The computing architecture of eCognition PIPEMON is based on eCognition enterprise. It can be executed on standalone workstations with minimum installation time, or can exploit the computing power of distributed processors within an entity. The execution on these processors can be started and controlled from an intranet or internet based client.

This flexible architecture enables three main workflows:

1. The whole process can be executed on standalone workstations or parallel computing within one entity:

This approach gives maximum control on data and computation power, but requires the acquisition of data and complete software by one entity.

2. The process is executed on central servers, only the interactive process steps are performed on local workstations

This approach minimizes investments in data and software, but requires to have possibly sensitive data stored on external storages and reduces control on computational power.

3. The process is split in two parts:

• calculations on non sensitive data are performed on centralized data processing centers,
• upgrade with sensitive information is executed on local workstations together with the manual interaction to integrate soft knowledge and perform post processing.
  
  This approach reduces the investment in data, which can be used by other users as well and such lead to cost reduction (of course dependent on pricing model of data provider) and reduces investment in and maintenance of software (sharing of cost for modules on server).