ILDG quick links

ILDG plone site

Browse ensembles

USQCD portal

LDG portal

Japan portal

UKQCD portal

Download configurations

get a DOEgrids certificate

register with ILDG VO

Install the ILDG Client Tools from LDG

Virtual Data Toolkit (Globus)

ILDG development

ILDG-11 workshop

ILDG-middleware archive

Datagrids for Lattice QCD" Nucl. Inst. and Methods in Physics Research, Vol. 559, Issue 1 (2006) 57-61

ILDG INCA monitor

ILDG member sites

CSSM

Japan

LDG

UKQCD

USQCD

International Lattice Data Grid

The International Lattice Data Grid (ILDG) is an international organization which provides standards, services, methods and tools that facilitates the sharing and interchange of lattice QCD gauge configurations among scientific collaborations, by uniting their regional data grids.

The ILDG promotes a common schema to markup meta-data (e.g. physics and algorithmic parameters) that describe ensembles of gauge configurations. Each regional grid catalogs the meta-data for the ensembles and gauge configurations they wish to share. Through each of the ILDG web portals, a user can search any or all of the regional meta-data catalogs which implement the ILDG web service interfaces.

Gauge configurations are provided by the ILDG in a standardized file format which is compatible with the SciDAC QIO input/output library. Member collaborations within the ILDG are developing scripting and GUI based tools to simplify downloading data files. Each file to be downloaded is specified by a unique logical file name (LFN) that is part of the data file's meta-data. Further, each regional grid maintains a file catalog (also known as a replica catalog) that relates the logical file name to one or more copies of the data file in the regional grid's storage resources. The tools being developed will automate contacting the regional grid replica catalog and executing the file transfer.

Obtaining files from the ILDG has the following prerequisites:

  • You must obtain a Grid Certificate that is your "electronic identity"
  • You must register this Certificate in the ILDG Virtual Organization VOMS server
  • You need to have access to a machine that has the various grid and file transfer clients installed

Establishing your grid identity

Obtain a personal grid certificate. Members of the USQCD Collaboration need to apply for a grid certificate from doegrids.org. Select "new user" from the menu bar and fill out and submit the application form. Once your identity is verified, you will be invited to return to doegrids.org to retrieve you certificate. You must use the same computer, account and web browser as before to obtain your certificate. You will need to export a copy of your certificate from the web browser in order to use it with grid tools, on other computers or with other web browsers. Do not forget your certificate's pass phrase: there is no way to reset or recover a forgotten pass phrase!

Create a key pair from your certificate for GLOBUS. You need to prepare a key pair from your grid certificate and install a copy of the key pair in your account on computers where you will be using ILDG grid applications. Contact your system administrator to find a system that has GLOBUS tools installed. The key pair is typically installed in the directory ~/.globus. YourCert.p12 is the file containing the copy of your certificate exported by your web browser.

To extract your certificate:
openssl pkcs12 -in YourCert.p12 -clcerts -nokeys -out $HOME/.globus/usercert.pem
To extract your encrypted private key:
openssl pkcs12 -in YourCert.p12 -nocerts -out $HOME/.globus/userkey.pem
You must protect your private key:
chmod go-rw $HOME/.globus/userkey.pem

Files usercert.pem and userkey.pem may be copied to other systems where you wish to use grid tools.

Join the ILDG Virtual Organization. Membership in the ILDG VO allows grid sites to verify your affiliation with the International Lattice Data Grid Organization. Joining the ILDG is a prerequisite for being able to download gauge configuration files. You must have your personal grid certificate loaded in your web browser before you begin the ILDG VO application. During registration you will declare you affiliation with one of the regional member grids of the ILDG. Affiliation with a regional grid may give you an extended set of access permissions within that grid.

Browsing meta-data catalogs

Several ILDG web portals that allow a user to browse the regional meta-data catalogs are being developed. Users are invited to try any of the available portals and use the one that best suits their needs. Here is a list of existing portals (circa January 2008):

Other gauge ensemble archives that are not part of the ILDG are listed below. Users will find directions on how to download files at each of the sites.

Ensemble and Configuration Filenames

An ensemble and individual configurations are identified respectively by Markov chain (MC) uniform resource identifier (URI) or a logical file name (LFN) within an ILDG meta-data catalog. Markov chain URIs of the form:

mc://[REGIONAL GRID]/[GRID SPECIFIC ENESMBLE ID]

whereas LFNs are of the form:

lfn://[REGIONAL GRID]/[REGIONAL GRID SPECIFIC FILENAME]

mc://USQCD/MILC/asqtad/2_plus_1_flavor/4096f21b708m0031m031

denotes one of the MILC "fine" ensembles, and

lfn://USQCD/MILC/asqtad/2_plus_1_flavor/4096f21b708m0031m031/series_0/l4096f21b708m0031m031a.204

denotes configuration number 204 of series 0 in this ensemble.

