LSC Data Analysis Software Working Groups

Navigation

DASWG
LSC
LIGO


DASWG LAL Doxygen

Docs

How-to
Minutes
Technical
Software Docs

Download

Browse CVS
Repositories

Participate

Change Control Board
Edit these pages
Sub-committees
Mailing List
Telecon

Projects

DMT
geopp
Glue
LAL Home Page
LALApps Home Page
LDAS
LDG Client/Server
LDM
LDR
LIGOtools
MatApps
Metaio
Onasys
Online
OSG-LIGO

Agenda and Minutes, 20 March 2008

Agenda: Thu, 20 March 2008 (joint CompComm/DASWG f2f)

 

Where: The SCR (351 W Bridge--Science Conference Room) starting at 2PM
PDT.

Dial in instructions: 

Reservationless-Plus Toll Free Number (US / Canada): (866) 380-5536
Reservationless-Plus International Dial-In Number:(816) 249-4731 
Conference Code: 3829781554

Numbers from outside the US

France: 08-009-101-24
Italy: 800 870 318
Germany: 0 800 182 1591
UK: 0 808 234 7914


14:00 (30') Review projects and needs for software/LDG support
data reading time format jamie
[HTML , Patrick to lead discussion]

There was a lot of discussion of trying to define an API for data
access. Shourov and Jamie will come back with a short document
explaining the issue so that we can decide where this should be put. 

Warren gave a demo of the new authentication infrastructure. There
were lots of questions and they were put offline.


15:30 (30') Network NDS status and plans
[John Z]

John explained ths status of this project.  Stuart summarized the
picture that has been developed for the CDS side.

IFO -- DC --- broadcast 1/16 --- framebuilder --- QFS --- LDAS
        |                                          |
        | --------------------------------------- NDS++ -- off-site
        |                                                      |
        | --------------------------------------------------- NDS 
                                                               |
                                                         control room

People who should be talking to each other are Alex, Josh, John, Lisa,
Dave, Rolf, Shourov, Greg, ....

Authentication needs to be easy to use for everybody

Keith asked if there are any sticking points:  JZ replied
authentication, backward compatability; JZ would like to define an
API.

Stuart explained that Rolf and Alex are working on the DC part above

Milestones:  

API -- 2 weeks ------------- 2 April
non-authenticated service -- 1 May

clients are:  DTT, DMT, MatLab, ligoviewer, ....

JZ, Duncan, Josh, Warren, Lisa, Dave, Rolf ....


14:30 (30') Review of S6 data generation, publication 
[PPT , Greg]

RDS generation:
* will be done the same way as it was in S5
* scripts that drive publishing will include error checking
* jobs running should repeatedly try for data
* RDS data showed up within 10 minutes of real time
* raw data from reflective memory into archive

h(t) generation:
* Xavi thinks we should generate it for all times
* Virgo would like long frames
* Proposal for faster h(t) generation inside of DMT for online analysis
* could provide which channels, sample rates are available for times

Stuart asked about extracting metadata from frames as they are going
by:  agreed to extract, but not design service

15:30 (30') Review of segment database status and plans
[PPT , Duncan, Ping] 

Kipp asked for decoupling of protocol from the python class for
segments. 

Ability to be able to extract a snapshot

Peter raised the question about science mode based on state vector or
based on lock-loss. 

Proposals:
* use multiple channels to identify the start/end of science
* maybe add a channel to RDSs which correctly reflect this

Shourov wants a dump of the daily segs for quick end user access
- would like to have the segments written to a web server from publish script

Shourov wants to have an unauthenticated segment server

Peter raised issues of daily dumps; wasteful in terms of space, and
time of transfer


16:15 (30') Review of monitoring project status and plans
[Filippo]




16:00 (30') Infrastructure for online analysis in S6
[PDF , Stuart]

END -- items after this will be handled at the next telecon

17:15 (30') Workflow tools and projects: ability to use OSG
[Britta]

17:45 (15') Version control and bug tracking
[Kipp and Jamie]

15:00 (30') Data distribution status and plans
[PDF , Scott]

(30') Cluster OS upgrade timeline and path
[Stuart]

(30') Archival system for processed results; update and planning
[Brian (remote participation)]


Other things:

* h(t) generation and distribution
* Event database/web areas
* Data discovery tools
* Authentication and authorization
* Software maintenance, support and release
* Metadata services
* LDG Client/Server

Minutes

Tentative attendees:

Xavier Amador (maybe)
Stuart Anderson
Warren Anderson
Patrick Brady
Duncan Brown
Kipp Cannon
Shourov Chatterji
Filippo Grimaldi
Jonah Kanner
Scott Koranda
Ed Maros
Greg Mendell
Adam Mercer
Peter Shawhan
Keith Thorne
Igor Yakushin
John Zweizig
Jonah


Friday, March 21 2008

Meeting with Rolf, Keith, Kipp, Josh, Duncan, Stuart, Shourov, Lisa,
Xavi, John Z

Agreed to have API team (Alex, Josh, John, Lisa, Dave, Rolf, Shourov,
Greg, .... Warren and Duncan for security) define API in about 2 weeks

* Update from John next week at DASWG
* Draft API 2 weeks, 2 April

Summary of discussion:

Xavi will deliver and insure code is available for 
RDS ... Greg will make that work
DMT ... generation will use same code base (Who?)
Real Time .... or is this the same as DMT? (Xavi)




2. Xavi

1. Xavi to understand the latency of the data generation 
2. If we can achieve a latency ~X sec, try to get hardware to do generation

$Id: 080320.html,v 1.22 2008/05/14 18:47:11 patrick Exp $