I'm working on connecting the 390 Web Server to CICS/390 transactions, DB2/ESA and IMS. This enables a more secure, faster pathlength for client browsers to access data stored on their MVS systems and run transactions.
Related areas of work are tools to convert HTML to SQL, DL/1 and CICS execs as well as the conversion in the opposite direction. Another part of the 390 Web work is accepting distributed method calls from client objects to our MVS Web and DSOM (Distributed ORB) servers.
My interest is in using object technology to encapsulate legacy data. I wrote a position paper describing the basic ideas for a recent ACM SIGCOMM workshop.
Here is a brief summary of this paper:
In the world of information services heterogeneity is inevitable: not only because of the huge amount of legacy data which exists on a wide variety of different platforms in different formats; but also because there is unlikely ever to be a format, protocol or platform which is suitable for all applications. This paper explores how heterogeneity is managed within WWW and CORBA (Common Object Request Broker Architecture). It shows how the concept of object wrapping is useful for incorporating external information services and providing an integrated view of that information. In doing this work we are developing a programming model for WWW which is very close to CORBA. We conclude with some recommendations about the kind of facilities middleware needs to incorporate to support easy integration of information services. We begin by comparing the CORBA and WWW models, and go on to show how object wrapping can be applied to both.
I am working on the Harness family of systems that are based on the object encapsulation of heterogeneous data. I am also working on the Information Repository Definition Language (IRDL) aimed at providing declarative support for building information repositories.
The InfoHarness information integration platform is aimed
at providing integrated and rapid access to huge amounts of heterogeneous
information independent of the location and representation of data.
InfoHarness provides advanced search and browsing capabilities without
imposing the burden of restructuring, reformatting or relocating information
on information suppliers. It is currently being productized at Bellcore.
GeoHarness is a public version of InfoHarness aimed primarily at providing access to NASA's geospatial data. The GeoHarness research is being performed under the USDAC project supported by a NASA three-year $2.3m Cooperative Agreement. The GeoHarness demo is currently available (click here to bypass the welcome/setup screen). The following research papers describing the ideas behind InfoHarness and GeoHarness are currently available:
I am working on the WONDER system that generically and reliably allows access to Legacy Datasets on Mainframe, as well as other database environments ... (We use a package called VARS, this allows for Variable persistence in non-homogenious environments; ie: SUN, IBM Mainframe, Novell, MicroSoft SQL Server to name a few of current implementations).
The VARS format allows us to transparently transport huge amounts of binary data from one platform to another, whether it be textual articles with graphics, or numeric data stored in simple tables or complex matricies with extended imbeded functions; all can be easily transported to the WEB server for conversion/integration into the HTML application.
What this boils down to is not only access to legacy data, but an actual
legacy application that has been migrated to the WEB.
Webmaster: Leon Shklar