A configuration LFN must be mapped to a storage url (SURL) which encodes a file in a storage device before the configuration can be downloaded. This mapping is performed by a grid replica catalog. For example, A SURL corresponding to the LFN above is:

srm://fndca1.fnal.gov:8443//pnfs/fnal.gov/usr/lqcd/USQCD/MILC/asqtad/2_plus_1_flavor/4096f21b708m0031m031/series_0/l4096f21b708m0031m031a.204.ildg

The ILDG has developed some client tools (ildg-get) that allows one to interrogate the file catalogue and automate a download

Sample Download Session

On a system where Globus client and the ildg-get tools are installed

Signing on to the Grid - Initializing a Proxy

# create a grid proxy certificate
$ grid-proxy-init
Your identity: /DC=org/DC=doegrids/OU=People/CN=Frodo Baggins 60510
Enter GRID pass phrase for this identity:
Creating proxy ......................................................... Done
Your proxy is valid until: Wed Sep  22 08:31:25 3019

# check certificate validity
$ grid-proxy-info
subject  : /DC=org/DC=doegrids/OU=People/CN=Frodo Baggins 60510/CN=proxy
issuer   : /DC=org/DC=doegrids/OU=People/CN=Frodo Baggins 60510
identity : /DC=org/DC=doegrids/OU=People/CN=Frodo Baggins 60510
type     : full legacy globus proxy
strength : 512 bits
path     : /tmp/x509up_u11211
timeleft : 11:52:49
      

Using ildg-get to download metadata

The following invocation will download the Metadata XML for a given MC URI

$ ildg-get -mdc-only mc://USQCD/MILC/asqtad/2_plus_1_flavor/2464f21b676m005m050
      

This will create a file called 2464f21b676m005m050.xml containing the ensemble metadata

One can also download metadata for a configuration file using ildg-get:

$ ildg-get -mdc-only lfn://USQCD/MILC/asqtad/2_plus_1_flavor/2464f21b676m005m050/series_0/l2464f21b676m005m050.978
     

This will create a file called l2464f21b676m005m050.978.xml containing the configuration metadata

Using ildg-get to interrogate the replica catalogue

ildg-get can be used to map a configurations LFN to its storage URL. This involves the tool interrogating the file catalog of the regional grid in question. For example the command:

$ ildg-get -fc-only lfn://USQCD/MILC/asqtad/2_plus_1_flavor/2464f21b676m005m050/series_0/l2464f21b676m005m050.978
      

returns the result

srm://fndca1.fnal.gov:8443//pnfs/fnal.gov/usr/lqcd/USQCD/MILC/asqtad/2_plus_1_flavor/2464f21b676m005m050/series_0/l2464f21b676m005m050.978.ildg
      

At this point, one can download the file using a tool like srmcp, for example:

$ srmcp \
   srm://fndca1.fnal.gov:8443//pnfs/fnal.gov/usr/lqcd/USQCD/MILC/asqtad/2_plus_1_flavor/2464f21b676m005m050/series_0/l2464f21b676m005m050.978.ildg
   file:////data/l2464f21b676m005m050.978.ildg
	   

Using ildg-get to download the file

The ildg-get utility supports a direct download of a file from its LFN. Essentially combining the two steps of the File Catalogue lookup and the srmcp invocation into one command. For example:

ildg-get lfn://USQCD/MILC/asqtad/2_plus_1_flavor/2464f21b676m005m050/series_0/l2464f21b676m005m050.978
      

should download the data file to the local directory as l2464f21b676m005m050.978.ildg.

Verifying the file using ildg_cksum

Once the file is downloaded the integrity of its data can be verified by computing a checksum on the file, and comparing this to the checksum in the metadata. The configuration file XML contains a tag <crcCheckSum> for example for the configuration discussed above, the metadata file l2464f21b676m005m050.978.xml shows that the CRC checksum of the data is:

<crcCheckSum>2119168653</crcCheckSum>

We can compute the checksum on the downloaded configuration itself using the ildg_cksum utility

$ ildg_cksum ./l2464f21b676m005m050.978.ildg   
2119168653
       

and we can see that the checksum from the downloaded file matches that in the metadata document

usqcd-webmaster@usqcd.org