36
ARTIFICIAL INTELLIGENCE
companies tend to add to existing systems rather than
optimising the existing system. All this puts extra pressure
on data centres and intensifies long-standing issues. So, to
keep pace, data centres will need to evolve.
AI in the
data centre
By Guy England, Director,
Lenovo DCG
www.lenovo.com/gb
Is the use of artificial intelligence the logical
progression for driving efficiencies in the data centre?
Over recent years, the appetite for artificial intelligence (AI)
has grown dramatically as businesses race to adopt the
latest technology to get ahead. From retail to agriculture, AI
is already being used in countless industries and it’s likely
that this trend will continue.
According to research conducted by Infosys, 76% of
IT decision makers believe that AI is a long-term strategic
priority for innovation. So it’s not surprising that worldwide
spending on cognitive and artificial intelligence systems
is predicted to reach $57.6 billion in 2021, or that the
technology is on the lips of media and influencers.
That said though, introducing technologies such as AI
isn’t simple and doesn’t happen overnight. Data centres
are vital for storing the data needed to power AI. So
naturally, growth in this technology will require more from
data centres which can be problematic, especially for those
facilities already working at high capacity. This forces data
centres to add more servers and hardware, which ultimately
becomes complex to manage.
As with much technology underpinning business, many
data centres run on inefficient legacy technology because
Going green
Arguably, the most pressing of these issues is energy
consumption as data centres now account for up to 6% of
global electricity use. This is a concern for business costs
and there is now increased pressure to take corporate
responsibility for environmental concerns. Organisations
including Greenpeace have been calling for tech companies
to become more energy conscious for some time, and now
they are honing in on data centres’ carbon footprint.
More data centres are moving to renewable energy
sources, although this isn’t feasible for many – particularly
smaller ones. Switching to renewable energy only addresses
part of the problem, and as the demand for computing
power continues to grow all options should be explored.
AI offers opportunities to improve efficiency and reduce
energy consumption by using existing data and real-
time monitoring. With AI, workloads can be distributed
across servers to maximise productivity and solve network
congestion issues. AI can also be used to control the data
centre environment in real-time, such as cooling systems, to
reduce energy consumption. Google is already implementing
AI to monitor the data centre environment and reported
that DeepMind AI was responsible for reducing Google’s
data centre cooling bill by 40%.
Step up security
Another critical area under pressure in the modern data
centre is security. In the cybersecurity battle, the stakes for
data centres are particularly high. They’re highly complex
infrastructures and have extensive levels of encryption, so
preventing data breaches in an ever-evolving IT landscape
requires constant vigilance.
The solution again may lie in AI. Implementing AI
systems can offer a more flexible and sophisticated solution
to data security as well as the possibility of reducing
reliance on human intervention. The nature of AI allows it
to adapt far more quickly than humans as well reducing the
man-hours spent on round-the-clock monitoring issues and
decrease the risk of human error.
With all these data centre opportunities unlocked by AI,
the concept of a human-free data centre is closer than many
imagine. Some companies like Litbit are already trialling-AI
driven robots to assist with data centre management and
hardware maintenance further cementing the ties between
AI and data centre.
Looking forward
The demand for AI won’t be slowing down any time soon.
This will only increase the demand for the physical space to
store the huge quantity of data needed to run AI, which will,
of course, increase the demand for data centres. Data is the
lifeblood of AI meaning the increase in data centre capacity
in the future will be mandatory.
Getting a flexible, future-defined data centre
infrastructure in place is crucial to prepare for AI. Once the
foundation is in place, servers can learn from the data they
process, creating the cycle that improves AI and data centre
ecosystems endlessly. n
www.networkseuropemagazine.com