GE Digital

Unlocking Business
Value Through
Industrial Data
Management
Author: Shefali Patel
Contributing authors: Jeremiah Stone, Scott Duhaime and Venkat Eswara

Contents
03 The Opportunity and Potential of Industrial Big Data
05 Industry Mega Trends and Challenges for Big Business
08 GE’s Approach: Maximizing Value from Machines and Enterprise Data
12 Rethinking Data Management—Control to Cloud
14 Customer Case Studies
16 Conclusions

2 of 17 Unlocking Business Value Through Industrial Data Management

The Opportunity and Potential
of Industrial Big Data
The industrial network of connected and in-memory computing, distributed Based on research from the Center of
machines with data analytics is creating databases and data science, along with Digital Business at MIT Sloan School of
a new frontier of competitiveness proliferation of sensing and controls, have Management, data-driven decision-makers
for information-centric modern given rise to "big data"—fundamentally (DDD) demonstrate 4% higher productivity,
organizations. We are citizens of changing the ways in which individuals, 6% greater profitability and 50% higher
a data-driven century, in the early organizations and machines interact market value from IT, according to Eric
stages of a digital industrial revolution. with their intelligent environments. Brynolofsson. Data-driven decision-
Orchestration of digital machines will Convergence of these physical and making also improves other performance
lead to unprecedented change in the way digital improvements has paved a way measures such as a) return on assets,
companies perform, compete in their for the Industrial Internet, driving b) return on equity, c) asset utilization
industries, and unlock new value from their efficiency gains across the production (output per total assets) and d) market
data and business models. value chain. value (market-to-book ratio).
Technological progress in information At the heart of this revolution is an
systems, broadband, communications invaluable resource: data that can be
and artificial intelligence is generating converted rapidly into insights, leading
enormous wealth across industrial and to smarter business decisions and
consumer sectors. Advancements in cloud increased automation.

3 of 17 Unlocking Business Value Through Industrial Data Management

We are citizens of a data-driven century,
in the early stages of a digital industrial
revolution.

With lower cost of sensors, we can now tools proliferate, they are changing how At GE, we are developing a better
measure many things that we could we behave, automate machines and reap way—a better way to manage industrial
not before, and therefore control our benefits of digitization.1 big data that triggers insights. We are
technology more precisely and less However, despite the promise of big in the early stages of a long journey
expensively than ever before. Utilization data, companies struggle to exploit its of discovery and invention, taking a
of Internet technologies has made it value. Why? Abundant data by itself solves longer-term view to strategic data
easier to access data. Data can help us nothing. Its sheer volume and variety management and its technologies that
make better predictions and take smarter exceeds human capacity to configure it translate to business advantage.
actions. We can be collectively objective, efficiently. Inherent challenges tied to
rather than individually subjective. We can evolution and integration of information
do so in areas where we formerly acted
based on intuition and assumption rather
and operation technology make it difficult
to glean intelligence from unorganized
Abundant data by
than by data and analysis. As big data data, compromising digital literacy. itself solves nothing.

What if invisible insights into your business became visible? What if terabytes of data pulsing through your operations were
captured and stored securely in the field and in the cloud so that it could be accessed in real time? Imagine an enterprise
world with a single source of truth—a panoramic view of your entire fleet and operation with the ability to zoom into
powerful telescopic detail in a cost-effective way, and a workforce focused on resolving issues, innovating and collaborating
across silos.
This paper is about Industrial Data Management: its challenges and opportunities, with a central focus on foundational technologies that
lay the groundwork to win in the future of the Industrial Internet.

