Deployment architecture decisions

Before installation, you must know how many hosts are needed in your Nuance network and which components to run on each. Your design gets incorporated into your license file: if you change the deployment architecture in the future, it is likely you'll need to re-generate the license file.

Basic decisions	Description
Minimal installation	Install all components on a single host. This is called an All-in-one architecture. This deployment is good for initial testing, beginning development, and small deployments. You do not need additional planning: proceed to Install third-party software.
Expanding capacity	Install components on multiple hosts, and run different components to on each. This deployment expands capacity for optimal performance, and for larger volumes of traffic. Determine which components to run on each host. See Distributed architecture. In addition, you can expand capacity after the initial installation by running more instances of any component on any host. (This does not require additional installations.)
Managing the system	Decide how to administer and operate the system. You can provide your own tools or install the Nuance Management Station.
Estimating load	Be aware of the load placed by each component on each host, especially when preparing for testing and production. See Required hardware and memory.

Note: Most customers design their installations with help from Nuance. The Professional Services organization can estimate memory, processing power, and network loads based on your application's purpose and estimated volumes. Further, you might evolve the topography after testing the application and conducting phased deployments. The important factor is that you have a plan to run specific components on specific hosts. Having this plan gives you the host and port information needed to connect the components together.