By Azmi Jafarey Computerworld | May 4, 2015 3:00 AM PT
Gartner has predicted that by 2020 the Internet of Things will grow
to 26 billion objects. (This excludes smartphones, tablets and PCs,
which will account for a separate 7.3 billion devices, Gartner adds.)
With these kinds of staggering numbers, there is a disruption in the
making -- and we CIOs need to be ready for it.
What are the "things" that we should be prepared for? At one level,
all sorts of familiar "dumb" devices -- the toaster, light bulb,
refrigerator, faucet -- will be 'smartened' with real-time sensors
responding to internal or external data, and will be able to
communicate.
Even more exciting are a class of totally new things ---
clothes with embedded sensors, earphones that measure heart rate and
temperature, smart watches that look for presence -- creating ripples of
data around both inanimate and animate objects. In the future it will
be tough to have a heart attack in private, or even to lose a dog.
As the IoT explodes with sensor costs coming down and
capability going up, one should expect, at least initially, a fair
amount of heterogeneity. There will be sensors that read and transmit,
but cannot be controlled. But in some cases there will be truer
bi-directional control. There will be different mechanisms for
communication between certain clusters and classes of things -- think
RFIDs: very different from how, say, a network of servers communicates,
or how kitchen devices might poll each other to compute a shopping list
for you. Standards will undoubtedly evolve,
but these are likely to be sets of standards for particular verticals
like medical devices or the automotive industry. IP-addressable sensors
and sets of sensors will make the "internet of the physical world"
happen.
But what will IoT be for? At the consumer level, we are headed
towards hyper-awareness at multiple levels: Personal health statistics,
environment optimization, social presence relay and detection,
behavioral prediction (like personal spend preferences and triggers) and
the like.
At the business level there will be two imperatives. For
those manufacturing physical goods, there will be the pressure for
"smart everything" -- what should be measured and why, how the data
should be used and when, and how such sensors can be made virtually
invisible.
The second imperative, and this will be for all: How can IoT
data be used to understand and optimize business processes, tools,
communications and buying and selling behavior?
Ultimately the game is one of competitive advantage, and using IoT to advantage will be a key skill required of CIOs.
For CIOs, the biggest challenges will be the quantity,
collection, analysis and purposeful utilization of near-real-time or
real-time data from numerous heterogeneous sources. Big Data has emerged
at just the right time for this. But the harvesting of data from
inexpensive sensors -- many of which will fail, be in error, need
recalibration in different environments or may not have been activated
-- will require intelligent handling of large data volumes.
Even without reference to IoT, SanDisk, as an example, is predicting a
14-fold increase in enterprise data to be managed by 2020. IoT
multiplies this challenge.
For companies looking to make an impact In the IT world,
there are clear and open frontiers to a wide array of both simple and
complex sensors to detect and correct device failures for IoT, or better
still prevent them. There are obvious needs for large volumes of sensor
data to get to the right place securely for analysis and optimizing of
objectives.
Accompanying this are concerns about privacy, security and
theft, especially since many of the 'things' entering a business may be
from multiple unknown consumer sources. (If you thought BYOD was tough
on CIOs, wait until your employee's shirt wants to adjust the
thermostat!).
At a more mundane level, as machines communicate
with software, today's concerns about user experience will be replaced
with concerns about efficiency and effectiveness of the back-end. New
licensing models are also likely to evolve -- clusters of machine
interactions differ significantly from users interacting with software.
The cloud will play a big part in machine interactions, particularly for
transmission, storage and analysis, since local read/transmit or
read/act/transmit will be the most common states.
We have highlighted heterogeneity as pervasive, at least at
the start of the Internet of Things. This does not mean that "seamless
integration" will not be expected of the CIO! Personal, home and work
environments will be expected to connect in rich ways without
interruption. APIs and extensions from the sensor manufacturers,
communications standards and protocols will all help. But the work ahead
is fairly formidable.
With all of the above, how does the role of the CIO and IT
change? I have a fundamental belief that I will re-state here: "If you
treat IT as a commodity that is what you will get. If you treat it as
the creative edge of your business, you have a weapon like no other."
Nowhere will this be truer than in how different companies
approach IoT. The laggards will view IoT as an infrastructure issue: Get
things talking, collect data, send it off for data warehousing and
analysis. The leading IT departments will embrace IoT as a green-field
for partnership with the business to explore how new business models and
predictive customer knowledge can evolve.
Azmi Jafarey has over 25 years of experience as
an IT leader. He was CIO at Ipswitch, Inc. for the last nine years. In
2013, he was awarded CIO of the Year by Boston Business Journal and Mass High Tech. He can be reached at azmijafarey@gmail.com.
Comments
Post a Comment