4 of 17 Unlocking Business Value Through Industrial Data Management

Industry Mega-trends and
Challenges for Big Business

Before we discuss industrial big data solutions, let us review some

of the mega-trends that are impacting the state of the market.
They create both challenges and Disruption of industries The future of work
opportunities for progress, while shaping From domains as diverse as entertainment, The nature of how work gets done today
our choices for technology investments. retail and transportation, digital is also evolving with demographic,
These major market trends fall into innovations have disrupted industries all globalization and technology changes.
three different areas, all relevant to around us. Missing the digital beat is falling The composition of today’s workforce
digital transformation: prey to disruption. Disruption happens is ever so diverse, with digitally savvy
slowly, as centralized incumbents millennials working side by side with more
The rise of big data and the Internet get displaced by nimble companies experienced knowledge workers.
of Things that have commercialized cheaper, As technologies further penetrate the
With the cost of computing, bandwidth more convenient and widely available workplace, they will create new ways
and sensors decreasing multifold in recent digital technologies to underserved of organizing work and dispersing
years, there has been an explosion of consumers. It creates new markets. knowledge across distributed teams.
embedded devices that can communicate At GE, we have disrupted ourselves by With virtually every device in the workplace
with one another and churn out volumes infusing an ethos of entrepreneurship, emitting data, it will take new ways to
of data. Big data can be defined as high- reinventing our business model, leading the manage data-driven organizations and get
volume, high-velocity and / or high-variety practice of the Industrial Internet for our the most of out of our human capital.
information assets that demand cost- customers, and creating GE Digital.
effective, innovative forms of information
processing that enable enhanced insight,
decision-making and process automation.2 Why IoT Is Happening Now
The physical world is being digitized. Smart
objects linked through wireless networks Cost of Sensors Cost of Processing Cost of Bandwidth
Avg. Cost $1.30
that carry information are forming a
system called the Internet of Things (IoT). 0.60 60x 40x
The basis of IoT, the real-time dynamic
analysis of data, is challenging business Over the past 10 years Over the past 10 years Over the past 10 years

models built on static and rigid

information architectures. We need fluid Cost of RAM Disk Storage Cost of Flash Memory

business models based on agile software

platforms to catch this technology wave 21x 17x 646x
and maximize its value before it diminishes.
Over the past 10 years Over the past 10 years Over the past 10 years

Source: Goldman Sachs Global Investment Research, John C. McCallum Research, TCG Advisors
Industry and
Customer Challenges
Taken together, these trends are carrying us forward. They are also
creating new challenges for business leaders. We have grouped them
into technology, business and organizational challenges.

Technology challenges
Islands of disparate data challenges. When the data is Technology integration
scattered throughout plant and the
Keeping up with a flood of Management of industrial technology
enterprise, integrating and analyzing
information is difficult. Most has traditionally been split between
it manually becomes resource-
companies struggle with data deluge two separate fields: information
intensive and tedious. This has an
driven by lower cost storage, sensing technology (IT) and operations
opportunity cost. By the time
and communications technologies. technology (OT). IT worked from top
data is organized, its value
But a few have figured out how down, deploying and maintaining
may have been lost and the
to exploit their data. Big data data-driven infrastructure largely to
personnel too fatigued to derive
that is neither structured nor management side of business. OT
any insights.
contextualized is strenuous built from ground up, starting with
to cost-effectively store and machinery, equipment and assets,
analyze in its entirety through To extract meaning and and moving up to monitoring and
traditional computing approaches. value from data, new industrial control systems. With
systems are required to smarter machines, big data and
What causes handle the challenges the Industrial Internet, worlds of IT
data islands? posed by the volume, and OT are converging. Traditional
velocity and variety of enterprise data management, such
Data islands are created as a
as ERP or CRM, is being dwarfed
byproduct of operational and project these big data sets.
by operations data due to sheer
moment-in-time decisions not
made in the context of a larger data
strategy. Layering of legacy systems Islands of Disparate Data
conjoined with newer technologies
and lack of governance for data When everything’s an island, it’s hard to be intelligent
systems also results in data-
related islanding of internal
departments and work groups.
It can lead to limited purview
and inhibit collaboration. Data
gets siloed, whether it is enterprise
data, equipment data inside an
organization, or data across
different organizations.
This fragmentation makes data
discovery difficult and presents
Unplanned Downtime Wasted Analytics Lost Productivity Missed Opportunities
complex technical and organizational
volume and variety. But most together and find common ground
of this data is still in the dark. IT to develop a new infrastructure.3
and OT, developed separately
with independent systems
architectures, need to come

