keeping your cloud infrastructure healthy with the internet of things
TRANSCRIPT
Table of Contents
Chapter 1What Connected Machine Intelligence Means For Businesses Today
Chapter 2Challenges for Data Centers
Chapter 3Cloud Operations: Better With Predictive Analytics
Chapter 4VMware Can Help
1
4
6
10
14
IntroductionThe Internet of Things is Fast Becoming a Business Reality
1
Imagine if global logistics systems could
precisely track incoming shipments, giving
complete visibility into operations from
receiving raw goods to products leaving
the manufacturing floor. Or if sensors
on retail shelves could notify the local
warehouse when more stock is needed at
a particular store, enabling global inventory
to manage itself.
Welcome to the business Internet of
Things (IoT).
In the not-so-distant future, every object,
many elements in the natural world, and
even every person could have intelligence
embedded in them and be networked.
Everything will be capable of being
monitored, of contributing data to complex
Introduction
User-facing cloud services
Remote storage device (fixed or mobile)
Wide-area link
Service layer, cloud storage, analytics
Sensor Multi-sensor security device
T Local scanning device
Multiple independent remote network switches in
one location
Multiple remote wireless access points
User access via desktop device
User access via mobile device
Server (on-premises)
Local area links
Service provider access
The Anatomy of Machine to Machine Networking
Source: The Executive’s Guide to the Internet of Things, TechRepublic, 2013
2
systems, and of communicating with
everything else to drive a whole range
of possible outcomes.
At the heart of the IoT is data. Lots of
data. Think of all these devices sending
out constant signals. All that data needs
to be captured, stored, and analyzed.
Enterprises seeking to tap into this data
feed require more resources: more servers,
more storage, and greater networking
bandwidth, possibly even more data
centers. Furthermore, data mining and
data analytics techniques become part of
a Big Data strategy aimed at interpreting
these puzzle parts.
Cloud is absolutely necessary to make
sense of the IoT. Given how much data
the IoT will generate, businesses will
simply not be able to harness it without
a cloud infrastructure.
But cloud alone is not enough. The
IoT requires efficient and effective
management of the underlying cloud
infrastructure, as well as predictive
analytics to understand what the data
is telling you.
The solution? Cloud operations
management solutions with built-in
analytics capable of interpreting streams
of both structured and unstructured
data. That is required to make sense
of the IoT as well as keep the cloud
infrastructure running efficiently and
always available.
Introduction
4
The idea that the things around us would
be embedded with intelligent sensors and
able to communicate with each other—and
us—first appeared in the late 1980s. An
engineer at Procter & Gamble coined the
phrase “the Internet of Things” in 1999.
Today, the IoT generally refers to devices,
systems, and services that are connected
via a broad array of protocols, domains,
and applications.
More than three-quarters of companies
today are either actively exploring or
already using the IoT, according to research
by the Economist Intelligence Unit1, which
also found that within three years, almost
all companies expect to be using the IoT
in some capacity.
The number of connected devices is expected
to grow rapidly. IDC2 predicts the installed
base of things connected will be 212 billion
by the end of 2020. And according to the
McKinsey Global Institute3, the IOT could
have an economic impact of $2.7 trillion to
$6.2 trillion annually by 2025.
The IoT has the potential to transform
a broad range of industries. Here are just
a few of the potential applications:
Smart utilities
By 2025, placing sensors in the power
grid could result in savings of $200 billion
to $500 billion annually by pricing
electricity based on peak usage times,
according to the McKinsey Global Institute4.
Sensors could also be used to monitor
performance of utility networks in real
time, and could enable electric companies
to detect failures immediately and begin
putting fixes in place. One Midwest utility
Chapter 1
5
has been able to offer 100-percent uptime
to commercial customers through its
smart grid installation, according to the
McKinsey Global Institute.
Smart transport
Shippers are ahead of the pack when
it comes to using sensors to track raw
materials, parts, and goods through
production and distribution. Using
RFID tags that emit radio signals that
can be used to pinpoint the location
of products, logistics pros can identify
bottlenecks in production lines,
schedule trucks to take away finished
goods, or track the location of shipments
in transit in real time.
Smart jet engines
The jet turbine engine manufacturing sector
has been a leader in developing IoT products.
