A good low risk strategy is to start with a small service capacity and establish a plan for scaling up: adding servers, disk storage, and moving from single servers to clusters with network load balancing. Such a scale-up can usually be accomplished with no code changes and conceptually simple physical and operating systems upgrades or changes. Once a family of interoperable building blocks has been selected, the main design, implementation, support, and maintenance benefits are derived from reconfiguring them as needed to adapt to changing service delivery volumes. The most capable configurations will group functionality with different usage characteristic on dedicated servers or clusters. This allows the service level to be precisely tuned to the actual usage.
In these three chapters we examine deployment configurations for a navigation service provider supporting the full range of MAGIC Services with configurations ranging from a single server (the Small Space Server or the Position Proxy Server) to a fully disaggregated collection of dedicated function servers.