You are on page 1of 6

Next-Generation Enterprise Architectures

Common Vernacular and Evolution Towards Service-Orientation ✰

Mohsen Moghaddam a, b, *, C. Robert Kenley a, b, Julia M. Colby a, b, Marissa N. Cadavid Berns a, b, Randy Rausch c,
Joel Markham c, Wesley M. Skeffington c, John Garrity c, Alok R. Chaturvedi a, d, Abhijit V. Deshmukh a, b
a
GE-Purdue PRIAM, Purdue University, West Lafayette, IN, USA
b
School of Industrial Engineering, Purdue University, West Lafayette, IN, USA
c
GE Global Research Center, Niskayuna, NY, USA
d
Krannert School of Management & Department of Computer Science, Purdue University, West Lafayette, IN, USA

Abstract—Industry 4.0 is opening new avenues for reconfigu- systems, and developing a path towards defining new standards
rable and information-centric integration of enterprise functions (if needed). The NGEA project has been defined in three phases:
and control systems. Most of the current approaches (e.g., ISA-95)
view enterprise architectures in a pre-defined, monolithic, and hi- 1) Evaluate existing and emerging standards, define com-
erarchical sense. To enable more innovative, personalized, and ef- mon vernacular, educate on how the standards interact
ficient manufacturing processes, however, such ‘tree-like’ archi- and overlap, summarize strengths and weaknesses;
tectures must turn into decentralized and cyber-physical networks 2) Define gaps, if any, in existing/emerging architectures,
of ‘things’ and ‘services’. This article investigates the emerging In- clearly define the unsolved problem;
dustry 4.0 paradigms and architectures (e.g., IIRA, RAMI4.0), 3) If necessary, design and publish a new architecture.
their commonalities, limitations, and evolution towards modular
and service-oriented architectures (e.g., ISO/IEC 18384:2016). This article encompasses parts of Phases 1 and 2.
The goal is to identify and analyze the existing technical and tech-
nological gaps, and provide recommendations for developing new The Purdue Enterprise Reference Architecture (PERA) [5],
reference models/architectures for next-generation enterprises. developed by Prof. Ted Williams at Purdue University, served
as the basis for the industry-wide standard for enterprise integra-
Keywords—Service-oriented architecture; Modular design; In- tion, ISA-95 (IEC 62264:2013 [6]), and heavily influenced the
dustrial internet of things; Cyber-physical system; Industry 4.0 development of other frameworks (e.g., SCOR; OAGIS; DIRA;
MESA) and architectures (e.g., ISA-99 for security; ISA-101 for
I. INTRODUCTION human-machine interfaces). The task of information integration
The fourth industrial revolution, or Industry 4.0, is charac- also extends to device communication protocols, SCADA, and
terized by a series of tools and technologies that enable seamless industrial control functions, such as MTConnect, OPC UA,
connectivity between physical systems and their ‘digital twins’. Modbus, Profinet, Foundation Fieldbus, and ISA-100 for wire-
The vision of Industry 4.0 is to leverage these new capabilities less communications. To fully enable the capabilities of Industry
(e.g., networked/digitized processes; cyber-physical systems) to 4.0 systems, however, new frameworks and architectures are re-
better address the dynamic, unpredictable, and complex nature quired that support dynamic composition of functionality based
of next-generation manufacturing systems [1], and emerging on a modular and service-oriented perspective of an enterprise.
challenges such as agility [2], interoperability and cyber-secu- The Industrial Internet Reference Architecture (IIRA) [7]
rity [3]. A prerequisite for development of an ‘unprecedented’ and the Reference Architecture Model Industrie 4.0 (RAMI4.0)
system like this is an architecture—a ‘blueprint’ that “provide[s] (DIN SPEC 91345:2016 [8]) are amongst the pioneering refer-
current or future descriptions of a ‘domain’ composed of com- ence models for next-generation enterprises. The vision of IIRA
ponents, and their interconnections, actions or activities those is to create a broad industry consensus for higher interoperability
components perform, and the rules or constraints for those ac- and enhanced development of Industrial Internet systems. IIRA
tivities” [4]. relates to a wide range of industries including energy,
The General Electric-Purdue University Partnership in Re- healthcare, manufacturing, and transportation. On the other
search and Innovation in Advanced Manufacturing (GE-Purdue hand, RAMI4.0 proposes a 3-dimensional reference model de-
PRIAM) has recently launched a new project on Next-Genera- termining the Functional Layers of Industry 4.0, as well as the
tion Enterprise Architectures (NGEA), with the goal of investi- lifecycle, value stream, and equipment hierarchy models.
gating the fitness of current enterprise architectures for infor- RAMI4.0’s primary focus is on manufacturing. Due to the sig-
mation-centric integration of enterprise functions and control nificant overlap between the frameworks and capabilities of
IIRA and RAMI4.0, and their considerable similarity with ISA-

