the plant historian as a cloud...
TRANSCRIPT
By Harry Forbes
ARC STRATEGIES
NOVEMBER 2012
The Plant Historian as a Cloud Application
Executive Overview .................................................................... 3
Cloud Computing ........................................................................ 4
The Business Value of Cloud Computing ......................................... 8
Data Historians as a Cloud Application .......................................... 11
Historian Supplier Cloud Strategies .............................................. 14
Recommendations ..................................................................... 23
VISION, EXPERIENCE, ANSWERS FOR INDUSTRY
ARC Strategies • November 2012
2 • Copyright © ARC Advisory Group • ARCweb.com
Cloud-Computing Characteristic Explanation
On-demand self-service
A consumer can unilaterally provision computing capabilities, such as server time and network storage, as needed automatically without requiring human interaction with each service provider.
Broad network access
Capabilities are available over the network and accessed through standard mechanisms that promote use by heterogeneous thin or thick client platforms (e.g., mobile phones, tablets, laptops, and workstations).
Resource pooling The provider’s computing resources are pooled to serve multiple consumers using a multi-tenant model, with different physical and virtual resources dynamically assigned and reassigned according to consumer demand. There is a sense of loca-tion independence in that the customer generally has no control or knowledge over the exact location of the provided resources but may be able to specify loca-tion at a higher level of abstraction (e.g., country, state, or datacenter). Examples of resources include storage, processing, memory, and network bandwidth.
Rapid elasticity Capabilities can be elastically provisioned and released, in some cases automatically, to scale rapidly outward and inward commensurate with demand. To the consumer, the capabilities available for provisioning often appear to be unlimited and can be appropriated in any quantity at any time.
Measured service Cloud systems automatically control and optimize resource use by leveraging a metering capability1 at some level of abstraction appropriate to the type of service (e.g., storage, processing, bandwidth, and active user accounts). Resource usage can be monitored, controlled, and reported, providing transparency for both the provider and consumer of the utilized service.
Essential Characteristics of Cloud Computing (Source: NIST)
Cloud Computing Deployment Models
ARC Strategies • November 2012
Copyright © ARC Advisory Group • ARCweb.com • 3
Cloud computing is a new but rapidly growing computing model. It is
potentially disruptive to both software business models and enterprise IT
practices. Within manufacturing, ARC believes that the plant historian
software suite is likely to be an early adopter of the cloud model, but both
end users and suppliers see reasons for a gradual rather than rapid adoption.
Executive Overview
Cloud computing is a new computing model characterized by the intensive
use of remote networked computing resources. Cloud architectures are es-
pecially useful for scaling up computing tasks that can be parallelized.
Though relatively new in the commercial/industrial space, cloud technolo-
gy already has huge industry backing because it has been developed in
support of the internet applications of major software and internet firms
such as Google, Amazon, Apple, and Microsoft. Today, cloud data centers
alone comprise an $80-$90 billion annual market.
Cloud represents a disruption to established mod-
els of software development, software deployment,
support, and pricing. It also greatly lowers barriers
to entry in the software market.
Cloud computing also threatens to disrupt enter-
prise IT operations. Cloud offers enterprises more
rapid deployment, greater scalability (up and
down), reduced capital investment, and easier
support for mobile devices. Widespread cloud
adoption will cause enterprise IT organizations to act more as intermediar-
ies rather than internal project and IT asset managers. The business value
proposition for adopting cloud technology in an enterprise application de-
pends upon how strategic and differentiating the application is. Cloud
computing generally improves service for strategic applications and lowers
costs for non-differentiating applications.
In manufacturing, the plant data historian application has evolved signifi-
cantly. Beginning in the late 1970s early digital automation systems
delivered improved plant control and performance, but did not enable
plantwide digital data history or analytics. These first became available in
the late 1980s with the explosion of computer networking technology. Plant
historian software became a critical application and the first one that usual-
ly ran “above” automation networks on IT standardized platforms and
networks. Historian data became a foundation for many other critical ap-
plications and analytics. Such historians can now span many production
units and multiple plants.
Though the plant historian was the first operational technology (OT) appli-
cation to move from an automation system to an IT network, the historian
ARC Strategies • November 2012
4 • Copyright © ARC Advisory Group • ARCweb.com
presents several barriers for implementation in the cloud. Software suppli-
ers are eager to exploit potential cloud benefits such as better support for
mobility, inexpensive and large capacity resources, and easier data sharing.
However, today’s historian suppliers also deliver a large variety of other
software solutions that leverage historian data. For many customers (and
suppliers) these “other” applications are of equal or greater value than the
historian itself. Historian software has to evolve in a way that does not dis-
rupt the value being delivered by these applications. Many manufacturing
software suppliers are now taking the first steps in that journey.
Cloud Computing
Cloud computing is a new computing model that features the intensive use
of remote computing resources (processing, storage, networking, and soft-
ware). The salient feature of cloud computing is that these resources are
delivered to end users as a service over a network, typically the Internet.
Within cloud computing, the available computing resources are virtualized.
Virtualization separates an IT resource from specific physical hardware.
Virtualization can be applied to any IT resource; including servers, storage,
desktops, and networks. For example, server virtualization enables multi-
ple “virtual servers” to run on one physical server. This permits greater
server flexibility and greater server resource utilization. Virtualization pro-
vides large gains in available capacity, resource utilization, and energy
efficiency since it greatly reduces the number of physical resources required
per unit of virtual capacity.
