§Start small with an implementation which addresses the
most critical needs,
defined either geographically or by category of information
□Focus
first on known critical data‐elements
first to ensure that you have the capacity, ability, and availability of those items, before
expanding
□Take
a lesson from the search engine industry – great systems are built slowly over time from small, hardened implementations
□Ensure
that core competencies are answered before trying to do everything
§Build for scalability
□Avoid
situation where a system is built to perform very well for a test setup, but does not scale well in the real-world
□Leverage
technologies such as clustered databases, virtual warehousing, virtual servers, etc.
§Use multiple servers to distribute load for real time
databases
□Databases
can grow quite large very quickly
□Easier
to solve the problem once approximate data sizes and elements have been defined