This work is sponsored by the GE-Purdue Partnership in Research and In- 95, it is important to evaluate their relative strengths and weak-
novation in Advanced Manufacturing (PRIAM). nesses with respect to different enterprise integration capabili-
* Corresponding author, ties. In doing so, one can identify gaps in the state-of-the-art that
mmoghadd at purdue.edu, (765) 430-4957. need to be filled, exposing possibilities for new standards.

l-))) 

Authorized licensed use limited to: Steve Bishop. Downloaded on June 07,2020 at 21:13:18 UTC from IEEE Xplore. Restrictions apply.
The purpose of this article is to systematically compile the Language, ISO/IEC 19501:2005) for constructing common ob-
inventory of existing industry-wide enterprise reference models ject/information models associated with different manufacturing
and architectures, particularly IIRA and RAMI4.0, and evaluate scenarios. The equipment hierarchy levels in ISA-95 are as fol-
their relative key strengths and limitations vis-a-vis (1) the In- lows (IEC 62264-1:2013 [6]):
dustrial Internet-enabled sensing, communication, computing,
and integration capabilities, and (2) the evolution of systems ar- x Enterprise. A collection of sites and areas, responsible
chitecting methodologies and practices, from hierarchical Struc- for high-level, strategic decisions.
tured Analysis models to Service-Oriented Architecture (SOA). x Site. A physical/geographical/logical grouping of areas,
Section II provides an overview of the equipment hierarchy production lines/units, and process cells, responsible for
model of ISA-95 and the I4.0 Component model of RAMI4.0, rough-cut planning and scheduling.
and the path towards developing a standardized ‘Internet-of- x Area. A physical/geographical/logical grouping of pro-
Things’ (IoT) [8]. Section III discusses the functional hierarchy duction lines/units, process cells, and storage zones, with
models of ISA-95, IIRA, and RAMI4.0, along with their capa- well-defined manufacturing capabilities and capacities.
bilities, and mapping for creation of an ‘Internet-of-Services’
(IoS) [9]. Section IV summarizes the discussion, and provides x Work center and unit.
recommendations for future research and development. (Note: - Production unit. Continuous manufacturing equip-
The focus of this work is primarily on manufacturing; however, ment.
the concepts and architectures discussed here are also applicable ▪ Unit. Lower-level elements and modules.
to other sectors.) - Production line. Discrete manufacturing equipment.
▪ Work cell. Lower-level elements for flexible rout-
II. INTERNET OF THINGS
ing of work (if applicable).
Industry 4.0 is based on inter-networking of ‘things’ (e.g., - Process cell. Batch manufacturing equipment.
sensors; computing devices; robots; machines; organizations) ▪ Unit. Lower-level elements for flexible routing of
supporting a wide range of interaction, communication, and work within process cells (if applicable)
computing services. Revisiting the definition of enterprise archi- - Storage zone. Material handling/storage equipment.
tecture in Section I, these ‘things’ refer to the interconnected ▪ Storage unit. Lower-level units with finer details
‘components’ performing different actions or activities follow- and changeable location (e.g., an AGV).
ing certain rules or constraints. This section presents an inte-
grated approach, based on the RAMI4.0 standard DIN SPEC ISA-95 proposes a monolithic and hierarchical model for
91345:2016 [8] as well as the Equipment Hierarchy model of systematic representation of components in a manufacturing en-
ISA-95 (IEC 62264-2:2013 [6]), for standard representation of terprise (generalizable to other industrial domains). However,
Industry 4.0 Components (henceforth, I4.0 Components). In this the inherent complexity, uncertainty, and dynamicity of Industry
context, an IoT (Internet-of-Things) refers to a network of I4.0 4.0 systems call for higher ‘agility quotient’; i.e., ability to ef-
Components, configured following certain rules/conditions for fectively accommodate time-pressure (responsiveness), recover
on-demand provision of services. (Services are further discussed from damage (resilience), handle inadequate options (flexibil-
in Section III.) ity), invent new solutions (innovativeness), handle a variety of
situations (versatility), and self-configure (adaptation) [2]. This,
A. ISA-95 Equipment Hierarchy Model in turn, necessitates modular design—characterizing an Industry
ISA-95 categorizes components in a role-based hierarchical 4.0 system as a network of distributed, self-contained, and au-
manner, representing the relationships between different hierar- tonomous modules (e.g., agents [10]; holons [11]), with inte-
chy levels along with the areas of responsibility associated with grated local collaborative autonomy and global orchestration
each level (FIG. I). ISA-95 applies UML (Universal Modeling [7].