Delivery as a service is another important aspect of the cloud model. Re-
gardless of the deployment model, cloud computing models are defined by
sets of services implemented via application programming interfaces
(APIs). Since the computing resources are remote and virtualized, cloud
APIs must present these resources at a higher level of abstraction.
Cloud APIs enable problems to be solved in a parallel fashion by dividing
the computing tasks and allocating these tasks to a large number of proces-
sors, usually referred to as a cluster. This contributes to cloud computing’s
scalability, or “elasticity.” While many cloud APIs are currently available,
several initiatives are now underway to develop standardized APIs for spe-
cific cloud computing platforms and develop open source implementations
ARC Strategies • November 2012
Copyright © ARC Advisory Group • ARCweb.com • 5
of these. However, since with cloud computing the API essentially defines
a product (service), standardization of cloud APIs is problematic, although
some cloud APIs may become widespread and popular. This situation is
similar to the state of APIs for internet servers where the wide variety of
available APIs range from commercial to open source and from very broad
to very narrow scope.
It’s important to understand the distinction between cloud computing and
outsourcing or remote hosting models. Cloud requires some level of soft-
ware re-design, while the other technologies can simply re-implement
existing practices at a different (and more advantageous) location.
The US National Institute for Standards and Technology (NIST) has devel-
oped an excellent definition that classifies cloud computing by three
properties; its characteristics, service models, and deployment models (see
following tables and on inside cover).
Software as a Service (SaaS)
The capability provided to the consumer is to use the provider’s ap-plications running on a cloud infrastructure2. The applications are accessible from various client devices through either a thin client interface, such as a web browser (e.g., web-based email), or a pro-gram interface. The consumer does not manage or control the underlying cloud infrastructure including network, servers, operating systems, storage, or even individual application capabilities, with the possible exception of limited user-specific application configuration settings.
Platform as a Service (PaaS)
The capability provided to the consumer is to deploy onto the cloud infrastructure consumer-created or acquired applications created using programming languages, libraries, services, and tools support-ed by the provider.3 The consumer does not manage or control the underlying cloud infrastructure including network, servers, operating systems, or storage, but has control over the deployed applications and possibly configuration settings for the application-hosting envi-ronment.
Infrastructure as a Service (IaaS)
The capability provided to the consumer is to provision processing, storage, networks, and other fundamental computing resources where the consumer is able to deploy and run arbitrary software, which can include operating systems and applications. The consumer does not manage or control the underlying cloud infrastructure but has control over operating systems, storage, and deployed applica-tions; and possibly limited control of select networking components (e.g., host firewalls).
Cloud Computing Service Models (Source: NIST)
ARC Strategies • November 2012
6 • Copyright © ARC Advisory Group • ARCweb.com
Private cloud The cloud infrastructure is provisioned for exclusive use by a single or-ganization comprising multiple consumers (e.g., business units). It may be owned, managed, and operated by the organization, a third party, or some combination of them, and it may exist on or off premises.
Community cloud
The cloud infrastructure is provisioned for exclusive use by a specific community of consumers from organizations that have shared concerns (e.g., mission, security requirements, policy, and compliance considera-tions). It may be owned, managed, and operated by one or more of the organizations in the community, a third party, or some combination of them, and it may exist on or off premises.
Public cloud The cloud infrastructure is provisioned for open use by the general public. It may be owned, managed, and operated by a business, aca-demic, or government organization, or some combination of them. It exists on the premises of the cloud provider.
Hybrid cloud The cloud infrastructure is a composition of two or more distinct cloud infrastructures (private, community, or public) that remain unique enti-ties, but are bound together by standardized or proprietary technology that enables data and application portability (e.g., cloud bursting for load balancing between clouds).
Cloud Computing Deployment Models (Source: NIST)
Cloud Maturity from Internet Applications
Much of what makes up cloud is already mature technology that is being
delivered today in high volume. Much of cloud technology has already
“crossed the chasm”. This is because cloud technology was initially devel-
oped to provide the extreme scalability that was required for successful
Internet applications. Google search is the canonical cloud application. But
Google, despite its success in internet search, was not simply a search com-
pany. Rather, the more accurate model for Google, even from its early
days, was that of a platform company that ran many different applications
on its own distributed (cloud) platform. These applications included
Google Earth, Blogger, Gmail, Picasa, YouTube, Reader, and others, with
Google Search being the most popular.
The need for extreme scalability of internet applications is not unique to
Google, but is also required by other major software and internet firms.
Amazon needs scalability for its massive on-line sales operation; Apple for
iTunes and for supporting hundreds of millions of consumer devices. Mi-
crosoft ran Hotmail, Bing, and now Skype. And the explosive growth of
Facebook, NetFlix, and other firms forced these companies onto cloud plat-
forms. Why? Because scaling up an application by a factor of 10,000
represents a serious technical challenge. Commercial software generally
cannot scale to this degree. These internet applications need to scale by
ARC Strategies • November 2012
Copyright © ARC Advisory Group • ARCweb.com • 7
Barriers to entry for cloud software- providers are very low,
since they can leverage established cloud platforms.
adding servers (and adding them by the hundreds or thousands), and so
require a virtualized, highly elastic, managed execution environment – a
cloud platform.
