I know only this architecture (not sure with A and B type)
The Relays just contain L1 and L2 of the eNodeB , not the complete LTE eNodeB stack.
-> Layer 3 (RRC) is terminated in Relay Node for the (Relay Node UE) interface.
-> L1/L2 signaling (CQI, HARQ, etc) is performed between Relay Node and UE.
-> Relay Node has their own PCI (physical cell identity).
There are two different types of Relay's.
Type-I relay will help the UE to access the eNodeB that are far from the coverage of the eNodeB. So the Type–I relay will transmit the common reference signal and the control information for the eNB and will help the eNodeB in extending the signals and service coverage area. Type-I Relay mainly perform IP packet forwarding in the (layer 3) and can make some contributions to the overall system capacity by enabling communication services and data transmissions for remote UE units.
Type-II Relay can help a local UE unit, which is located within the coverage of an eNB and has a direct communication link with the eNB, to improve its service quality and link capacity. So a Type-II RS does not transmit the common reference signal or the control information, and its main objective is to increase the overall system capacity by achieving multipath diversity and transmission gains for local UE unit Increase the throughput