Business challenges
New sources of revenue There is a need to link analytical to be able to monitor assets in real
systems to operational systems. time and ensure all assets (across all
and profitability
Today, most business analytics do plants) are performing at an optimal
With a highly volatile market not support any connection back to level. They need increased visibility
environment and costs of the originating systems of the data. and better insights that can be acted
maintaining aging infrastructure, Analytics are on an island, as well, upon. This enables them to detect
companies are continually inhibiting the ability to take action in anomalies and fix issues before
challenged to sustain their a reliable and effective way due to they occur, yielding no
profitability by finding new sources the onus on the individual to connect unplanned downtime.
of revenue. Manufacturers are the worlds in their brain and connect Asset Performance
seeking ways to lower capital the systems and workflows via their Management and operations
expenditures, and they need own initiative optimization software can
a single source of the truth to
provide operators with
help them make the right decisions
Asset-level visibility answers on what equipment
for improved performance, while
Improved capacity utilization is one is most important, how it
mitigating risk from
of the great benefits of state-of-the- should be maintained and how
unexpected incidents.
art information systems. To achieve unexpected failures can
production targets, operators need be avoided.

Organizational challenges
Aging workforce and of younger employees. Preparing attacks pose a range of threats—
for this impending change by using from personal devices to corporate
knowledge capture
digital technologies can ease IT systems—making individuals and
Aging of the workforce is impacting the transition.3 institutions vulnerable to financial
a number of industries. Retirement and physical harm.
of experienced workers is
Cyber security There is growing awareness and
expected to create a skills gap.
As billions of assets get smarter, paranoia among stakeholders, and
While younger generations of
network and store information an urgency to mitigate these risks.
workers will bring new skills, it
on the cloud, they will be exposed Vendors are deploying solutions
is crucial that the knowledge
to digital privacy risks. Just in the to prevent these cyber events and
and experience accumulated
last year, there have been several protect against digital crime. As we
by more senior workers is
cases of data breaches causing invest in digital technologies, cyber
captured and made accessible
significant damage to all parties security capability must be part of
to the new workforce. Inability to
involved (companies and their supply the selection criteria.
institutionalize this knowledge can
be detrimental to the apprenticeship chains, as well as consumers). Cyber
GE’s Approach: Maximizing
Value from Machines and
Enterprise Data
Industrial Internet technologies can turn the challenges
discussed earlier into opportunities for improved productivity.
Many industrial companies have already
started their digital journeys towards
Reaching maturity involves five
stages with corresponding technology
5 stages of Industrial
Industrial Internet maturity. Technologies components that allow an enterprise to Internet maturity:
including enterprise data management connect, monitor, analyze, predict and connect, monitor, analyze,
and predictive analytics that we have optimize their assets and operations.
been deploying for our customers are This can be accomplished on the premises predict and optimize
now seeing double-digit performance and in the cloud, depending on the current
gains across the following sectors: power configuration of your data architecture.
generation, oil and gas, transportation and
health care.

8 of 17 Unlocking Business Value Through Industrial Data Management

The first step of the Industrial
Internet maturity model is to GE Industrial Data Maturity Model
connect all critical assets across the
enterprise. This is not a trivial task as
we are referring to hundreds of discrete
instrumented components with different
communications, networking protocols
and underlying data formats, all generating
terabytes of data. It requires a highly
scalable and fault-tolerant software
Optimize
system that enables cost-effective
Predict

Value
data storage, data visualization and
Analyze
analytics. This foundation of the
Industrial Internet starts with Monitor
data management. Connect
Historically, it was possible for a single
machine or a handful of machines to
receive and store data. However, the Capability

