' in 1991, our
design and
'ersion of the
operational
It of Delta-X
management
pport, project
ntrol. Issues
>valuation are
[IVES
jects are:
ulti-database
it provides
ferent GISs :
elta-X); and;
/Ser as a front
icilitates the
1s databases,
fic datasets,
rage, format,
t, etc.
ities
nalities are
inctionalities
The clients
the database
d as a server
. The key
ace is that it
; server. The
ther Motif or
s capable of
note X client
which a user
er is able to
tify a dataset
r a dataset is
-set owners,
ng feature of
allow spatial
selecting the
the search.
Browser to
ine (E-mail,
information
Invoking Delta-X from a user workstations
first establishes connection to a default
DxServer. Delta-X provides functions for:
» Displaying selected or all features of a
vector map, the image or even a digital
elevation model of a coverage.
« Retrieving a specific data-set, either
vector, raster, etc. and translating to the
user's specific format.
» Administrative functions such as
registering and de-registering a database,
starting up the dxServer after failure,
authorizing users, etc.
+ Monitoring the progress of data-being
retrieved and the state of a ongoing
transaction.
Although Delta-X is not a Geographic
Information System, it has the potential of
being used in conjunction with other user
libraries and software for full GIS application
development. In particular, it provides the
opportunity for those interested in exploiting
parallel and distributed processing capability to
develop algorithms for their specific
applications.
3. OPERATIONAL FRAMEWORK
The work on this project stared with an
evaluation of the computing systems, GIS
software packages and database management
technologies. There were a number of
emerging technologies and we opted for
Commercial of the Shelf Components. The
functionalities of several DBMS' were studied
and Ingres RDBMS was selected. A spatial
topological data model was developed and
implemented.
The environment we considered for the system
implementation included:
* large organizations which often use
different (and incompatible) GISs to store
their spatial data in different databases,
further GISs are nodes connected to a
Delta-X server in a LAN,
* anetwork of wide area networks (WAN)
connecting multiple Delta-X servers; and
* individual users (Delta-X clients), who are
authorized to connect over a LAN or a
WAN to at least one Delta-X server.
A Heterogeneous GIS environment was
selected due to the wide variety of GIS
Software and hardware platforms available on
the market. The approach we took in our
349
solution was to consider the databases residing
in various agencies to be autonomous. The
availability of autonomous databases over a
WAN to users has many implications. First,
the dataholders must be part of a federated
environment and the data users must have
authorization to access data in the federation.
The federation must allow browsing of
databases so that users can identify the data
required and request transfers. Second, the
data holders manage and control their spatial
data and metadata, so that data will be up to
date, consistent, error free and secure from
unauthorized changes. In this environment,
the user can browse the metadata database for
data sources. Data transfer from "donor' to
user 'target' systems is executed by the Delta-
X.
4. MANAGEMENT PERSPECTIVE
AND TECHNOLOGY SUPPORT
One of the cost effective methods the Division
used to see if the users' needs are satisfied
was to implement a research prototype system
based on the requirement analysis and to test
the functional capabilities of the design. This
way, the concepts that were established earlier
could be tested on a real system to make sure
they work. This allowed our team to install
the prototype on a local area network (LAN),
make effective assessment of the development,
conduct performance analysis and establish
requirements such as storage and display
capability. The results of these tests made it
possible for further system enhancements
towards developing an industrial strength
prototype system that will meet the challenges
of data exchanges over a wide area network
(WAN) without rehashing much of the code
developed for the research prototype.
Moreover, the rationale that was followed was
to use existing technology as much as
possible. To illustrate this in the context of
Delta-X, we decided to use PHIGS, RPC,
Open Windows, MOTIF and INGRES among
other products. The integration of these
technologies into Delta-X was necessary for
the effective performance of the system. The
conformance to such standard packages helped
the system designers to make it highly
portable.
5. RESOURCES
A project of this nature needs a funding
strategy to encourage the developers to reach
the goals and even surpass them. The limited