Layers Enterprise ISA-95

Business Site
Functional
Area
Information
Communication
Process Production Production Storage
Integration Unit Line
Cell Zone
Asset
Unit Unit Work Cell Storage Module

Lower Level Lower Level Lower Level Lower Level


Resources in Resources in Resources in Resources in
Batch Ops. Continuous Ops. Discrete Ops. Handling Ops.
RAMI4.0

FIGURE I. THE PHYSICAL ARCHITECTURES OF ISA-95 (ROLE-BASED EQUIPMENT HIERARCHY MODEL) AND RAMI4.0 (HIERARCHY LEVELS).



Authorized licensed use limited to: Steve Bishop. Downloaded on June 07,2020 at 21:13:18 UTC from IEEE Xplore. Restrictions apply.
B. I4.0 Component
RAMI4.0 builds its component hierarchy model upon ISA-
95 and ANSI/ISA-88 standard for batch control (IEC Consumer

Mgmt. & Security


61512:1997 [12]), by incorporating three additional elements:

Development

Governance
Information
Integration
Field device; Product; Connected world (FIG. I). Field device Process
refers to the smaller elements (e.g., sensor) of a large control
device (e.g., a machine). Consideration of product implies that Services
in addition to factory machinery and operators, RAMI4.0 takes
into account the outcomes of manufacturing processes as com- “Things”
ponents of the system. External collaborators (e.g., suppliers;
manufacturers; customers) are also considered by adding con-
nected world on the top of the equipment hierarchy model.
FIGURE II. THE SOA REFERENCE MODEL (ADAPTED FROM ISO/IEC
In addition to the equipment hierarchy, RAMI4.0 also ad- 18384:2016).
dresses continuous data acquisition and improvement through-
out the lifecycle of components (e.g., products; orders; equip-
ment; factories), as well as the associated value streams (FIG. I). of components and functions. The NGEA project proposes an
The lifecycle model is based on IEC 62890 [13], where the dis- analog for these architectures in the context of Industry 4.0; an
tinction between component ‘type’ and ‘instance’ is the key— architecture for dynamic mapping of IoT and IoS (FIG. II)—an
type-related examples include prototype design, development, SOA adapted from ISO/IEC 18384:2016 [9].
testing, and validation; while instance-related examples are A service refers to “a logical representation of a set of activ-
sales, delivery, data collection, and after-sale services. Virtual- ities that has specified outcomes, is self-contained, may be com-
ization of components is the main enabler of the value stream posed of other services and is a ‘black box’ to consumers of the
model. It enhances visibility across different domains, thus ena- service” [9]. Going back to the definition of enterprise architec-
bling systematic elimination of non-value-adding processes. ture in Section I, a service refers to the set of actions/activities
(This is further discussed in Section III.) that the components of a system perform, under certain rules or
The unique aspect of RAMI4.0—corresponding to the notion of constraints. Thus, services and I4.0 Components complement
modular design—is the I4.0 Component, which may refer to a each other, under the umbrella of an SOA (FIG. II). This section
group of hardware, software, ideas, or concepts, arranged on a briefly discusses the Functional Hierarchy model of ISA-95, the
