Documentos de Académico
Documentos de Profesional
Documentos de Cultura
Architecture Framework
MODAF
By:
Jason Caruso
Tefe Jakova
Kristin (Kramer) Mills
Aaron Varrone
CIS 695 Enterprise Architecture
Dr. Richard McCarthy
Spring 2010
Quinnipiac University
Page 2 of 66
TABLE OF CONTENTS
Overview and Scope of MODAF .................... 3
History of MOD and MODAF ....................... 28
Strengths of MODAF ................................. 31
Weaknesses of MODAF ............................. 32
Comparison of MODAF to Other Frameworks 32
Compliance Issues ................................... 37
MODAF Users .......................................... 38
Upcoming MODAF Versions ....................... 44
Analysis of Enterprise Architecture ............. 44
References.............................................. 64
Page 3 of 66
Operational Views (OVs) are used to describe the tasks and activities,
operational elements, and information exchanges required to conduct
business and operational activities
Service Oriented Views (SOVs) are used to describe the services, (i.e.
what is provided to the customer), required to support the tasks and
activities described in the Operational Views
Page 4 of 66
Systems Views (SVs) are used to describe what happens when the
Operational and Service Orientated Views are implemented and,
thereby, define the solution.
Acquisition Views (AcVs) are used to describe what is needed and how
long it will take for the projects that will deliver the solution.
All Views (AVs) are used to provide a summary of the contents of the
architecture and any related terms/characters/abbreviations used.
Strategic Viewpoint (StV)
The strategic viewpoint defines the overall vision of the organization in order
to support military operations and is broken down into six views. The
strategic views are focused primarily on capability management. A
capability is described by MODAF as a specification or classification of the
ability of the enterprise (Ministry of Defence, 24 November 2008).
Capabilities are stable over time, but the requirements of capabilities can
change. A requirement capability says that a certain capability must be
achieved within a time period (Ministry of Defence, 24 November 2008).
While there are six views in the strategic viewpoint, they are not all
required. An organization may only need to use a few of them. Since
military capabilities change over time, views typically refer to a particular
period in time. If a capability undergoes a significant change, it is necessary
to reassess the views.
The first is StV-1 (Enterprise Vision), which provides the overall strategy for
the framework. A capability vision shows what is required to meet the
organizations strategy. MODAF version 1.2 calls for a diagram to show an
organizations goals and capabilities. An example is shown below.
Page 5 of 66
(MODAF, 2008)
While the StV-1 view only describes high level goals, the StV-2 view is
referred to as the capability taxonomy and breaks down each capability into
a hierarchy. Each capability describes what is to be done, but does not
specify how something is to be completed. Capabilities are often broken
down into structured lists and an approach such as unified modeling
language (UML) may be used (MODAF, 2008).
StV-3 is the capability phasing view, which deals with the realization of
capability during different time periods. This view strives to look for
differences or overlapping areas of capability, by plotting projects and
support over time. The following diagram shows where capabilities are
missing over a period of two years.
Page 6 of 66
Page 7 of 66
Page 8 of 66
Page 9 of 66
Page 10 of 66
Page 11 of 66
The OV-5 Operational Activity Model describes the operations that are
normally conducted in the course of achieving a mission or a business goal
(Ministry of Defence, 12 February 2009). OV-5 describes the different
business processes and workflows, and gives a clear definition of roles and
responsibilities. OV-5 is usually coupled with OV-2. OV-5 focuses on the
operational activities while OV-2 focuses on the operational nodes. OV-5
was created to describe military activites, but it is well suited to work with
non-military activities as well. It is used to describe the operations or
actions that are normally conducted in achieving a mission or a business
goal (Ministry of Defence, 12 February 2009).
OV-6 is really made up of 3 different views. OV-6a which is the Operational
Rules Model, specifies operational or business rules that are constraints on
the way that business is done in the Enterprise. OV-6b is the Operational
State Transition Description which is a graphical method of describing how
an operational node or activity responds to various events by changing its
state. Finally, OV-6c is the Operational Event-Trace Description, and it
provides a time ordered examination of the information exchanges between
participating operational nodes as a result of a particular scenario. The
three views describe the dynamic behavior and timing performance
characteristics of an architecture. Views OV 1-5 gives us knowledge of the
operational nodes, activities, and information exchanges (Ministry of
Defence, 12 February 2009). But OV-6a,b,c shows what happens when an
activity occurs in the architecture and how that architecture behaves. When
we send a message from node A to node B, it is important to know what
kind of response we will receive.
