Star Student Project
virtual interrupt latency may be important, for example, in real-time systems
based on CPU partitioning or preemptive scheduling. Simply combining two
independent interrupt coalescing schemes introduces additional latency for
delivering a packet. In the multiple tasklet solution, CPUs receive interrupts
and process the packets from dedicated queues or functions (distributed across
queues based on connections) in parallel. In the multiple Star Student Project kernel threads or high-priority threads
solution, each thread of the back-end driver can be bound to a specific virtual
CPU.