Can You explain the Data Flow Architecture used in AAM?

A customer typically uses Adobe Audience Manager as follows:

• Customers log into the Audience Manager web interface to configure rule sets using defined attributes, such as signals, traits, and segments.
• Then rules are then sent to a control database and distributed throughout the Audience Manager network via proprietary configurator software.
• When a user visits the customer’s website, Audience Manager then places code on the computer to create behind-the scenes communication and data collection with Adobe and its partners.
• Edge data centers, hosted on Amazon Web Services (AWS), process all incoming data into segments, based on the customer-defined rules.
• These data centers collect the external data, marry the incoming data in the DCS to any prior user data stored in the PCS, process the information, add cookie values to each user, and send report data.
• Once Audience Manager processes the data, it returns it back to the customer via a distributed search index that provides real-time segment estimates to the customer.
• Audience Manager also enables customers to track partner data. Partners are third-party entities that share inbound and outbound audience data with the customer.
• As with other incoming data from the customer’s website, Audience Manager collects inbound data from the customer’s partners.
• The DCS receives all incoming external data from both partners and the end user, and the PCS in the Edge data center attempts to match the stored two weeks of historical prior trait data to incoming DCS data, thereby building a more robust web surfer profile. When required, the DCS
  pushes data to an external real-time reporting function.

Adobe Audience Manager Data flow