| • |
|
|
|
|||
|
||||
Meeting with Ethan Alpert and Dave Brown of SCD for a discussion and demonstration of their work on dataserver project.
Access to products on their server is mainly through a Web interface based on the PMEL Live Access Server (LAS). Client software, such as Ferret and NCL actually runs on the server. Access to the functionality of these clients is through a Web interface constructed with a combination of HTML, Javascript, Java, and Perl CGI scripts. The main components of their server systems are:
The user can use the system to download data files (or parts of data files) through this interface for analysis on her own workstation, but no emphasis has been placed on having clients on user computers access the data from servers via DODS (or ADDE).
Regarding the availability of NCL and NCAR graphics, both products are now freely available. However, at the present time, the source code is available for NCAR Graphics but only binaries are available for NCL. Source availability for NCL is planned but only after the proper packaging and documentation is developed. Apparently this is true of Ferret also.
SCD is working with Brian Doty to incorporate a version of his DODS/GRADS server into the system, but that is a ways off.
An interoperable set of dataservers based on the GLOBUS Toolkit and software tools developed at DOE labs and supercomputer centers. Much of the work was done for the PCMDI (Program for Climate Model Diagnosis and Intercomparison) system. The Earth System Grid enables users to ask about the availability of data within the system, then the system itself figures out which files satisfy the request and automatically FTPs the files to the requester's computer.
A key component of the Earth System Grid is an authentication system based on the same certification system used by the Access Grid. This is based on work at SDSC and is the same system being incorporated into UCAR and is under study for use in DLESE and NSDL. This is an important common component of all these systems because it would allow a community participant to have one account for access to all the different facilities for which she is authorized.
All these data access systems also need data discovery facilities. The Grid has a mechanism whereby data contributors can manually enter metadata about their datasets. This data is then made available at a central site via LDAP. That's the part of the system that Ethan will be working on to augment the system so metadata will be available on each server as well as at the central site. They are also interested in XML representations for the metadata in individual datasets. Not much consideration has been given to discovery metadata standards and guidelines such as the FGDC standard and the GCMD DIF specification.
| Contact Us • Site Map • Search • Terms and Conditions • Privacy Policy • Participation Policy | |||||
|
|||||