As cloud software platforms evolved, the explosive growth of web-oriented
companies created what is now a huge market for new data centers. These
data centers, often owned and financed by internet companies, require bil-
lions of dollars of capital investment. Current estimates of worldwide data
center investment are in the range of $80-90 billion per year (See:
http://www.arcweb.com/strategy-reports/2012-03-22/abb-initiative-
targets-the-data-center-1.aspx). Software companies, forced to invest bil-
lions in the “brick and mortar” of data centers, used their cloud software
platforms to generate additional revenue from their excess data center ca-
pacity. The close partnership between Amazon and NetFlix illustrates this
phenomenon. These platforms offer a huge business opportunity to the
market “survivors” as the market for cloud platforms consolidates.
Cloud Disrupts Software Development and Business Models
The cloud also threatens to disrupt traditional seat-based business models
for software pricing. While cloud does not mandate any change to these
models, it generally represents a shift away from IT pricing models domi-
nated by fixed costs, toward those containing a higher portion of variable
costs. Pay-for-use pricing models will fit cloud applications much more
easily than seat-based pricing.
The cloud’s technical disruption to software development and deployment
will be even more pronounced. For decades, software development has
meant coding critical applications from scratch and linking application code
with packaged libraries to create a deliverable for installation on a support-
ed operating system. Cloud development changes the deployment
environment for both platform as a service (PaaS) and software as a service
(SaaS) models and gives developers any number of new choices for devel-
opment tools and deployment platforms.
Cloud also opens a whole new field for software develop-
ers who create APIs linked to cloud services rather than
following traditional software distribution models. Barri-
ers to entry in this market are very low, since the cloud
service developers exploit the elastic properties of cloud
ARC Strategies • November 2012
8 • Copyright © ARC Advisory Group • ARCweb.com
technology in their own products. In theory, using third-party cloud ser-
vices should enable ISVs to focus on their core application logic. But
software developers also face a host of new challenges. These include find-
ing and testing cloud service options, implementing and testing new
applications, and support throughout their product lifecycle. Most likely,
PaaS providers will facilitate the choice of cloud services but this may not
effectively limit developer choices. Cloud represents a whole new ball
game for developers.
The Business Value of Cloud Computing
The technology disruption that cloud computing represents centers on IT
rather than OT. Cloud represents a new model for IT, challenging an organ-
ization’s established IT practices in every area; new platforms, new
services, new deployment, licensing, and support models. Cloud forces an
IT organization to examine each application within its portfolio and re-
evaluate the way the application is currently deployed and supported.
Furthermore, IT organizations have cultivated and developed many of the
skill sets that the cloud computing model threatens to disrupt. For exam-
ple, the practice of procuring, commissioning, and managing various
servers within an enterprise has been a critical IT skill for decades. Cloud
offers an alternative to this practice, whereas IT outsourcing simply re-
locates existing services to lower cost regions.
From a business standpoint the cloud model offers several potential ad-
vantages over the traditional IT service model:
Rapid deployment – Within many large enterprises it can take several
months to specify, procure, install, and commission a dedicated server.
With this in mind, organizations may move the procurement steps earlier in
a project, when less accurate preliminary sizing information is available.
Regardless, the cloud model offers resources that can be deployed more
quickly.
Elasticity – The concept of elasticity expands on scalability in that it implies
a service that can both grow and shrink to match demand. This attribute is
most valuable for customer-facing and internet applications, which may
exhibit huge demand variances. The “poster child” for cloud elasticity is
ARC Strategies • November 2012
Copyright © ARC Advisory Group • ARCweb.com • 9
probably the internet-based NetFlix
streaming service, which streams
video on demand to many millions of
US households, representing a sig-
nificant fraction of total US internet
traffic during prime time viewing
hours.
Reduced capital investment – The
Cloud model shifts costs from CapEx
to OpEx. It reduces the high initial
outlay required for servers, software,
support staffing, administration, and maintenance. In its place, cost behav-
ior approaches pay-per-use. The challenge here is that fixed costs such as
staffing and administration must be allocated across a portfolio of applica-
tions and these costs will not respond linearly to each adoption of a cloud
model.
Yet another cost advantage is that large corporations and IT shops can pur-
chase cloud services in volume. This enables an IT organization to contract
in advance with its internal customers for rapidly deployable resources.
This option changes the relationship between IT and the rest of the organi-
zation. Instead of managing IT capital projects, an IT organization can
aggregate the internal demand for new computing resources, and offer “in-
stantly available” spare capacity to its internal customers on a contingency
basis. In effect, instead of purchasing IT resources, the IT organization acts
as an intermediary between its own organization and third-party cloud
service providers.
Mobility – The cloud model can form part of the answer to the huge explo-
sion of smartphones and tablet devices that now confront IT organizations.
Since cloud services are by nature external to an enterprise, the gap that
develops between internal and external capabilities becomes less relevant
for cloud-based applications. While there are other solutions for mobility
problems, cloud-enabled location-independent support for user mobility is
one potential advantage of cloud services.
Cloud and the Business IT Portfolio
Businesses operate and maintain a portfolio of IT applications to support
their operations and management. Businesses that manage this portfolio
Cloud Resources Can Match Demands
ARC Strategies • November 2012
10 • Copyright © ARC Advisory Group • ARCweb.com
Cloud computing drivers vary greatly between strategic
differentiating applications and non-strategic applications.
The particular IT applications that are strategic differentiators
vary across vertical industries and enterprises.
effectively not only control their IT costs, but also better
support their company’s mission. ARC research indicates
that leading businesses often classify the applications in
their portfolio based on the degree to which they provide
them with a competitive differentiation or advantage. In
different businesses and industries, these differentiating applications can
vary greatly.
