FEATURE: ENTERPRISE DATA CLOUD /////////////////////////////////////////////////////////////////////
The public cloud has now been experienced
by a vast number of organisations, who
value its simplicity and elasticity. However,
unexpected operating costs and vendor lockin
have prompted enterprises to opt for some
other cloud infrastructure models that would
allow both choice and the ability to run
demanding workloads no matter where they
reside and originate, from the Edge to AI.
Same problems, new challenges
The most valuable and transformative
business use cases – whether it’s IoTenabled
predictive maintenance, molecular
diagnosis or real-time compliance
monitoring – do require multiple analytics
workloads, data science tools and Machine
Learning algorithms to interrogate the same
diverse data sets to generate value for the
organisation. It’s how the most innovative
enterprises are unlocking value from their
data and competing in the data age.
However, many enterprises are struggling for
a number of reasons. Data is no longer solely
originated at the data centre and the speed
at which Digital Transformation is happening
means that data comes from public
clouds and IoT sensors at the Edge. The
heterogeneity of datasets and the spike in
volumes that is leading to real-time analytics
means that many organisations haven’t yet
figured out a practical way to run analytics
or apply Machine Learning algorithms to all
their data.
Their analytic workloads have also been
running independently – in silos – because
even newer cloud data warehouses and
data science tools weren’t quite designed
to work together. Additionally, the need to
govern data coming from disparate sources
makes a coherent approach to data privacy
nearly impossible, or at best, forces onerous
controls that limit business productivity and
increases costs.
Back to the drawing board
Simple analytics that improve data visibility
are no longer enough to keep up with the
competition. Being data-driven requires the
ability to apply multiple analytics disciplines
against data located anywhere. Take
autonomous and connected vehicles for
example, you need to process and stream
DATA IS NO LONGER SOLELY
ORIGINATED AT THE DATA CENTRE
AND THE SPEED AT WHICH DIGITAL
TRANSFORMATION IS HAPPENING
MEANS THAT DATA COMES FROM
PUBLIC CLOUDS AND IOT SENSORS
AT THE EDGE.
Romain Picard, Regional Vice President
South EMEA at Cloudera
real-time data from multiple endpoints at
the Edge, while predicting key outcomes and
applying Machine Learning on that same
data to obtain comprehensive insights that
deliver value.
The same applies, of course, to the needs
of data stewards and data scientists in
evaluating the data at different times in
the processing chain. Today’s highestvalue
Machine Learning and analytics use
cases have brought a variety of brand-new
requirements to the table, which have to be
addressed seamlessly throughout the data
lifecycle to deliver a coherent picture.
Enterprises require a new approach.
Companies have grown to need a
comprehensive platform that integrates all
data from data centres and public, private,
hybrid and multi-cloud environments. A
platform that is constantly informed about
the location, status and type of data and
can also offer other services, such as data
protection and compliance guidelines, at
different locations.
The rise of the enterprise
data cloud
Since enterprises undergoing Digital
Transformation are demanding a modern
analytic experience across public, private,
hybrid and multi-cloud environments, they
are expecting to run analytic workloads
wherever they choose – regardless of where
their data may reside. In order to give
enterprises flexibility, an enterprise data
cloud can empower businesses to get clear
and actionable insights from complex data
anywhere, based on four foundational pillars:
1. Hybrid and multi-cloud: Businesses
have grown to demand open architectures
and the flexibility to move their workloads
to different cloud environments, whether
public or private. Being able to operate with
equivalent functionality on and off premises
– integrating to all major public clouds as
well as the private cloud depending on the
workload – is the first ingredient to overcome
most data challenges.
2 Multi-function: Modern use cases
generally require the application of multiple
analytic functions working together on the
same data. For example, autonomous
vehicles require the application of both
real-time data streaming and Machine
Learning algorithms. Data disciplines
– among which edge analytics, streaming
analytics, data engineering, data
warehousing, operational analytics, data
science and Machine Learning – should all
be part of a multi-functional cloud-enabled
toolset that can solve an enterprises most
52 INTELLIGENTCIO www.intelligentcio.com