The OV-7 Information Model addresses the information perspective on an
operational architecture (Ministry of Defence, 12 February 2009). It
documents the business information requirements and structural business
process rules of the architecture (Ministry of Defence, 12 February 2009).
An OV-7 is very useful when different organizations use the same terms to
refer to completely different things.
Service Oriented Viewpoint (SOV)
The third viewpoint that differentiates MODAF from other frameworks is the
service oriented viewpoint. It identifies views that can be used in a service
oriented architecture and was derived from the NATO Architecture
Framework (Ministry of Defence, 27 November 2008). The service oriented
viewpoint consists of seven views.
The first view is SOV-1, which is Service Taxonomy. The NATO Architecture
Framework (NAF) equivalent is NSOV-1. The Service Taxonomy view
Page 12 of 66
The following UML diagram shows how a service hierarchy might be set up.
Page 13 of 66
Page 14 of 66
SOV-4b is the Service State model. It shows the different possible states
that a service may have and is used in service specification (Ministry of
Defence, 27 November 2008). SOV-4b can be illustrated with UML or by
another modeling language. The objects represented by this view are only
the service and the service state machine. The last view on the fourth level
is SOV-4c, the Service Interaction Specification. This view shows the
interactions of services with outside elements, the sequence of these
elements and any dependencies between the interactions. It is usually
signified through a UML sequence diagram, with objects such as the
services, interfaces between services, lifelines, and the consumers (Ministry
of Defence, 27 November 2008). The following example shows a sequence
diagram.
Page 15 of 66
Page 16 of 66
network components like switches and routers. SV-2 focuses on the physical
characteristics of a link by specifying attributes. The goal of SV-2 is to show
how systems are connected, what interfaces each system exposes (ports),
the type of hardware interface used and the information transmitted across
the interface. SV-2 expands on SV-1 by giving more detail of the physical
characteristics of those interactions specified in SV-1 that are between
systems. The first SV-2 view, SV-2a System Port Specification, which
specifies the ports on a system and the protocols used by those ports when
communicating with other systems. SV-2a is used to fully describe the
interface protocols and hardware specifications of each port on a system.
SV-2a lists the names of the ports, the interface protocol used, and the
physical port specification. The second SV-2 view is SV-2b System to
System Port Connectivity Description, which specifies the communications
links between systems and may also list the protocol stacks used in
connections. SV-2 gives a precise specification of a connection between
systems. Lastly, the third SV-2 view is SV-2c System Connectivity Clusters,
which define how individual connections between systems are grouped into
logical connections between parent resources (Ministry of Defence, 12
February 2009). An example of the SV-2c is shown below.
Page 17 of 66
gives a description of the necessary data flows that are input (consumed) by
and output (produced) by each resource. It makes sure that inputs receive
the information that they are requesting and at the proper level of detail.
SV-4 is the behavioral counterpart to the SV-1 (in the same way that OV-5
is the behavioral counterpart to OV-2) (Ministry of Defence, 12 February
2009). A simple example of an SV-4 is shown below.
Page 18 of 66
The sixth view is the SV-6 Systems Data Exchange Matrix view. It specifies
the characteristics of the system data exchanged between systems with the
focus on data crossing the system boundary (Ministry of Defence, 12
February 2009). SV-6 is the physical equivalent of the logical OV-3 table
and provides detailed information on the system connections that implement
the information exchanges specified in an OV-3. If the information is nonautomated, it is captured in the SV-1 and SV-3 only. SV-6, like SV-5, uses a
tabular format to present the data. The focal point of SV-6 is on how the
system data exchange is affected, in system-specific details covering
periodicity, timeliness, throughput, size, information assurance, and security
characteristics of the exchange. Also represented in the table are the
system data elements, their format and media type, accuracy, units of
measurement, and system data standards. The seventh view is the SV-7
Resource Performance Parameters Matrix. SV-7 depicts the performance
characteristics of a Resource (ie. system, role or capability configuration). It
builds upon SV-6, and therefore should be developed in conjunction. SV-7
takes the information presented in SV-1 and expands upon it by depicting
the characteristics of the resources shown in SV-1 (Ministry of Defence, 12
February 2009). To help facilitate a mission or goal, SV-7 is used to
communicate which characteristics are considered most crucial. The format
used in SV-7 is also that of a table.