dynamic basis to provide different services on demand. Functional Viewpoint of IIRA, and the Functional Layers of
RAMI4.0 defines certain characteristics, in order for a group of RAMI4.0. The goal is to generate a ‘superset’ of enterprise func-
‘things’ to be regarded as an I4.0 Component: Identifiability, tions, which can then be formalized as standard services, com-
communication capability, compliant services, states and se- pliant with ISO/IEC 18384:2016 [9].
mantics, virtual description, safety, security, quality of service, A. ISA-95 Functional Hierarchy Model
nestability, and separability [8]. RAMI4.0 introduces the notion
ISA-95 defines five hierarchy levels for different functions
of an ‘administration shell’ as the logical unit of an I4.0 Com-
of a manufacturing enterprise; from business planning to logis-
ponent, responsible for virtual representation, interaction with
tics, manufacturing operations management, and shop-floor
the system, and resource management. RAMI4.0 also defines a
control. The functional levels are (IEC 62264-1:2013 [6]):
‘CP’ classification, which classifies components in an Industry
4.0 system with respect to two key characteristics [8]: x Level 4. Business-related functions such as production
scheduling, shipping, and inventory control.
x Communication capability (C). The ability of digital
communication (e.g., via Fieldbus or TCP/IP). x Level 3. Process-related functions such as workflow con-
- Classes: None; Passive; Active; I4.0 compliant. trol, detailed scheduling, and reliability assurance.
x Presentation (P). The ‘publicity’ of a component in the x Level 2. Process monitoring functions such as supervi-
information system. sory/automated control of processes.
- Classes: Unknown; Anonymous; Individually x Level 1. Sensing and process manipulation functions.
known; Entity. x Level 0. The physical processes.
The majority of the high-level functions in the Functional
III. INTERNET OF SERVICES
Hierarchy model of ISA-95 (e.g., Levels 3 and 4) are processed
The purpose of every sensor, device, machine, robot, soft- at the Enterprise, Site, or Area levels [6], and functions or com-
ware, or simply, every ‘thing’ in an Industry 4.0 system is to munications are not able to jump between non-adjacent levels.
perform certain functions. That is, parallel to the physical archi- This indicates the centralized and monolithic nature of decision-
tecture of an Industry 4.0 system, characterized by IoT, is a func- making and control in ISA-95, which is contradictory to the
tional architecture, which we characterize as an IoS, or an Inter- ideas of agile and modular design discussed thus far.
net-of-Services. The mapping between the physical and func-
tional architectures in the classic Structured Analysis techniques B. IIRA Functional Viewpoint
[4] is performed by an allocated architecture (a.k.a., configura- IIRA is composed of four viewpoints addressing the con-
tion architecture). The allocated architecture provides a set of cerns of different stakeholders compliant with ISO/IEC/IEEE
rules for governing the topology, interactions, and independence 42010:2011 [14]. Considering local collaborative autonomy and