Because jet turbine engines are so expensive,
and because so much is at stake in ensuring
their safe operations, manufacturers have
built sensors into the engines that transmit
information about the engine’s condition to
the ground in real time. Now, instead of
scheduled maintenance, manufacturers offer
airlines “predictive maintenance,” which is
performed when an engine actually needs it.
Chapter 1
1 The Internet of Things Business Index: A Quiet Revolution Gathers Pace,” Economist Intelligence Unit, 2013.
2 “Internet of things: $8.9 trillion market in 2020, 212 billion connected things,” Larry Dignan, ZDNet, October 3, 2013. [IDC numbers]
3 Disruptive technologies: Advances that will transform life, business, and the global economy,” McKinsey Global Institute, 2013.
4 “All Things Online,” by James Manyika and Michael Chui, McKinsey Global Institute, September 23, 2013.http://www.mckinsey.com/insights/mgi/in_the_news/all_things_online.
7
Today, the IoT has developed to the
point where it is relatively inexpensive
and simple to embed microprocessors
in devices. The problem is, what happens
to all that data?
For the IoT to work, three steps are
necessary:
1. The data must be captured by an
intelligent sensor in a device
2. That data must be transmitted
across a network and aggregated
into a repository
3. The data must be analyzed so action
can be taken on it
Steps 2 and 3 are where today’s
bottlenecks exist. For starters, the sheer
volume of data generated by the IoT is
going to raise challenges for data centers.
It makes sense: the data generated by
the IoT will expand exponentially as the
number of connected devices multiply.
That’s because handling this explosion
of information is going to be both costly
and performance-challenging. Data
security and privacy issues are also likely
to grow in significance as more consumers
engage with IoT-based products. Beyond
storing, securing, and analyzing this data,
companies should also consider how they
manage the commercial sharing of the
data as the IoT becomes a platform for
trading information.
All this points to needing a central
repository of data (i.e. a cloud)—which
in turn requires a robust approach to
cloud operations.
The cloud provides a centralized place to
Chapter 2
8
collect and analyze data; however,
unabated demand for resources can
create numerous challenges for data
centers. It will force IT professionals
who manage data centers to “completely
rethink” the way they handle capacity
across all layers of the IT stack, according
to a recent report by Gartner1.
Cloud can bring together disparate data
streams. It can handle the enormous
volumes of data that the IoT will create.
It can scale as needed, providing massive
cost efficiencies compared to the
traditional data center.
But although the cloud might work for
taking care of Step 2 (finding a central
repository to store all this data), what
about Step 3—taking away actionable
insight from all this data?
Advanced analytics—specifically, predictive
analytics—are also needed. Predictive
analytics is a type of advanced analytics
that makes predictions about unknown
future events by analyzing current and
past data. Predictive analytics can also help
companies understand what is happening
from indirect evidence.
Techniques from the fields of data mining,
process modeling, artificial intelligence,
and machine learning are applied to
uncover patterns in both structured and
unstructured data to determine what is
an “anomaly” versus what is “normal.”
The indirect capabilities of predictive
analytics shouldn’t be underestimated.
For example, sensors on an airline’s
turbine engine send in data that
shows the aircraft is burning fuel at
Chapter 2
1 “Gartner Says the Internet of Things Installed Base Will Grow to 26 Billion Units By 2020,” Gartner Inc., December 2013.
9
a higher-than-normal rate. Perhaps a jet
engine has a failing part or simply needs
maintenance?
The combination of the IoT and predictive
analytics is a particularly potent one
because it allows enterprises to anticipate
issues. With intelligence embedded in
the everyday objects of business, and
the ability to leverage that intelligence to
project likely future events, enterprises
are transformed from being reactive to
proactive, whether it involves keeping
cloud infrastructure healthy and avoiding
costly outages, anticipating when
equipment needs maintenance before it
breaks down, or foreseeing inventory or
raw goods shortages before that impacts
manufacturing.
Chapter 2
11
As IoT applications become more integral to
businesses, keeping the cloud infrastructure
that supports them healthy becomes vitally
important.
Despite the accelerating popularity of
virtualization and cloud technologies, these
infrastructures have turned out to be much
more complex and difficult to manage than
generally anticipated. IT professionals must
be able to see into all the different layers in
a cloud. Problems can occur in an internal
physical network, or in a virtual network
managed by a virtual switch. CPU utilization
can be so high that it slows down service
on either a physical or virtual machine. But
a high CPU utilization on one application
could be normal for another application,
so you need a way to dynamically make
that determination.
