1. Distributed Data Servers and Web Interface in the Climate Data Portal Willa H. Zhu Joint Institute for the Study of Ocean and Atmosphere University of Washington (PMEL/NOAA)

2. Outline  Climate Data Portal (CDP) Overview  Distributed Object System (what, why, how)  CDP System Architecture  CORBA Client/Server Application  CDP Data Servers  CDP Client implementation  CDP Client Web Interface  Feature Plans

3. Climate Data Portal Overview  It provides consistent and uniform access to geographically distributed in-situ data collections  Use of CORBA and LDAP services provides instantaneous network “awareness” of the distributed data servers  Sophisticated and customized Java tools facilitate collections typical of in-situ observations  Selected data from distributed data servers can be plotted together for comparison  Dual access: Desktop Java application (with on-line help and tutorial) Web Interactive Interface  Distributed Object System

4. What is Distributed Object System  Objects are software units that encapsulate data and behavior  Systems that feature these objects are termed distributed object systems Applications and interfaces are  remotely accessible  easily customized Software used to create distributed server objects are  independent from the architecture  independent from hardware platform used  totally transparent to their clients

5. Why Distributed Object Systems  Divide client/server applications into self- managing components that can interoperable across networks and operating systems  Centralize data access and retain localized data management for heterogeneous data sets from different geophysical locations  Require no data being transported to a central data repository

6. Distributed Object Technologies complete distributed object platform operates across networks, languages, component boundaries, and operating systems network transparency automatic recognition of server supplements Java with a rich set of distributed services CORBA (Common Object Request Broker Architecture) CORBA provides a facility to establish the remote communications with distributed objects and to handle network interactions in passing data between objects.

7. Climate Data Portal Architecture PMEL CORBA Server TAO Data support In-situ Data support UH CORBA Server Sea Level Data support NODC CORBA Server GTSPP Data support Object Request Broker (Internet Inter ORB Protocol) CORBA Client Java Servlet Client support Web Browser Web Server HTTP CORBA Client Java Application Data Data Servers Data End Users Distributed Object System

8. CDP Data Servers Dual access from Java desktop application Web client interface  CORBA tool  ORBacus  fully compliant with the CORBA specification  free for non-commercial use and available with complete source code http://www.ooc.com  In-situ data collection presently available:  Global sea level data (realtime) – University of Hawaii  Global Temperature-Salinity Profile Program (GTSPP) – NOAA/NODC, Silver Spring  TAO El Nino buoy data (realtime) – NOAA/PMEL, Seattle  PMEL insitu data collections – NOAA/PMEL, Seattle  Data Supported:  NetCDF data format  MySQL database maintaining meta-data

9. CORBA Client/Server Application Client Stub Files (Java) Object Request Broker Server Skeleton Files (Java) Client Implementation IDL Compiler IDL File (interface specification) Object Implementation Development steps : IDL = Interface Definition Language (ORBacus, a CORBA tool, is used)

10. CDP Java Desktop Client Application  Provides higher degree of interactivity  Supported for most commonly used workstations and operating systems  On-line help and tutorial  Supported by other Java tools (Ndedit) for data sub-setting  Easily download and install ftp://www.epic.noaa.gov/java/CDPclient/install.html

11. CDP Client Web interactive interface http://www.epic.noaa.gov/cdp  Java Servlets provide HTML front-end to the CORBA server  Java Applet and JavaScript enhance the user interactivity  Java Scientific Graphic Toolkit (SGT) is used to generate plots  Result pages are dynamically created on- the-fly

12. Listing/Plotting options:  View data collection properties  View data item properties  Plot/Co-plot profile data  Plot/Co-plot time series data  Make section 2-D plot  List data in ASCII format  Download data in netCDF format CDP Web Client Interface Features Data search and access by choosing : keywords data servers data collections data items geographic location depth range time range

13. CDP Web Data Selection Locate data collections:  keywords  geographic location  time range

14. CDP Web Data Selection (continued)  View Collection Properties  Retrieve data items: Specify data collections Location ranges (lon,lat,depth, time)

15. View item properties Select items from item list Specify access options CDP Web Data Selection (continued)

16. Select variable Select subset CDP Web Data Selection (continued)

17. CDP Web Results Pages (profile plot) Profile Plot Showing co-plot of temperature data from two resources:  NODC GTSPP Silver Spring  PMEL TAO Seattle

18. CDP Web Results Pages (Time Series Plot) Time Series Plot

19. CDP Web Results Pages (Section Plot) 2-D Section Plot

20. CDP Web Results Pages (Data Listing) Data Listing

21. CDP Web Results Pages (File Downloading) Select variables and subset Download netCDF file

22. Future Plans  Include more data servers Intl. Pacific Research Center in-situ collection – University of Hawaii WOCE hydrographic data – Scripps Institution of Oceanography WOCE Time Series data – NOCD, NOAA WOCE Acoustic Doppler Current Profiler data – University of Hawaii In-situ data collection - Marine Environmental Data Service (MEDS) Canada  Use LDAP (Lightweight Directory Access Protocol) directory server for CORBA data server registry for Web  Add data sub-setting capabilities for Web access  DODS proxy server - selected DODS datasets http://www.epic.noaa.gov/cdp Willa H. Zhu (willa.zhu@noaa.gov)