Authorized licensed use limited to: Steve Bishop. Downloaded on June 07,2020 at 21:13:18 UTC from IEEE Xplore. Restrictions apply.
optimization through global orchestration as the two emerging x Business. The functions mapping the business mod-
thrusts in Industry 4.0 systems, the Functional viewpoint of els/processes, defining rules and regulations, and orches-
IIRA specifies five major functional domains as follows (FIG. trating services in the Functional layer.
III): x Functional. The functions enabling formal description,
x Business. The functions enabling end-to-end operations modeling, and integration of services, associated with the
of an Industry 4.0 system; Business layer and the corresponding processes.
- e.g., enterprise resource planning; customer relation- x Information. The functions supporting event pre-pro-
ship management; lifecycle management; billing; cessing, execution of rules, data analysis, quality assur-
planning and scheduling. ance, and provision to the Functional layer.
x Application. The functions implementing the application x Communication. The functions supporting communica-
logics that realize business functionalities; tion for the Information layer, and providing services for
the control of the Integration layer.
- e.g., rules; models; activity/workflows; application
programming interface; user interface. x Integration. The functions providing asset information,
enabling computer-aided control, event-generation, con-
x Information. The functions supporting data provision nectivity, and component virtualization.
and deployment;
x Asset. The functions physically performed by the com-
- e.g., data gathering; syntactical and semantic transfor-
ponent or connecting the component to the virtual world
mation; storage; quality processing; distribution.
by passive means (e.g., UPC barcode).
x Operations. The functions operating components
throughout their lifecycle; D. Services
- e.g., provisioning; deployment; management; moni- The functional hierarchy models proposed by ISA-95, IIRA,
toring; diagnostics; prognostics; optimization. or RAMI4.0 may be sufficient for representation of system func-
tions and their correspondence with the traditional, hierarchical
x Control. The functions enabling the industrial control structure of components. However, considering the definition of
system; IoT in the context of Industry 4.0, as a network of self-contained
- e.g., sensing; actuation; communication; abstraction; and autonomous I4.0 Components, these hierarchical models
virtualization; analytics; asset management. must be transformed into analogous ‘network models’, com-
posed of self-contained and autonomous groupings of functions;
C. RAMI4.0 Functional Layers i.e., services. ISO/IEC 18384:2016 [9] provides a comprehen-
The four major aspects of RAMI4.0 are vertical integration sive taxonomy of services (mainly for software engineering ap-
of components and services, end-to-end engineering throughout plications), classified as domain-specific (e.g., process services;
value streams and component lifecycles, horizontal cross-enter- information services; business services) and domain-independ-
prise integration, and human-in-the-loop orchestration of value ent (e.g., security services; infrastructure services; life cycle ser-
streams [8]. With this in mind, RAMI4.0 specifies the following vices). Service-orientation offers several unique benefits com-
six major functional layers (FIG. III): pared to the traditional, hierarchical models:

① Hardware; So ware; Human Resources; Ideas; Concepts ⑥ Semantics; Syntaxes; Persistence; Storage; Quality Processing; Analytics; Distribution
② Sensing; Actua on; Virtualiza on; Modeling; Execu on ⑦ Engines; Ac vity Flows; Workflows; API & UI; Service Modeling
③ Communica on; Interfacing ⑧ Specifica on of Rules & Models
④ Provisioning; Deployment; Asset Mgmt.; Monitoring; Diagnos cs; Prognos cs ⑨ Service Orchestra on; ERP; CRM; Lifecycle Mgmt.; Billing; HRM; Planning; Scheduling
⑤ Data Collec on

Business Plan/Logistics Functional Domains


Production planning and scheduling, Business
engineering design, purchasing, customer Layers
order handling ⑨

Business
Information

Application
Operations


Functional

Mfg. Ops. Mgmt. ⑤
Information
Production, maintenance, quality Communication
testing, inventory and material flow ④
Control ③ Integration
management ②
Asset
Sensing Actuation

Batch Continuous Discrete


Control Control Control Physical Systems

ISA-95 IIRA RAMI4.0

FIGURE III. THE FUNCTIONAL LAYERS OF ISA-95, IIRA, AND RAMI4.0 (DOTTED LINES: MAPPING BETWEEN LAYERS OF IIRA AND RAMI4.0).