For example, one major global retailer classifies supply chain management
(SCM), inventory management, and point-of-sale as the most critically dif-
ferentiating applications, even though only the point-of-sale application is
directly visible to customers. While large discrete manufacturers such as
major automotives are also very concerned about their supply chains, most
classify their product lifecycle management (PLM) applications as among
the most critical differentiators, especially as they increasingly harmonize
their designs and products globally. Global consultancies or service firms
often consider CRM and HR applications most critical for managing and
deploying their staff and services. Process manufacturers, operating capi-
tal-intensive plants, look to asset lifecycle management (ALM) asset
performance management (APM), and enterprise asset management (EAM)
as the most critically differentiating applications, because these are the ap-
plications that have the potential to improve capital asset utilization and
product costs.
Strategies for allocating IT investment to particular applications drive in-
vestment towards the differentiating applications. Cost reduction is the
major strategy for applications where strategic differentiation is not a fac-
tor. The emergence of cloud architectures adds a new di-
mension to application investment decisions. However,
the drivers for cloud adoption will be much different for
strategic vs. non-strategic applications. For strategic ap-
plications, the benefits of greater scalability, faster time to
market, and greater mobility must weigh heavily. For non-strategic appli-
cations, a higher value point must be reached either through lower costs or
via best-in-class solutions at a lower cost point than is possible with con-
ventional computing architectures.
ARC Strategies • November 2012
Copyright © ARC Advisory Group • ARCweb.com • 11
Early DCS implementations improved process control, not historical data management.
Data Historians as a Cloud Application
Continuous process manufacturers have captured, stored, and managed
historical plant measurement data for decades, but the technology, data
volume, uses, and importance of historical process data continues to grow.
Since production in process plants is usually measured as a continuous
flow rate of product (barrels/day, gallons/minute, liters/min, megawatts,
etc.) the on-line process measurements are the best monitor of actual pro-
duction. In its simplest form, the historical data record for process plants
consists of time series data for each historized measurement.
Stages of Historian Technology
The earliest capture of plant historical data was done with paper chart re-
corders. In fact, paper charts were the only plant data repository until
automation systems became digital in the 1970s. Plants maintained a store
of charts representing critical production measurements, with each circular
chart representing a single day’s history. Filing was relatively easy, but
retrieval, comparison, correlation, and analytics involved so much manual
transcription and inaccuracy that these were largely impractical.
The early digital DCSs did not greatly improve this situation, and in fact
often made it worse. The mission of these systems was to improve and
even optimize process control (and hence production), rather than manage
data. While process data could now be captured and
retrieved with relative ease, from the mid-1970s through
mid-1980s, historical data within a DCS was primarily
limited to supporting the HMI. Besides their proprie-
tary designs, these systems also had very limited data storage capacity (of-
ten less than the legacy paper charts!). In that pre-computer-networking
era, historical process data was essentially captive within the DCS.
That was not the only difficulty. Process plants often utilize different au-
tomation systems in different plants and tend to modernize these
automation systems on a unit-by-unit basis. This made it difficult, if not
impossible, to merge historical data across units with DCS-resident histori-
ans, especially if the DCSs came from different suppliers or were even
different generations of the same supplier’s system.
ARC Strategies • November 2012
12 • Copyright © ARC Advisory Group • ARCweb.com
Plant-wide Historians Enable Data Visualization and Analytics
Enter the Plant Data Historian
Computer networking developed immensely during the decade 1985-1995.
As a result, standalone DCSs could not keep pace with the compute power
and storage capacity of networked systems. Process engineers discovered
powerful software tools to analyze data, but were often frustrated by lim-
ited ability to access their plants’ historical data for analysis.
During this period, data historians emerged as critical applications in the
process industries. Independent software vendors pioneered these prod-
ucts, followed quickly by process automation suppliers. The primary
differentiator of these products was architectural. They were client/server
applications that often ran outside the process control systems on the
plant’s standards-based LANs. These plant historians could support high-
capacity data interfaces to many different DCS, PLC, LIM, and other pro-
cess data sources.
Plant historians quickly added APIs that served desktop clients. These cli-
ents could be PCs within the enterprise or third-party software
applications. Engineers found these new software tools immensely valua-
ble. Plantwide historians could capture and manage large amounts of
measurement data from all the units in a process plant. They enabled pro-
cess engineers to visualize and analyze this data as a time series because all
the data carried some form of time tag. Taking advantage of the digital data
capture capabilities of the DCSs, the process historian became the founda-
tional tool for many activities.
For the first time, engineers had the abil-
ity to see a unified view of their plant
history to support improvements. They
could quickly define an ad hoc set of
plant measurements and display it as a
time-series. Plant-wide historical data
also provided the foundation for a se-
cond phase of more advanced analytical
applications.
Easy access to plant-wide historical data provided immediate payback in
the form of equipment performance metrics. Modern historians with data
retrieval APIs enabled these metrics to become on-line calculations that
could monitor and detect problems; alerting the plant staff to significant
deviations or changes. Many other types of analytics could now be per-
ARC Strategies • November 2012
Copyright © ARC Advisory Group • ARCweb.com • 13
formed as well. Control loop analytics could search through historical data
and advise on improvements to controller parameter settings, but were also
capable of looking for relationships between the disturbances or oscillations
of plant variables. This was only possible because the historian aggregated
a much larger number of measurements. These extended across unit
boundaries and beyond the view of process automation systems.