growth in sheer volume of devices and

sensors coupled with the desire to perform
rapid data mining on larger amounts of traditional archive-and-ignore model of environments. While very efficient,
historical data (requiring that they be data management—and at a lower cost. most data historians in the marketplace
kept in memory or on a disk) necessitated They make it possible for structured are single-server solutions, limited
the development of systems with and contextualized data to be online by the memory and disk capacity of a
new technical approaches. Enterprise and available anytime for analysis single machine.4
Historian, designed and built by GE, and mining. They not only give data Our next-generation historian
addresses this challenge in a cost- scientists access but also a visual efficiently stores and performs
effective way.4 representation of what is happening to historical analysis on hundreds of
an asset, making it easy to spot terabytes of time-series data. It is
To illustrate this point, let us first think unusual conditions.4
through how industrial data becomes seamlessly scalable in how much data it
industrial big data. As evidenced by figure Data historians are special-purpose can store, with the flexibility to address:
1, a single machine in the manufacturing of database applications. They are designed • Consistent data collection, storage,
baby care products, which generates 152K to efficiently store and analyze large federation and data modeling across
data samples per second, provides 4 trillion quantities of time-series data with asset, site, plant and enterprise
samples a year. Added to this are other ultra-fast read and write performance.
They are capable of storing up to millions • On-premise execution with the ability
streams of data such as geolocation (GIS),
of data points per second, capturing to manage data at rest in the cloud
alarms, and maintenance and shift logs.
Big data systems that support handling streams of data in real time from sensors • Fast data modeling with data
of increasingly fast and large volumes located across a manufacturing facility, discovery models to process a variety of
of heterogeneous data can replace the power plant or other such sensor-rich data sets (structured, semi-structured
and unstructured)

Increase in Data Volume Over a Year

545 MILLION
4 BILLION
samples per hour
samples per shift

13 BILLION
samples per day

9 MILLION
samples per minute
4 TRILLION
samples per year

152,000
samples per second

Figure 1
Next, let us review underlying technical capabilities
and the business outcomes of data historians.

Data collection Historian, it cannot be edited or changed (except during a re-run

Data historians collect time-series data in real time from of data collection). It is secured in a highly trusted enterprise
distributed, disparately structured sources. With GE’s Historian, data store.
the first value customers have realized is through simplifying Data services
data collection and aggregation functions. A mix of redundant, A central enterprise data store built on near-real-time data opens
fault-tolerant collection options that can be paired with sites at up new opportunities to use current data in support of a range
different levels of automation, coupled with the ability to embed of systems and procedures. A key part of Historian’s enabling
pre-handling and calculation functions enable a more flexible way technology is a services-orientated architecture. It provides a
of maintaining data flows to the central operational data store set of data models that can be adapted to reflect common types
(ODS) with lower demand on IT and engineering resources. of assets (and related events, personnel and materials) deployed
Data storage across various operations. Source data elements from existing
Our customers are realizing value in the way GE's Historian stores automation and systems can then be mapped to the items
data. With no compression at all, it offers much higher disk space defined in the model. Users commonly take advantage of this
efficiency than a relational database (RDB). When using a 1% modeled Historian data to:
dead band compression, it delivers even greater efficiency for • Feed maintenance systems with asset usage data
enhanced performance and reduced maintenance. Efficient data • Deliver comparative views of asset and process health
storage and compression enable high performance and across like assets or installations, regardless of differences in
minimize maintenance. underlying automation and systems
Why is this important? Consider the big data example of the baby • Use data and calculations from Historian as the basis for
care line that collects 152K samples per second. Uncompressed, triggering a range of corrective or keep-running actions,
delivered through electronic work instruction systems like
this data could reach 40 TBs annually. With standard GE Data Workflow, or even as outputs back out to existing systems
Historian compression, this is easily reduced to 4 TBs. At a cost (SCADA being most common)
of $5,000 per terabyte, your storage budget for this one machine
shrinks from $200,000 to $20,000. In addition, once the data is in

Historian: Built for the Industrial Comparing the Disk Space Effiency
Internet and Big Data Solutions of an RDB vs. GE Historian†

Collection Storage Distribution

40
40

35
Effective bytes per sample

30

Process Data 25

Client based 20

15

10
6
5
Files 1.3
Stores 0
Analysis
Archives RDB Historian Historian
Context (no compression) (1%compression)

This data represents a specific test on 400,000 samples logged to a standard

†

RDB and Historian. Results will vary depending on the raw data set used and the
Server Data Web based RDB schema employed.