Authorized licensed use limited to: Steve Bishop. Downloaded on June 07,2020 at 21:13:18 UTC from IEEE Xplore. Restrictions apply.
x Interoperability. Addressing heterogeneity issues via A. Service-Oriented Architecture—IoS
open and standard service interfaces;
Various SOA reference models have been developed over
x Loose-coupling. Enabling intra/inter -enterprise compo- the past decade and implemented with significant financial suc-
sition, provisioning, and deployment of services; cess [16]; however, these are mostly designed for software en-
x Complexity-handling. Lowering system complexity via gineering applications. Common examples include the IBM Ser-
‘encapsulation of services as ‘black-boxes’ to users; vice-Oriented Modeling and Architecture [17], the Open Group
SOA Reference Architecture Technical Standard [18], the Ar-
x Stateless-ness. Focusing on the functionalities of an ap- rowhead framework [19], the OASIS Reference Model for Ser-
plication rather than the objects providing it; vice Oriented Architecture [20], and the General Technical Prin-
x Scalability. Supporting reconfiguration/expansion of an ciples of SOA (ISO/IEC TR 30102:2012, Annexes D & F) [21],
enterprise without major architectural changes; leading to the development of the international standard
x Reusability. Enabling processes composition via dy- ISO/IEC 18384:2016, Reference Architecture for SOA (FIG. II).
namic mapping of services to their ‘physical twin’; The manufacturing ‘analog’ of these developments is Cloud
Manufacturing built upon the Cloud Computing reference
x Standardization. Providing standard service descrip- model [22]. Nevertheless, the current Cloud Manufacturing ref-
tions, contracts, interfaces, and governance. erence architectures lack several key features of ISO/IEC
18384:2016 such as a standard and all-inclusive definitions of
IV. DISCUSSION various architectural capabilities/aspects, building-blocks, met-
rics, and service categories [9]. A recent Smart Manufacturing
The Industry 4.0 vision for cost-effective, agile, and cus-
System Architecture proposed by NIST [23] addresses some of
tomer-centric production cannot be fully materialized without
these concerns, while still noting many implementation chal-
interweaving IoT with a parallel IoS. This introduces two fun-
lenges. ‘Manufacturing adoption’ of ISO/IEC 18384:2016 is
damental questions:
therefore a foremost R&D/implementation effort, which must be
Q1 What changes if the enterprise resources and functions conducted consistent with DIN SPEC 91345:2016 [8], as well
are standardized as self-contained and autonomous as other emerging frameworks/architectures such as ManuCloud
components and services, respectively? [24] and Predix [25].
Q2 How can businesses make this transition from a central- B. Modular Design—IoT
ized and hierarchical architecture such as ISA-95 to a
modular and service-oriented architecture? An essential step in ‘manufacturing adoption’ of ISO/IEC
18384:2016 is to identify the analogs of software objects and
The changes resulting from modularization and service-ori- modules in a manufacturing environment. The recently pub-
entation (Q1) can be classified as the changes in the structure of lished RAMI4.0 standard, DIN SPEC 91345:2016 [8], provides
a system, the way it operates, and its expected outcomes. The the foundation for modularization of manufacturing via I4.0
first involves a transition from a hierarchical to a network struc- Components. Through reconstructing classic monolithic archi-
ture. The equipment/functional hierarchies will be replaced with tectures as reconfigurable networks of ‘Lego-brick’ (e.g., I4.0
networks of standardized components/services, where the links Components), modular design enables decentralization and de-
represent information carriers, edge connectivity, APIs, enter- mocratization of manufacturing to lower cost, improve effi-
prise service bus, etc. In line with this revolution, manual and ciency, and drive innovation. Major R&D and implementation
centralized workflows will be transformed into digital and de- efforts in this context therefore include (1) further development
centralized operations/decisions, supported by end-to-end col- of the notion of I4.0 Component proposed by DIN SPEC
laboration between all entities throughout the value chain. This, 91345:2016 to conform with ISO/IEC 18384:2016; (2) develop-
in turn, calls for new business models, analytical tools, compu- ment of passive and active infrastructures to enable plug-and-
tational models, and interaction protocols [15]. The expected play reconfigurability and module/infrastructure evolution [2];
outcomes of such a digitized and decentralized system will be and (3) validation through field experiments. The GE Power
better efficiency, agility, customer-centricity, cross-platform plants, for example, have already adopted the notion of modular
functionality, complexity-handling, and ability to interact with design for plug-and-play connectivity between manufacturing
external/heterogeneous systems. These outcomes all contribute modules (e.g., gas turbine; steam turbine; heat rejection system;
to the ever-increasing need for business agility [2]. electrical system; controls) [26].
This transition essentially requires a wide range of R&D and
C. Data-Driven Analytics
implementation efforts (Q2), which may vary depending on the
status of an entity in an Industry 4.0 environment (e.g., service Industry 4.0 is characterized by the notion of ‘digital twin’,
provider; physical resource provider; consumer). A preliminary an offspring of comprehensive digitization, ubiquitous connec-
step is to develop a modular and service-oriented enterprise ar- tivity, and big data. In this context, data-driven analytics are pro-
chitecture, as a ‘blueprint’ for guiding this transition, supported gressively playing a pivotal role in successful design and imple-
by new analytical tools and standards for emerging Industry 4.0 mentation of enterprise architectures. New analytical tools and
challenges such as interoperability and cybersecurity [3]. To bet- techniques are indeed required to address the emerging Industry
ter quantify the benefits and impacts of new and existing enter- 4.0 challenges such as cybersecurity (e.g., ISA/IEC-62443 [27];
prise architectures, however, more quantitative methods must be C2M2 [28]), safety, interoperability, connectivity [7], decentral-
developed. ized process control (e.g., agent-based modeling and simulation;