“Process identification” was one application used to design multivariable
control systems to support on-line optimization. Identification applications
create plant models using historical data. Depending upon the analytics,
the application will develop a model either by deliberately disturbing the
process or by relying on natural disturbances. The model is used to design
multivariable controllers and optimizations which are then deployed on-
line. The economic benefits of real-time
advanced control and optimization have
been huge, as is demonstrated by their
widespread use. An abundance of his-
torical data to use in the design of such
systems was a necessary condition for
their development.
Abundant historical data could also sup-
port production planning and scheduling
activities by comparing built-in assump-
tions against actual operating
performance. Data-driven analysis iden-
tified root causes of deviations between
predicted or ideal plant behavior and
actual performance. These analyses provided critical insights into how to
improve plant performance, but relied on extensive historical data services
as a foundation.
In summary, the domain of modern IT-enabled plant-wide data historians
now extends from the plant automation system to the desktop and from a
single process unit to many entire plants, arguably making them the most
critical component of the “OT” software portfolio for manufacturers.
Process Identification and Modeling Depend on Quality Historical Data (Source: Honeywell)
ARC Strategies • November 2012
14 • Copyright © ARC Advisory Group • ARCweb.com
“Plant” Historians Have Extended Their Reach to the Whole Enterprise (Source: OSIsoft)
Historian Supplier Cloud Strategies
Historian Is Only a Part of a Software Portfolio
Since the plant historian was really the first OT application to migrate “off
platform” from the process automation system to the plant and enterprise
network, it seems logical that historian might also be the first application to
move to a cloud computing platform. Why couldn’t historian suppliers
simply adopt a SaaS model and quickly migrate their customers to a cloud?
This question is much more complex than it may sound. Two factors great-
ly complicate the challenge.
Product portfolio integration - Historian software suppliers provide a large
number of other software products. These are not merely “add-ons” to the
historian, but are critical to both the supplier and the end user. Applica-
tions such as supply chain planning, real-time optimization, process design,
and process identification are common. In some cases, these applications
may be the companies’ flagship products even though they may rely on the
historian data interface. As a result, the software supplier could not mi-
grate the historian to a cloud platform without also integrating its entire
product portfolio with the cloud-based historian, a far more complex mi-
gration process.
Customer data storage preference – Data stored by the historian is both
highly critical and proprietary to a process manufacturing organization.
ARC Strategies • November 2012
Copyright © ARC Advisory Group • ARCweb.com • 15
Manufacturing organizations and plants with very limited IT
resources are better candidates for a cloud-based historian.
Many process industry firms are huge enterprises; super-major oil compa-
nies, national oil companies, global chemical/petrochemical manufacturers,
or large utilities. These companies are unlikely to be early adopters of an
architecture that would move their proprietary data to a service outside
their responsibility and control.
It’s not likely that software suppliers would try to force such a decision on
their customers. Rather, they could offer any new cloud-based product on
a private cloud that could be located with the enterprise.
This would enable their installed base to retain responsi-
bility for protecting their own proprietary data, but would
result in some trade-offs in terms of the elasticity and re-
mote management benefits of cloud architecture.
On the other hand, manufacturing operations that have very limited IT re-
sources are much better candidates for a cloud-based historian. Suppliers
that find this market difficult to serve may be able to gain some traction
here if cloud architecture enables them to provide a lower TCO for this
segment.
Publicly Announced Plans for Cloud Historians
A number of leading historian and automation suppliers have made public
announcements concerning their cloud strategy. Some cloud-based prod-
ucts have been released and some product roadmaps have been sketched
out. Other suppliers have not talked about specifics and some have been
entirely silent, at least in public.
The evolution toward cloud in the historian application is already common
enough that ARC sees the beginning of a trend, especially since market
leader OSIsoft has done probably the most talking. But OSIsoft is by no
means alone in looking to cloud as a way to add value to their historian of-
fering. Here is the current status based on the public information from
some major suppliers. ARC believes the suppliers who have kept quiet will
not be so for much longer.
OSIsoft Privately held OSIsoft pioneered the off-platform historian in the early
1980s. OSIsoft has always positioned its historian at the core of all the
products in its PI system software. Besides its long focus on the historian
application, OSIsoft has also been a close follower and partner of Microsoft
ARC Strategies • November 2012
16 • Copyright © ARC Advisory Group • ARCweb.com
Cloud architectures and deployments could enable easier sharing of “bulk” historical data with outside partners.
This is often a pain point in several manufacturing vertical industries.
since the mid-1980s. PI moved from its original VMS platform directly to
Windows servers, eschewing the UNIX platform period that most other
historian suppliers embraced temporarily.
OSIsoft has made public more detail on its cloud plans than most competi-
tors. It shared its view of cloud computing and long range plans in some
detail with customers and analysts at the OSIsoft Users Conference 2012.
The company mapped out five ongoing initiatives with respect to cloud. In
chronological order of development these are:
Mobility support – Breaking from its tradition, OSIsoft plans to add sup-
port for both iOS and Android client devices. Given Microsoft’s tiny share
of the smartphone and tablet market, this was the only way to enable PI on
most mobile devices. While in ARC’s view this is more a multi-platform
evolution than a cloud initiative, it reflects the reality that unlike its domi-
nance of the enterprise desktop, Windows is a rare exception in the mobile
device world.