Figure 2 Figure 3
Data access: On-premise to cloud This also allows IT managers creative options for long-term
Another two considerations for IT leaders are: a) the cost of storage. For instance, one customer, who keeps 30 days of
moving large data files across the network and b) long-term data historical data at the plant level, aggregates all plant data onto
storage requirements, specifically determining where the final a central Historian temporarily, and then moves all this data to
resting place of data should be. an HDFS cluster for their “data at rest” strategy. This approach
GE Historian solves for both these problems. has financial benefits. A terabyte that costs $5,000 to store
in a traditional manner may cost $1,000 in an HDFS cluster.
Per our previous discussion on data compression, consider With the server-to-cloud collector capability, the costs of highly
moving uncompressed archives of 10 GB or more across various compressed GE Data Historian file drops from $20,000 to $5,000
networks, say from a control network, through a firewall, into in HDFS.
a DMZ (demilitarized zone) or a business network, then again
across another firewall boundary, into an external facing network
or public Internet for cloud storage. Then consider the 10x
compression of GE Historian and the impact to your network of
Time to value: From calendar time
transferring 1 GB of archive files. to watch time
Beyond efficiency, GE Data Historian has a number of native Within the Enterprise Historian, the time-series data is stored
ways to move the data. From one Historian to another, one in an HDFS. The near-linear scalability of Hadoop allows our
can use a server-to-server collector for streaming data. Moving Enterprise Historian platform to scale out as the volume
from premise-based Historian to cloud-based time-series of data grows over time. Up to 20+ years of industrial big data
open source, Historian offers a server-to-cloud collector. To generated from the installed base of equipment can be stored
move larger files, one could use the Historian HD ingestion service online and mined on demand, replacing months of manual effort,
to move the tag configuration and archive data onto a Hadoop to explore much smaller data sets.
Distributed File System (HDFS) data lake on-premise or in
the cloud.

Industrial Data Management is the foundation for the Industrial Internet and goes beyond historian
time-series data. To effectively and efficiently deliver next-generation analytics and applications,
Industrial Data Management uses time-series data generated from machines and equipment in
conjunction with structured data such as enterprise resource planning (ERP), customer relationship
management (CRM), geolocation, semi-structured data such as machine logs and digital inspection
data, and unstructured data such as content (images, manuals, video, etc.). This is critical to bringing
a complete contextual and situational analysis for the assets. In addition, it becomes the basis to
empower personas beyond process engineers, such as data scientists and enterprise operations, to
unlock the value using advanced data management and analytic capabilities.

11 of 17 Unlocking Business Value Through Industrial Data Management

Rethinking Data Management
Data value chain from control to cloud.

We are learning a lot along with our Figure 4 below depicts how data Critical to this journey is to bring a wide
customers as we redefine GE as a data- management capabilities can traverse array of data together and relate it into
driven business. Beyond just assets, the enterprise—single-site, on-premise a common structure such as a data lake
our customers are looking at how to installations to multi-site, cloud-based in Predix™, our cloud platform for the
manage their data across the enterprise solutions. The cloud-based solutions go Industrial Internet, as shown in
and run analytical queries that provide beyond semi-structured time-series data figure 4. Business-level analyses based
immediate tangible insights to operations to include structured transactional system on statistical models can then be
and improve business metrics. To achieve data, and unstructured web and machine performed with Enterprise Historian
better business outcomes, digital industrial data, merging these data streams into a solutions as the foundational data layer,5
companies like ours are rethinking how single view. Evolving the use and capability as shown in figure 5.
they manage the data in the field in order of data historians to be cloud-based, is
to maximize the value of the analysis in central to rethinking data management.5
the cloud.5

