High-performance networks have developed significantly over the past several
years.  Many of these networks are capable of remote direct memory access
(RDMA) communication that involves movement of data between processor memories
without memory copies or remote host processor involvement.

Current high-speed networks deploy powerful communication processor(s) in the
network interface card (NIC) to handle processing of multiple incoming and
outgoing messages without interrupting the host processor. For example, the
Quadrics Elan4 network interconnect can do two overlapping DMAs and allows
multiple outstanding read/write transactions from and to the network
interconnect.

Switches used in the modern interconnects also have developed significantly in
the last several years, facilitating multiple communication paths between
network endpoints and hence roviding increased levels of concurrency and
redundancy communication between network endpoints. As a result, the modern
networks are very capable of handling simultaneous and concurrent data
movements.

ARMCI allows for its applications to utilize concurrency by its low-overhead
implementation of one-sided calls