Authorized licensed use limited to: Steve Bishop. Downloaded on June 07,2020 at 21:13:18 UTC from IEEE Xplore. Restrictions apply.
holonic manufacturing), and online/ evolutionary computing [2] R. Dove, “Agile systems-engineering AND agile-systems engineering”.
(e.g., machine learning; real-time optimization). INCOSE INSIGHT vol. 17, pp. 6–10, 2014.
[3] L.D. Xu, W. He, S. Li, “Internet of Things in Industries: A Survey,” IEEE
D. International Standards Transactions on Industrial Informatics, vol. 10, pp. 2233-2243, 2014.
[4] A. Levis, “Systems architecture,” in Handbook of Systems Engineering
Several organizations and working groups are currently de- and Management, Second Edition (A.P. Sage & W.B. Rouse., Eds.),
veloping different concepts/frameworks/architectures for Indus- Wiley, 2009, pp. 479–506.
try 4.0 enterprises; independent of each other. Additionally, [5] T. Williams, “The Purdue enterprise reference architecture,” Computers
there may be a need for a ‘super-architecture’ or ‘architecture- in Industry, vol. 24, pp. 141–158, 1994.
of-architectures’ that governs the interactions of different sub- [6] “IEC 62264:2013, ISA95 – Enterprise-Control System Integration,”
2013.
systems (e.g., organizational architecture, communications ar-
chitecture, functional architecture, service architecture, infor- [7] “Industrial Internet Reference Architecture (IIRA),” Industrial Internet
Consortium, 2015.
mation architecture, etc.) Existing frameworks such as The
[8] “DIN SPEC 91345:2016-04, Reference Architecture Model Industrie 4.0
Open Group Architecture Framework (TOGAF) [29], the Uni- (RAMI4.0),” DIN Deutsches Institut für Normung e. V., Berlin, 2016.
fied Architecture Framework [30], and OPC Unified Architec- [9] “ISO/IEC 18384:2016, Information technology—Reference Architecture
ture (UA) [31] may be regarded as different ‘facets’ of such a for Service Oriented Architecture (SOA RA),” 2016.
super-architecture, although this topic requires further investiga- [10] D. Trentesaux, “Distributed control of production systems.” Engineering
tion. As mentioned in Section I, this manuscript discusses some Applications of Artificial Intelligence, vol. 22, pp. 971–978, 2009.
of the work for Phases 1 and 2 of the three-part PRIAM NGEA [11] H. Van Brussel, J. Wyns, P. Valckenaers, L. Bongaerts, and P. Peeters,
project. The next phase of the project will plan to bring together “Reference architecture for holonic manufacturing systems: PROSA.”
a group of interested parties, from industry, government, stand- Computers in Industry, vol. 37, pp. 255–274, 1998.
ards organizations, and academia, to investigate the foundations [12] “IEC 61512, Batch Control,” 1997.
of next generation enterprise architectures, and initiate the pro- [13] “IEC 62890, Life-cycle management for systems and products used in
cess of creating new standards for enterprise integration and industrial-process measurement, control and automation,” 2016.
control that will fully leverage the capabilities offered by the In- [14] “ISO/IEC/IEEE 42010, Systems and software engineering—Architecture
dustrial Internet and cyber-physical systems. description,” 2011.
[15] D. Wu, M.J. Greer, D.W. Rosen, and D. Schaefer, “Cloud manufacturing:
E. Architecture Analysis Methods Strategic vision and state-of-the-art.” Journal of Manufacturing Systems,
vol. 32, pp. 564–579, 2013.