PI Coresight SaaS – Coresight is the primary data visualization tool within
the PI system. Creating a SaaS version of this product will enable it to per-
form on different platforms as well as remove the administrative burden
from customer IT departments. In most companies, desktop data visualiza-
tion is the most widely deployed historian application, so reducing IT
support associated with managing it will be important to customers who
often support hundreds or even thousands of seats.
PI Data Exchange Service – This product is used for large-scale historical
data sharing via bulk data transfers. There are several common scenarios
where this could be useful for PI customers. One example is joint venture
ownership of large oil and gas assets, where some owners are also competi-
tors. While all owners may have the right to historical data, only one firm
is responsible for operations and it may be unwilling to offer network ac-
cess to companies who are its major competitors.
A second scenario is data sharing between an own-
er-operator and EPCs or service suppliers. Other
scenarios are contract manufacturing (common in
chemicals and pharmaceuticals), unit or plant op-
erating contracts between owners and large
equipment OEMs, and data sharing to support per-
formance benchmarking. While sharing historical data could provide bene-
ARC Strategies • November 2012
Copyright © ARC Advisory Group • ARCweb.com • 17
fits in all these scenarios, the benefits may not be realized if they require
allowing others access to on-premise IT assets. A cloud version of such a
sharing service would not have this liability, facilitating data sharing and
collaboration.
PI System Monitoring SaaS – This application monitors the performance
of an operating PI system. OSIsoft prides itself on being more a software
supplier than a service provider, leaving the data-driven applications to its
customers and partners. But properly tuning and monitoring a PI installa-
tion is necessary for maximum value. IT staffs at larger customers have
developed expertise in this, but a SaaS implementation would enable all
sizes of customers to benefit from best practices in system self-monitoring
and tuning.
“PI in the Sky” – This is OSIsoft’s tongue-in-cheek term for a cloud-based
re-architecting of the entire PI historian, a long-term initiative for the com-
pany. At present, OSIsoft is only developing customer use cases and
defining the potential properties that such an architecture might exhibit.
Notice that among these initiatives, the customer needs that appear to be
driving the program are the need for data and applications on mobile (non-
Windows) devices, as well as the ability to share data securely with external
partners. Since only a defined subset of historical data needs to be shared,
the ability to move that data subset to the cloud allows manufacturers to
exclude partners from access to their proprietary systems and data. Public
or shared cloud historian architectures would not have this attribute, and
do not have the same priority in OSIsoft’s view.
OSIsoft’s distinction in the industrial software space is that its portfolio has
always centered on its PI historian. Other industrial software companies
lead with different types of products (e.g. HMI, visualization, simulation,
multivariable control, analytics, etc.). For these companies, while im-
portant, their historians usually play a support role, rather than a central
role.
Invensys Operations Management Invensys Operations Management (IOM) takes another approach. Unlike
OSIsoft, Invensys is not purely a software company. It also offers complete
solutions for process automation and safety. However, similar to OSIsoft,
Invensys has had a long and close partnership with Microsoft. Given that
much of its software runs very close to the manufacturing process, IOM
ARC Strategies • November 2012
18 • Copyright © ARC Advisory Group • ARCweb.com
For its customers, Invensys believes the initial benefits of
cloud will be realized by those operating widely geographically
distributed assets or plants.
makes a clear distinction between automation/control software and infor-
mation-oriented applications, and sees the latter as the real candidates for
employing cloud computing.
Invensys sees a combination of greater flexibility and lower TCO as the
drivers for cloud adoption. The company sees cloud allowing customers to
get started quickly and then scale up as demand changes. It also sees cloud
reducing IT management and server administration costs. The company
wants cloud to provide customers with a high-availability architecture and
enable collaboration by increasing accessibility to knowledge and visibility
across the enterprise. Again, the customer pain points targeted relate to is-
sues with IT costs and application lifecycle management (e.g., version
control, patch management, and application monitoring) as well as ad-
dressing the slow adoption rate of new IT products by process industry
plants that may not own the necessary and up-to-date infrastructure.
IOM has discussed its Cloud strategy with key customers. They believe that
the initial benefits to their customer base will be realized by customers op-
erating geographically distributed assets or plants. These occur in many of
their served industries including food & beverage, utili-
ties, and Oil & Gas, as well as small-to-medium sized
enterprises (SMEs) like facilities management, solar panel
installers, and municipalities. One classic scenario is a
food plant that is small and remote, and must operate
with only a skeleton IT staff. Oftentimes the manufactur-
ing engineer also serves as the IT staff, so local resources can be very lim-
ited. Cloud-based applications allow such plants to effectively use
advanced applications, without the need for a local IT group or local sup-
port personnel.
To date, Invensys has three specific product initiatives involving cloud:
SmartGlance SaaS – SmartGlance is Invensys Operations Management’s
visualization and reporting application for mobile devices that connects to
data sources (usually including the Wonderware Historian). The company
released a cloud-SaaS version in 2011 to simplify data connectivity and ad-
ministration. Targeting cost-effectiveness, it enables data delivery
anywhere, anytime (via a mobile device), because the connection point for
data has moved to the cloud. Invensys reports that it has 50 SmartGlance
SaaS customers under subscription.
ARC Strategies • November 2012
Copyright © ARC Advisory Group • ARCweb.com • 19
Wonderware Tiered Historian – A cloud-based future version of this
product would offer users a combination of in-plant and cloud services.
Invensys sees the benefits as greater enterprise access to data at a lower
TCO point. The company also sees applications where customers will find
value in the separation of historical data from in-plant control networks.
