INFOGRAPHIC
operations (AIOps) will be necessary to help
monitor complex environments across the
Edge, the core and the cloud.
2. Hardware-based composable
architecture will have less short-term
potential against commodity hardware
and software-based infrastructure
virtualisation. Continued improvements
in commodity hardware performance,
software-based virtualisation and
microservice software architectures will
eliminate much of the performance
advantage of proprietary hardware-based
composable architectures, relegating them
to niche data centre roles soon.
The advent of 5G is what AI-driven IoT has been waiting for. 2020
will see many players in the technology industry and business
community invest in building Edge Computing environments to
support the reality of AI-driven IoT.
These environments will make possible new use cases that rely on
intelligent, instantaneous and autonomous decision-making, with
low-latency, high-bandwidth capabilities. This evolution will bring
us to a world where the Internet will work on our behalf, without
even having to ask.
This AI-driven IoT innovation, however, will depend on a
massive prioritisation of Edge Computing, further disrupting IT
infrastructures and data management priorities.
As Edge devices move beyond home devices (such as connected
thermostats and speakers) and become more far-reaching (such
as connected solar farms), more data centres will be placed at
the Edge. Also, platforms such as Artificial Intelligence for IT
www.intelligentcio.com
Hardware-based composable architecture
is being hyped as the next evolution of
hyperconverged infrastructure (HCI). This
architecture enables CPUs, networking
cards, workload accelerators and storage
resources to be distributed across a rack-
scale architecture and to be connected
with low-latency PCIe-based switching. And
although composable architecture does have
potential, standardisation has been slow and
adoption has been even slower.
Meanwhile, software-based virtualisation
of storage, combined with software-based
(but hardware-accelerated) compute and
networking virtualisation solutions, offers
much of the flexibility of hardware-based
composable architectures today with lower
cost and consistently increasing performance.
This year, attempts to build a true hardware-
based rack-scale computing model will
no doubt continue and the space will
continue to evolve quickly. However, most
organisations that must transform within
2020 will be best served by a combination
of modern HCI architectures (including
disaggregated HCI) and software-based
virtualisation and containersation. n
INTELLIGENTCIO
31