Data Services Newsletter

Volume 3 : No 1 : March 2001

Data Handling Infrastructure at the IRIS DMC

Previous DMS Electronic Newsletter articles discussed how the DMC has shifted toward electronic reception and archiving of data at the DMC as well as reception of data in real time. (See December 2000 issue for articles). In many aspects, the IRIS DMC is attempting to position itself for even larger influxes of data from such projects as the USArray, the Advanced National Seismic System (ANSS), and data from the Ocean Sciences Community.

In October of last year, the software engineering group at the IRIS DMC began working closely with the staff at the University of South Carolina to integrate information at the IRIS DMC with the distributed computing environment being developed by Philip Crotwell (University of South Carolina) and others. The software being developed is using modern object-oriented approaches and incorporating advanced techniques for distributed computing. Most of this effort is leveraging the design work performed within the IRIS DMS as part of the FISSURES project.

Changing Access Patterns: “Just in Case” versus “Just in Time”

The banking industry wants to keep your money as long as they can; they make money using your money. Before the proliferation of ATM machines, customers would go to the bank periodically to withdraw cash, “Just in Case” they would need it. Since ATMs have become so accessible, we no longer feel as strong a need to have large sums of money in our wallets. Now bank customers know that they can get the cash they need “Just in Time” to make a purchase. The bank wins and the customer wins due to this change in access pattern.

Until now, seismologists would often request information from the DMC “Just in Case” they needed it later. An example of this “Just In Case” access would be recovering a dataless SEED volume and storing it on the researcher’s home computer, knowing that after seismograms were recovered the response information in the dataless SEED would be accessed to correct for instrument responses. The problem with this approach is that as new information about stations and their recording channels is acquired the metadata in the dataless SEED volumes grows stale and less reliable.

Using the new Data Handling Infrastructure being developed, the researcher will be able to access needed information directly from the Oracle Database when they actually need to use it; the “Just in Time” approach.

The IRIS DMS Data Handling Infrastructure

The new IRIS DMS Data Handling Infrastructure (DHI), is being developed to more closely link information (metadata) and waveforms stored at the DMC with the research seismologist’s desktop computing system. It is intended to ease the researcher’s task of insuring they have the most up to date metadata as well as ease the storage of the actual waveforms on the client’s own computer.

The software engineering group at the IRIS DMC (Rob Casey, Chris Laughbon, Sue Schoch and Sandy Stromme) has assumed the responsibility of developing and maintaining the Data Center Interface portion of the DHI. For the initial implementation they have developed object-oriented interfaces that allow access to

  • Event information
  • Station and channel information
  • Channel Response Information
  • Waveforms from the new FARM data products.

The first three types of information come directly from the DMC’s Oracle Database whereas the waveforms come off of the large disk-based buffer that holds the waveforms in the FARM products.

Philip Crotwell has played the central role in developing the object-oriented approach for the DHI. He is developing the client side software that will connect with the Data Center Interface discussed above. Easy access to the information types listed previously will be enabled through this application.

Waveforms will move from the DMC to the DHI Application where they can be displayed statically, scrolled on the screen to simulate data from an earthquake, or displayed as real-time feeds. After display, the waveforms can be presented to a user’s application for processing or simply be stored in a local directory system in a format of the user’s choice.

We anticipate the beta test of this system will take place during the month of March and be ready for presentation to the general IRIS community near the time of the IRIS workshop in Jackson Hole.

by Tim Ahern (IRIS Data Management Center)

18:12:24 v.22510d55