Skelta – Skelta is the Business Process Management (BPM) workflow prod-
uct from Invensys. Many corporate software roll-outs (of any software
product, not necessarily Skelta) are challenged to develop and implement
standardized corporate-wide deployments. The challenge is to develop a
good pilot program, and to make the pilot development process both wide-
ly collaborative and visible so that all corporate stakeholders have input
and buy-in to the pilot deliverables. Invensys believes that cloud will speed
up Skelta deployment, and will also offer a single point of administration
and workflow management after the roll-out. This is in addition to the
“normal” cloud benefit of better support for client mobility and remote ac-
cess.
Not-So-Public Plans of Other Suppliers
Other automation and software suppliers have not yet made specific public
announcements about their intentions for cloud computing, but some ARC
insights on how the cloud could fit and enhance their respective portfolios
follow.
ABB Automation and energy equipment supplier ABB has been publically quiet
about the cloud but in ARC’s view seems to have cloud capability in a
number of areas. Since 2010, ABB has executed a string of acquisitions that
include some substantial software firms (Ventyx, Mincom, IKS, and Obvi-
ent). These firms have been formed into a unit within ABB that includes
ABB’s [electric power] Network Management business, a large existing
ABB software organization. Thus, in public at least, ABB has been more
focused on absorbing and extending the reach of newly acquired software
products than on strategizing about software architectures. The company
usually provides an updated product roadmap at its annual “Automation
and Power World” customer event.
Looking beneath the surface, both Mincom and Ventyx already had sub-
stantial business providing hosted solutions for their customers, and ABB is
attempting to leverage this capability as more of its customers look to struc-
ARC Strategies • November 2012
20 • Copyright © ARC Advisory Group • ARCweb.com
ture their applications as managed services. ABB states that its historian
can use resources “at any point in the network,” but thus far this only been
implemented at the automation and enterprise network, not in a cloud.
Another hint may come from the ABB ServicePort. While not technically a
private cloud, ServicePort is a “service delivery device” that can provide a
number of automated support services that are sold on a subscription mod-
el. ABB has only delivered ServicePort in configurations that drop into a
plant site, since this is acceptable to the installed base of customers and be-
cause the types of services (control tuning, event notification, system
monitoring, etc.) are intimately connected to the automation systems. From
a technical standpoint, ARC sees no reason why the “back end” of this
product could not move to a cloud platform. Customer acceptance proba-
bly remains the reason for the current architecture.
AspenTech AspenTech is does not engage in the business of process control but has a
huge portfolio of OT and IT applications for the process industries, includ-
ing its InfoPlus 21 historian. AspenTech has publically discussed its cloud
strategy only in the broadest terms, saying “The ability of the technology
tool to run in a cloud and a range of other IT advancements is likely to im-
pact and change the way software engineering for process industries is
going to be in the future….We believe the focus of future development and
course will be in making data and models available to our customers
through the web and cloud.”
Aspen has already updated its business model and licensing technology to
a pay-for-use model and away from seat licensing. Aspen’s primary busi-
nesses are in process design and simulation, on-line advanced control and
optimization, and production planning and scheduling. Historian and data
visualization capabilities are necessary to support these areas, as well as
being an application in their own right.
Many of the business pain points for historian data (for example shared
plant ownership) impact both the design and operating phases of any plant,
and thus ARC expects Aspen to pursue cloud-based capabilities. However
given its broad product portfolio, large installed base, and the fact that ap-
plications like APC operate in close touch with process automation, ARC
believes Aspen will place a premium on keeping its deployment models
very flexible so as to provide a wide range of options to existing customers.
ARC Strategies • November 2012
Copyright © ARC Advisory Group • ARCweb.com • 21
Honeywell filed a US patent application for “Cloud Computing as
a Basis for a Process Historian.”.
Honeywell Process Solutions The portfolio of Honeywell Process Solutions (HPS) is much more oriented
to process automation. While so far silent about the cloud, like many au-
tomation suppliers, HPS has aggressively employed server and PC
virtualization within its new Experion PKS Orion to reduce the required
number of servers and PCs and to streamline administration and security
measures.
While Honeywell is publically quiet about any over-
arching cloud strategy, ARC notes that in 2010 Hon-
eywell filed a US patent application entitled “Cloud
Computing as a Basis for a Process Historian.” Given
that, it appears that the company is considering the concept of a cloud-
based historian, but has not publically disclosed any impact on Honey-
well’s existing PHD historian. PHD is a tiered historian product, so site-
wide or enterprise PHD implementations could, in theory, be located in a
cloud. Internally, Honeywell Process Solutions collaborates on many de-
velopment activities within the larger Honeywell ACS organization that
includes both Honeywell’s process and building automation businesses.
Rockwell Automation Another automation supplier example is Rockwell Automation, which has
developed cloud-based projects leveraging its relationship with Microsoft.
In Rockwell’s case the customer was a large drilling equipment manufac-
turer who needed visibility and analytics over many remote sites where its
Rockwell-automated equipment was being operated.
Analysis
A transition to cloud involves two major tasks; a re-architecting of software
and a choice of a deployment model. These two activities are largely inde-
pendent, meaning that private clouds within a plant’s Level 2-Level 3
systems are not revolutionary developments and ARC expects to see these
soon. Cloud deployment options will enable software and solution provid-
ers to more closely match their deliverables to customer preferences. Public
or hybrid cloud or hosted solution architectures are more a natural fit for
applications that operate over several production sites (production plan-
ning and scheduling is one). Here an “external” cloud is an advantage.
