Below are note from the above titled session as presented by DC Tardy Sr. Program Manager, Microsoft at MMS 2012 on April 17, 2012...

• Speaking of their intended goals of simplifying the hierarchy- CAS, Primary, Secondary and DP. SCCM 2012 should take your existing architecture and help to simply it with the same capabilities
• Moving from file based replication to SQL based replication
• CAS – Central Administration Site (not a central “site” as in 2007 but a roll up point for reporting and administration)
  • You need to have a CAS for every site server.
  • 100k is the upper limit of clients to sit on one primary site
  • CAS should be installed on a separate server
• While the limit is 100k clients per primary, you may want to add another to reduce impact of primary site failure (but remember that means another CAS server too)
• However you no longer need to create primary sites for some of the reasons you did before. Specifically you don’t need to create one for decentralized administration, logical data segmentation, client settings, language, content routing for deep hierarchies.
• A secondary site server will let you manage upward flow of WAN traffic and provides tiered content routing for deep network topologies.
• A distribution point is to be used when BITS is not enough control for WAN traffic, if you want to leverage multicast capabilities or provide App-V streaming.
• There is only one kind of distribution point type now
  • This role can be installed on clients and servers
  • You can throttle and schedule
  • Built-in PXE and Multicast support
  • Specify drivers for content storage
  • Note that IIS is required on all distribution points
  • It can co-exist on a secondary site server or may be remotely connected.
• There is a single tool is used for content pre-staging. You can export content into multiple files or a single file and move it to another DP (when setting up a new server, sneakernet, etc.). It will do a file diff and get only changes when changes are introduced in the future.
• New feature: Forest Discovery: Discovers site server’s forest and any trusted forests to get all AD information to pull them in for use in SCCM. If you have credentials you can add forests that are not trusted as well. Discovery gets: domains, IP subnets and AD Sites and supports boundary creation.
• There was then a demo and then lots of detail about the replication model and how the CAS, primary and secondary sites push data up and down. Data can be routed between secondary sites if you wish.
• DRS (Data Replication Service) handles on-way and bi-directional replication. SQL is used to handle replication now.
• You can view server architecture now in a Bing map with a new “Geographical View” feature
• You don’t need to create MOF files any more, you can browse WMI and choose what you want directly. If you have invested in MOF file efforts already, they can be imported.