Smallworld's architecture is designed to take advantage of
client/server-based computing, but is careful to keep the
amount of data that must be passed over the network to a minimum.
This is essential with the large volumes of data associated
with spatial information processing.
It does this by smart caching in the Smallworld client.
The result is that a large number of users can be supported
without difficulty on a single local area network.
For wide-area networks, each site can use a local cache
to ensure that data is transferred from the master database
only once, however many users need it. Alternatively, if communications
links are unreliable, a replica of the master database can
be used, and synchronized periodically.
Whatever the solution chosen, users are ensured unrestricted
access to the information they need.
A tiered approach
Smallworld 3 also structures information so that the time
taken to access it is largely independent of the volume of
data being handled, and it ensures that as much as possible
of the complex processing of spatial information takes place
in the client. The more clients there are, the more processing
power is available.
This means that a Smallworld installation has almost unrivalled
scalability, able to grow to support first hundreds and then
thousands of users at once.
The core Smallworld 3 technology can also be used to build
application servers that support thin clients such as field
units and Web browsers. This extends the system's inherent
scalability to provide a complete solution for deploying applications
across the enterprise.
SmallworldWeb is built
using these principles to provide access to corporate information
from as wide a user base as possible. SmallworldWeb is much
more than just a document viewer - it's a sophisticated extension
of Smallworld 3 designed for the corporate Intranet. It lets
users contact and interact with a live Smallworld 3 application
using industry-standard Web browsers.