Cloud encompasses three major deployment models (public, private, and
hybrid). Application and server platforms that are clearly OT rather than
ARC Strategies • November 2012
22 • Copyright © ARC Advisory Group • ARCweb.com
IT and run inside the plant increasingly resemble private clouds. ABB’s
ServicePort and the extensive use of blade servers and virtualization in new
DCS platforms like Honeywell Experion PKS Orion are just two examples.
ARC expects this trend to continue because cloud technologies enable sup-
pliers to leverage commoditized hardware, software, and networks to
deliver high-availability services.
Many “on-line” manufacturing applications seem at first glance to be un-
likely candidates for cloud computing. This is especially true for
applications like APC, which tend to be tightly coupled to the automation
system and must respond quickly to changes in the process and/or auto-
mation system states. Indeed, ARC believes that some applications will
become more tightly coupled with automation while others will more natu-
rally migrate to the cloud. On the other hand, production applications that
impact multiple plant sites, such as production planning and scheduling,
might benefit from the easier universal access enabled by cloud deploy-
ment. This is particularly true for manufacturers with numerous and
widely distributed plants (e.g., the food and cement industries) that need to
coordinate their operations across multiple facilities.
Historian applications have traditionally taken advantage of improvements
in compute capacity, network speed, and storage capacity. Because of this,
it makes good sense that historian applications will be among the first OT
applications to utilize cloud software architecture, though ARC expects
many manufacturers will initially adopt “private cloud” deployment mod-
els that represent less technological disruption and less perceived risk.
Indeed while historical data is a valuable resource and needs to be kept
confidential, key analytical results are far more strategic to manufacturers.
There are numerous parallels between developments in conventional com-
puting and the emergence of the cloud. The decades-long IT trend has been
toward greater reliance on remote networked computation and storage re-
sources. This trend has persisted from the client/server model of thirty
years ago through the development of the Internet and today manifests it-
self as the growth of cloud computing solutions.
ARC Strategies • November 2012
Copyright © ARC Advisory Group • ARCweb.com • 23
Recommendations
ARC recommends the following for plant owner-operators:
• The term “cloud computing” encompasses so many different options
that plant owner-operators should disregard the term itself, and focus
on what business value can be delivered by any “cloud-enabled” archi-
tectural changes in their applications, especially historians.
• Greater plant data visualization on (non-Windows) mobile devices is
likely to be an early cloud benefit.
• Another cloud historian benefit will be easier sharing of historical data
subsets with partners, enabling collaboration without the need for
granting access to the enterprise network.
• For specific applications, manufacturers should consider the pros and
cons of cloud computing based on whether or not the application is a
“strategic differentiator” for the firm.
• For strategic applications, firms should focus on how cloud could ena-
ble them to deliver greater value. For non-strategic applications, cost
reduction is the primary driver for change.
• Using cloud resources can rationalize IT equipment purchases by great-
ly simplifying system sizing considerations. The cost model for cloud IT
resources features less fixed cost and a more linear variable cost charac-
teristic. Cloud resources will also be available in huge (practically
unlimited) amounts at predictable prices, which are likely to decline
over time. Plant historian applications will be among the first to benefit
from these properties.
ARC Strategies • November 2012
24 • Copyright © ARC Advisory Group • ARCweb.com
ARC Strategies • November 2012
Copyright © ARC Advisory Group • ARCweb.com • 25
Analyst: Harry Forbes
Editor: Paul Miller Distribution: MAS and EAS Clients
Acronym Reference: For a complete list of industry acronyms, refer to our web page at www.arcweb.com/Research/IndustryTerms/
ALM Asset Lifecycle Management APC Advanced Process Control API Application Program Interface APM Asset Performance Management BPM Business Process Management CRM Customer Relationship
Management DCS Distributed Control System EAM Enterprise Asset Management HMI Human Machine Interface HPS Honeywell Process Solutions HR Human Resources IaaS Infrastructure as a Service IOM Invensys Operations Management
ISV Independent Software Vendor IT Information Technology LAN Local Area Network LIM Laboratory Information
Management NIST National Institute for Standards
and Technology OT Operational Technology PaaS Platform as a Service PLC Programmable Logic Controller PLM Product Lifecycle Management SaaS Software as a Service SCM Supply Chain Management TCO Total Cost of Ownership
Founded in 1986, ARC Advisory Group is the leading research and advisory firm for industry. Our coverage of technology from business systems to prod-uct and asset lifecycle management, supply chain management, operations management, and automation systems makes us the go-to firm for business and IT executives around the world. For the complex business issues facing organizations today, our analysts have the industry knowledge and first-hand experience to help our clients find the best answers.
ARC Strategies is published monthly by ARC. All information in this report is proprietary to and copyrighted by ARC. No part of it may be reproduced with-out prior permission from ARC.
You can take advantage of ARC's extensive ongoing research plus experience of our staff members through our Advisory Services. ARC’s Advisory Services are specifically designed for executives responsible for developing strategies and directions for their organizations. For membership information, please call, fax, or write to:
ARC Advisory Group, Three Allied Drive, Dedham, MA 02026 USA Tel: 781-471-1000, Fax: 781-471-1100 Visit our web pages at www.arcweb.com
3 ALLIED DRIVE DEDHAM, MA 02026 USA 781-471-1000
USA | GERMANY | JAPAN | KOREA | CHINA | INDIA | BRAZIL | ARGENTINA