One major challenge of determining the best enterprise ar- [16] “Reaping the Rewards of Your Service-Oriented Architecture.” IBM
chitectures developed for Industry 4.0 is the lack of quantitative Global Services, 2008.
architecture analysis methods. While Martin [32] identifies pos- [17] A. Arsanjani, “Service-oriented modeling and architecture,” IBM
sible architecture analysis methods, many are qualitative such as Technical Library, 2004.
focus group survey and Delphi methods. These methods make it [18] “Open Group Standard SOA Reference Architecture Technical
Standard,” The Open Group, 2011.
difficult to quantify the benefits of one architecture over another,
an important step in deciding which architecture to convert to [19] J. Delsing. IoT Automation: Arrowhead Framework, CRC Press, 2017.
for Industry 4.0. A more comprehensive analysis would be to [20] “OASIS Reference Model for Service Oriented Architecture 1.0,” OASIS
Standard, 2006.
experimentally test the various enterprise architectures in a man-
[21] “ISO/IEC/TR 30102, Information technology — Distributed Application
ufacturing environment. However, this would be extremely Platforms and Services (DAPS) — General technical principles of Service
time-consuming and expensive, considering the work involved Oriented Architecture,” 2012.
in setting up the proper hardware and software to represent mul- [22] X. Xu, “From cloud computing to cloud manufacturing,” Robotics and
tiple enterprise architectures. The development of a quantitative Computer-Integrated Manufacturing, vol. 28, p. 75–86, 2012.
architecture analysis method requires further investigation. In [23] Y. Lu, F. Riddick, and N. Ivezic, “The Paradigm Shift in Smart
addition to the project work previously mentioned, case studies Manufacturing System Architecture,” 2011.
will be conducted in Phase 2 as a method of providing a more [24] M. Meier, J. Seidelmann, and I. Mezgár, “ManuCloud: the next-
objective analysis of the existing and newly-created architec- generation manufacturing as a service environment.” European Research
tures. Consortium for Informatics and Mathematics News, vol. 83, pp. 33–34,
2010.
ACKNOWLEDGMENT [25] “Predix: The Industrial Internet Platform.” GE Digital, 2015.
[26] “Powering the World.” GE Power, 2015.
The authors would like to extend their gratitude to Dr. Albert
[27] “ISA/IEC 62443: Industrial Network and System Security,” 2010.
T. Jones from the U.S. National Institute of Standards and Tech-
[28] “Cybersecurity Capability Maturity Model (C2M2).” U.S. Department of
nology (NIST) and Dr. Eswaran Subrahmanian from Carnegie Energy.
Mellon University for sharing their insights and expertise, and
[29] “The Open Group. TOGAF 9.1.,” 2011.
to the GE-Purdue PRIAM for their support during the course of
[30] W. Pessemier, G. Deconinck, G. Raskin, P. Saey, H. Van Winckel, “UAF:
this research. a generic OPC Unifed Architecture Framework”, Proceedings of SPIE -
The International Society for Optical Engineering, vol. 8451, 84510P,
REFERENCES 2012.
[1] “Factories of the future.” International Electrotechnical Commission [31] “OPC Unified Architecture,” 2008.
(IEC), 2015. [32] J. Martin, “Architecture Evaluation Framework,” in Trade-off Analytics,
First Edition (G. Parnell), Wiley, 2016.



Authorized licensed use limited to: Steve Bishop. Downloaded on June 07,2020 at 21:13:18 UTC from IEEE Xplore. Restrictions apply.

You might also like