Документ взят из кэша поисковой машины. Адрес
оригинального документа
: http://www.adass.org/adass/proceedings/adass97/charys.html
Дата изменения: Fri May 15 22:17:26 1998 Дата индексирования: Tue Oct 2 01:29:33 2012 Кодировка: Поисковые слова: molecular cloud |
Next: Browsing the HST Archive with Java-enriched Database Access
Up: Archives and Information Services
Previous: The IUE Archive at Villafranca
Table of Contents -- Index -- PS reprint -- PDF reprint
Sumitra Chary and Panagoula Zografou
Smithsonian Astrophysical Observatory, AXAF Science Center,
Boston, MA 02138
1Trademark of Sun Microsystems
The AXAF archive architecture involves a number of servers, RDBMS servers for managing relational data and archive servers that manage datafiles. Access to datafiles is provided by existing client applications which send data ingest, search and retrieval requests in a 4gl language to an archive server (Zografou et al. 1997). An architecture was developed to build a powerful and uniform interface to the various servers through the World Wide Web. The advent of JDBC a specification for an API that allows Java applications to access various DBMS using SQL became the obvious choice to build this interface.
Figure 1 shows both the client-server architecture developed using the Database vendor API (Zografou et al. 1997) where both the client and the data servers communicate using the same protocol and the 3-tier architecture which allows a Java client applet to communicate with the data servers behind a firewall. The middle tier consists of Sybase jConnect for JDBC package. This includes the Cascade HTTP Gateway which is the gateway to the RDBMS and archive servers.
According to Java security restrictions, applets can only connect back to the host they were downloaded from. If the Web server and database server were on the same machine then the Cascade gateway is not required. But in this case the data server hosts are behind a firewall hence the Cascade gateway acts as a proxy and provides the path to the data servers. The HTML documents and Java code are served from the Apache Web Server tree. The Apache Server is also used to execute a cgi-script which performs transformations of input and output data when required. Hence the Apache and the Cascade run on the same host on different ports so that the applet downloaded from the Apache server tree can connect back to the same host at the port of the Cascade gateway.
Due to Java applet security restrictions, files cannot be created on
the client filesystem. Hence the result contents are sent by the applet
to a cgi-script on the Web Server host which creates a file on its
filesystem from the binary stream. This file is then FTP'ed to the FTP
server in a previously created location. The cgi-script informs the
client of this path which is displayed to the user.
In the future with JDK1.1 and a compatible Web browser it may be
possible to re-create files on the client side depending on
authentication and access control schemes. This would eliminate the
extra step of passing data to a cgi-script.
<HTML> <applet code = "ocat.class" codebase = "base URL of applet to be displayed" archive = ocat.zip width=700 height = 650> <param name=proxy value="IP address:port of Cascade Gateway"> <param name=host value="IP address of DataServer"> <param name=port value="Port of DataServer"> <param name=uid value=""> <param name=pass value=""> <param name=cgiserver value="IP address of Web Server"> <param name=cgiserverport value="Port of Web Server"> </applet> </HTML>
Zografou P., Chary, S., DuPrie, K., Harbo, P., & Pak K. 1998, ``The ASC Data Archive for AXAF Ground Calibration", this volume
Next: Browsing the HST Archive with Java-enriched Database Access
Up: Archives and Information Services
Previous: The IUE Archive at Villafranca
Table of Contents -- Index -- PS reprint -- PDF reprint