the past, capacity and throughput were the major challenges when targeting data
growth. Today, capacity and throughput are “commodities”, while latency has become
the new performance metric.
system approach breaks new grounds toward dramatically lower latency levels and
therefore much greater latency performance.
Defined Network platforms cannot solve the latency problem, because software
only solutions as well as converged and hyper-converged systems are not intrinsically
efficient and introduce additional latency in inter-node communications.
a world in which data growth is disruptive, the number of devices generating
and transmitting data will continue to increase. Therefore, response time is
levels of applications performance can be easily achieved with low latency
platforms, on which new applications like machine learning and artificial
intelligence can perform on much grander scales than ever before.
is the most critical performance factor because it directly affects system data
exchange time. In fact, latency means lost time; time that could
have been spent more productively producing computational results, but it is
instead spent waiting for I/O resources to become available.
is the “application stealth tax”, silently extending the elapsed times
of individual computational tasks and processes, which then take longer to
execute. Once all such delays are accounted for, the overall system performance
can drop significantly. This is why when building or buying a new system
latency performance matters – and not just at the storage level, but across the
system, through system inter nodes and software applications.
A3Cube’s Fortissimo™ Foundation utilizes hardware-based native memory mapping capabilities in combination with a sophisticated shared memory management architecture, which together provide a robust, cost effective, low latency, high bandwidth and extremely user friendly I/O access technology solution, with full support for direct memory and remote I/O mapping.