at nchc.org.tw; asim at lbl.gov; kju at kisti.re.kr
> Cc: Allen Luniewski; gsm-wg at ogf.org
> Subject: RE: Data Transfer Chapter from OGSA Data Architecture Document
> 
> Dear GSM WG,
> 
> We haven't received any comments in response to our request below.
> Please could you read our Data Transfer chapter and send us feedback?
> 
> Best wishes,
> 
> Dave Berry
> Deputy Director, Research & E-infrastructure Development National
e-Science
> Centre, 15 South College Street
> Edinburgh, EH8 9AA                  +44 131 651 4039
> 
> 
> 
> > -----Original Message-----
> > From: Dave Berry
> > Sent: 21 October 2006 12:57
> > To: shoshani at lbl.gov; peter.kunszt at cern.ch; makeit at eyou.com;
> > a00ymw00 at nchc.org.tw; asim at lbl.gov; kju at kisti.re.kr
> > Cc: Allen Luniewski; gsm-wg at ogf.org
> > Subject: Data Transfer Chapter from OGSA Data Architecture Document
> >
> > Dear GSM WG,
> >
> > The OGSA Data Architecture WG held a F2F at the end of August.  At
> > that time, we worked with the DMI WG to clarify what the data
> > architecture calls out in the area of data transfer.  We would like to
> > share what the architecture document now says about data transfer.
> > This extract is below.  Your feedback will be most appreciated.
> >
> > Dave Berry
> > Deputy Director, Research & E-infrastructure Development National
> > e-Science Centre, 15 South College Street
> > Edinburgh, EH8 9AA           Tel: +44 131 651 4039
> >

--
   gsm-wg mailing list
   gsm-wg at ogf.org
   http://www.ogf.org/mailman/listinfo/gsm-wg


-- 
Ravi K Madduri
The Globus Alliance | Argonne National Laboratory
http://www-unix.mcs.anl.gov/~madduri