Data Value Chain: From Asset to Enterprise, On Premise to Cloud

Integration to Innovation

Predix Industrial Cloud

ERP / CRM

Site Plant Enterprise

Figure 4
This data set can be looked at as a flow Outcomes from the data value chain with our ability to ask “what if” questions.
horizontally across the enterprise and into approach — from the control system to the We are able to model our assets and
a cloud-based environment, where that cloud — include: drive predictive analytics using our
data can be managed in much more cost- • A dramatic decrease in cost by SmartSignal technology, most notably
effective ways. A key dimension in control managing operational data more in the area of early anomaly detection in
to cloud is integration with systems in the effectively through a decrease in cost of asset performance. We are able to see
control environment at the asset layer, with IT infrastructure and operations deviations from normal or desired asset
operational systems and with enterprise • An increase in speed of deployment, behavior long before they become visible
systems. This is part of the reason why the speed of development of new analytical to standard operational systems.
queries and applications. A unified,
horizontal flow is important, as well—a systemic approach that results in Continued maturity expands into
single source of data across layers of repeatable speed and time to value operations optimization and business
responsibility, rather than competing sets optimization capabilities.5
• A decrease in complexity with focus
of authoritative data. on optimizing at an infrastructure level
Industrial companies desire to have a capability within the field
common data management capability. At Companies that can leverage this data
GE Digital, we are finding that our historical value chain horizontally across their
capabilities in control systems and data businesses, from their machines to the
management and processing capabilities cloud, will reach new levels of efficiency
in the field are complemented very well and business performance. There are
with our capabilities in the cloud. We are other possibilities starting to emerge, as
starting to design this data flow holistically well. As we start to facilitate a greater flow
with Predix. of data and the right data into the cloud,
the quality of our analyses improve along

A Scalable Framework for Industrial Analytics

Engineering Models That Continuously Increase Insights in Each Asset

to Deliver Specific Business Outcomes.

Inspection Data (e.g., Borescope)

Operational State Data (e.g., temp) Asset Life Data (e.g., cycles to failure)

Asset Performance
Management
Data Lake and Predix Cloud Performance Models Life Models
Operations Optimization
Business Optimization

Online Sensor Data Field and Service Data and Actions

Data Layer Model Layer Application Layer

Figure 5
Customer Case Studies

At GE’s Remote Monitoring and Diagnostics Higher-quality analytics engineers with less historical knowledge
(RM&D) Center in Atlanta, we are Data-driven decisions are correct and are more effective, improving productivity
leveraging our Data Historian capabilities optimal. We can now use larger data sets by nearly $9 million.
to collect data generated from 1,600 to create rules, thus reducing rule errors on Enabled growth
gas-fired turbines around the globe. unseen issues.4 Digitized knowledge is allowed to focus
The total output of these turbines can more on creation of new digital products
Increased customer satisfaction
support the annual energy needs of over and services.
Our customers are experiencing our
60 million homes. The data feeds a central
greater speed and consistency in addition Subjecting machinery data
GE Historian cluster. A team of more than
to savings through higher-quality analytics, Subjecting machinery data to analysis
20 M&D engineers analyzes this data to
data-driven analysis and decisions. and data mining operations has yielded
assist customers in enhancing their asset
The questions posed and the answers significant amounts of productivity
reliability and performance 24/7/365. This
available to this team via the GE Historian for GE and business benefits to GE's
data is then fed into a big data Historian
infrastructure have generated customer customers through better management
running in a Hadoop file cluster, where this
savings estimated at more than of their equipment (avoiding unplanned
team runs complex analytics across 100
$100 million. downtimes). To date, the team has led
million fleet operating hours.4
Elevated team effectiveness 15 patents on the Enterprise Historian
Digitized knowledge reduces training needs system and adjacent technologies.4
Business outcomes of and increases staff flexibility. Experienced
engineers with valuable knowledge
GE Historian at our RM&D can now be more productive, and new
Center include:
Increased productivity
We have reduced decision times and
minimized effort spent managing data.
GE Remote Monitoring and Diagnostics Center
This time to find cross-fleet patterns and
creating test rules, which used to be in
weeks or months, has reduced to minutes
or hours.
Reduced costs
Independent software development costs
have been reduced by $3 million.

