The Shared Resource Interconnect (SRI) connects modules with native SRI interfaces such as CPUs and high performance modules such as the DMA with memories and peripherals. The SRI crossbar connects all the agents in one SRI domain with a deterministic performance objective. The fabric is extended with bridges, to connect one SRI domain to another as well as to other components and interconnects operating on different protocols.

The SRI crossbar supports concurrent transactions between different SRI master and SRI slave agents. When transacting to the same slave, the master can achieve very high utilization over the link by employing pipelining. In addition, a master can conduct read and write transfers to different slaves in parallel.