Additionally, each possible root cause to
a problem requires a completely different
type of remediation. Predictive analytics
can ensure higher availability, better
capacity management, and better
compliance by helping IT understand
the complexity, dependencies, and
interrelationships in all these different
layers. Even the most experienced network
or storage engineers find it difficult to do
this with traditional tools.
What’s needed: a best-in-class cloud
operations management solution
with predictive analytics that pulls
together all these disparate elements,
giving businesses a unified view of their
operations. Since operational data—both
structured and unstructured—flows in
from all layers of the cloud, having this
Chapter 3
12
centralized view is critical to meeting the
demands of the IoT.
Detecting the cause of problems and
remediating them immediately is also
critical for maintaining a healthy cloud.
A cloud operations management solution
provides the ability to perform root-cause
analyses on any issues that occur in a
broad range of independent resources—
storage, network switches, and the like.
A centralized view provides one diagnosis
of the problem, reducing mean time to
identification (MTTI) and suggesting a fix.
With predictive analytics, such a solution
could even detect potential issues before
they become full-blown problems.
A leading cloud operations management
solution provides the following:
• 360 degrees view into environment.
Businesses get complete visibility
into their cloud environments
without having to manually compile
a list of symptoms from a variety
of different cloud components. A
cloud operations solution provides
self-learning analytics, dynamic
thresholds, and smart alerts, which
improves availability and uptime of
production systems.
• Improved compliance. Cloud
operations make it easy to meet IT
Chapter 3
13
policies and regulatory mandates
such as PCI, HIPAA, and SOX. With
this enhanced visibility, you can better
enforce such policies and mandates
across all infrastructure assets. It
also correlates performance with
configuration and compliance
data to have a single operations
console across private, public, and
hybrid clouds.
• Lower costs. By automating many
of the previously manual cloud
management tasks, businesses
improve team efficiency and achieve
overall greater operational efficiency.
This results in improved virtual
administrator-to-infrastructure
ratios, and significantly reduces
OpEx. CapEx can also be cut by
reclaiming overprovisioned capacity,
and by forecasting future needs more
accurately.
Sounds overwhelming, but predictive
analytics in a cloud operations solution
can also help make sense of all the
data coming in from the IoT. Yet, only
13 percent of enterprises surveyed by
Gartner1 in March 2013 had a predictive
analytics system in place. That will change:
73 percent of companies intend to increase
spending on predictive analytics in the
next two years.
Chapter 3
1 “Gartner Researchers: Predictive Analytics to Gain Traction in Business,” by Ian B. Murphy, Data Informed.com, March 19, 2013
15
VMware is on the forefront of cloud
operations management solutions, helping
enterprises both large and small manage
their cloud infrastructures in a way that
reduces both OpEx and CapEx, minimizes
downtime, and enhances capacity planning.
The Internet of Things is about feeding
Big Data and using cloud operations
to harvest it. Cloud is required to
manage it cost-effectively at acceptable
performance levels. And predictive
analytics help businesses get the most
out of the data—both structured and
unstructured—that is generated by this
brave new world of connected machines.
Keeping cloud operations healthy
is critical. VMware is developing
solutions that can help companies
get a handle on the operations and
analytics challenges that will inevitably
accompany the growth of the IoT.
These include:
• Managing cloud operations
intelligently, from storage through
applications
• Getting a full 360-degree view
of all data
• Resolving cloud operations problems
faster using all the structured and
unstructured IT data available
The IoT promises to help organizations
reap huge dividends in operational
efficiencies; discovering new revenue-
enhancing business models; and making
customer service improvements, among
other benefits.
Chapter 4
16
But these benefits will only be possible with
cloud—and a cloud operations management
solution that enables organizations to
keep their cloud infrastructures healthy.
VMware stands at the vanguard of this
next evolution of the Internet, and is
making the investments designed to help
businesses thrive during the transition.
Chapter 4
To explore cloud operations solutions further, please see the following resources:
VMware vRealizeTM Operations Insight
Forrester Consulting Report: Total Economic Impact of VMware vCenter Operations Management Suite
Management Insight Study: Businesses Experience Significant Operational and Business Benefits from vCenter Operations
Predicting the Health of Your Cloud
Understanding Real-Time Log Analytics