Decreased infrastructure costs scaling

well over $1+ million include:
Data collection: From batch to real time
Storage size: 10x reduction
Database cost: 4x reduction
App server cost: 4x reduction
Data retrieval: 10x improvement
Software dev: Internal to COTS Figure 6
(commercial off-the-shelf)
Case studies below demonstrate how
we are partnering with customers in
different industries

One of the world’s A global manufacturer A major US

largest independent of consumer municipality with 43
oil and natural gas packaged goods political subdivisions
exploration and captures machine data in over 150 serving more than a million
production companies line-level GE Historians for local customers is leveraging GE Historian
was challenged with disparate analysis. This data is first aggregated to support their wet weather
islands of information. They had into a Plant Historian, which management and prevent sewer
grown rapidly through acquisition, generates manufacturing metrics overflow. With this solution, they
which resulted in no central data like operating efficiency, downtime will meet EPA regulations for water
source or a way to collect and report and waste. This data is then fed quality and gain visibility to 120
this data. They chose GE Historian as into a big data Historian running on remote pump and lift stations,
the central data store, and they now Cloudera to run fleet-level queries, providing an energy management
collect information from over 20,000 taking pressure off the production solution for seven major wastewater
wells. They are able to generate system and reducing the individual treatment plants. With a five-year
daily operating reports from the storage requirements at the line and capital budget of about $908 million,
information received into this single plant levels, where the customer a 1% decrease in operating costs will
trusted data store. They worked keeps only the last 30 days of data. save them nearly $2 million per year
with GE to create custom collectors This customer expects to save and avoid government fines for
from different SCADA systems. They millions annually in reduced storage non-compliance.
are also automating the modeling and infrastructure costs.
required to add a new asset to the
Historian database in watch time,
not calendar time, saving hundreds
of man-hours in labor per year.

Oil and Gas Exploration Consumer Goods Manufacturing Waste-water Management

and Production
Conclusion

Industrial companies have begun an exciting digital journey. At the heart of this transformation is the power of data analytics to unlock
new sources of value. However, the challenges of big data, threat of digital disruption and changing workforce dynamics are real. In order
to exploit the fast-moving technology wave of the Industrial Internet, companies need to think strategically and holistically
about the foundational elements of their data architecture, starting with Industrial Data Management.
GE has invested in a software solutions portfolio to provide our customers the building blocks of achieving Industrial Internet maturity.
On-ramping with cost-effective data management technologies that can aggregate, store, analyze and visualize terabytes of data pulsing
through assets and systems across the enterprise is a critical foundational step. These technologies give business leaders and operators
a single source of truth to improve asset-level visibility, cross-operation performance, knowledge capture and employee collaboration.
Having these capabilities in the field and in the cloud sets up the enterprise to extract value from insights that would have otherwise
remained hidden within islands of dark and disparate data.
We have deployed data management, predictive analytics and advanced control systems that are yielding operational improvements
and increased productivity for our customers. As a first mover of the Industrial Internet, our mission is to continue to lead the market
with our technologies, our domain experience and an unparalleled ecosystem of partners.

References
1 McAfee, Brynjolfsson. "Big Data: The Management Revolution," Harvard Business Review, 2012.
2 "Gartner Definition of Big Data." http://www.gartner.com/it-glossary/big-data
3 Annunziata, Bell, Buch, Patel, Sanyal. "Powering the Future: Leading the Digital Transformation of Energy," 2015.
4 Puig, Lu, Interrante, Pool, Aggour, Botros. Enterprise Historian for Efficient Storage and Analysis of Industrial Big Data, 2013.
5 Littlefield. "Will the Data Historian Die in a Wave of IIoT Disruption?" LNS Research, 2015.
http://www.geautomation.com/blog/rethinking-data-management-and-cloud-horizontal-data-value-chain-part-1
6 Stone. "Rethinking Data Management and the Cloud: Horizontal Data Value Chain, Part 1," GE, 2015.
http://blog.lnsresearch.com/will-the-datahistorian-die-in-a-wave-of-iiot-disruption
For more information visit www.ge.com/digital

16 of 17 Unlocking Business Value Through Industrial Data Management

Imagination at work

GE Digital
2623 Camino Ramon
San Ramon, CA 94583
www.ge.com/digital

GE Digital is a subsidiary of the General Electric Company. The GE brand, logo, and Lumination are trademarks of the General Electric Company. © 2015 GE Digital