The eighth view is the SV-8 Capability Configuration Management view. SV8 presents a whole lifecycle view of a resource, describing how its
configuration changes over time (Ministry of Defence, 12 February 2009).
SV-8 s main function is to show change over time. Sv-8, when linked
together with other evolution views such as AcV-2, StV-3 and TV-2, provides
a rich definition of how the enterprise and its capabilities are expected to
evolve over time (Ministry of Defence, 12 February 2009). These types of
views are good for presenting a potential future snapshot to management.
The ninth view is the SV-9 Technology and Skills Forecast. This view defines
the underlying current and expected supporting technologies and skills
(Ministry of Defence, 12 February 2009). SV-9 will show future trends in
technology and skills that may directly impact the architecture. This view is
used to forecast future investments in technology that will be needed. Given
the future-oriented nature of this product, forecasts are typically made in
short, mid, and long-range timeframes (Ministry of Defence, 12 February
2009). An example of Sv-9 may show the forecast for an operating system
like Microsoft; where the newest version may be in the short-term and
future upgrades mid and long term forecasts.
The tenth view is SV-10 which is actually made up of three models. Each of
the three views may be used to describe the dynamic behavior and
Page 19 of 66
Page 20 of 66
The final view in the SV grouping is the SV-12 Service Provision. SV-12
specifies configurations of resources that can deliver a service and the levels
of service those resources can deliver in different environments (Ministry of
Defence, 12 February 2009). SV-12 is related to the Service Oriented Views
(SOVs) in MODAF because the SV-12 shows how those SOVs affect the
architecture. SV-12 provides the mapping from services to the resources
that provide those services. The service attributes defined in SOV-1 can be
given values in SV-12 and can therefore relate to the environment under
which those values are true (Ministry of Defence, 12 February 2009). The
diagram below shows a UML representation of SV-12:
Page 21 of 66
Page 22 of 66
Page 23 of 66
Page 24 of 66
Page 25 of 66
Page 26 of 66
November 2008). One can also use it as a place to record any findings and
notes while constructing the architecture. AVs also contain a glossary of
terms used, to help others easily understand and translate in the future.
MODAF states that AV is critical whenever an architecture is created or
modified because it is where the records and notes are stored, so the
information is available for future access and modifications (Ministry of
Defence, 19 November 2008). The AV viewpoint is made up of two views.
The first view is AV-1 Overview and Summary Information view. AV-1, as
the name implies, is a summary of all the information related to the
architecture in an executive-level designed for quick reference and
comparison. AV-1 includes assumptions, constraints, and limitations that
may affect high-level decisions relating to an architecture-based work
program (Ministry of Defence, 19 November 2008). AV-1 also serves as the
planning guide for when the architecture is being built, and answers the
who, what, where, when, why and how of the plan. AV-1 is critical when
building a MODAF architecture because it keeps everything on plan and in
order, similar to blueprints for a house. AV-1 provides a summary of these
descriptions: Architecture Project Identification, Scope, Purpose and
Perspective, Context, Status, Tools and File Formats Used, Assumptions and
Constraints, and Data Completed (Ministry of Defence, 19 November 2008).
AV-1 may also include Findings and Costs. AV-1 can be a dynamic
document as the architecture is being built, however a final draft should be
completed once the architecture is finished to document the final product.
An example of AV-1 is shown below.
Page 27 of 66
Page 28 of 66
Page 29 of 66
Page 30 of 66
MODAF 1.0
(MODAF, 2005)
MODAF VERSION HISTORY
There have been several small revisions and one major revision to MODAF
since its introduction in 2005. Minor revisions usually involve changes to
architectural elements of a view, changes to the name of a view, or additions
or changes to the overall meta-model that do not dramatically change the
purpose of a particular view (Ministry of Defence, 18 May 2009). Major
revisions can involve the addition or removal of views or viewpoints, major
changes to the overall meta-model or the renaming of a viewpoint (Ministry
of Defence, 18 May 2009).
In April of 2007, Version 1.1.001 made some minor revisions to section OV2 and changed the title of the section from Operational Node Connectivity
Description to Operational Node Relationship Description. Version 1.1.002
changed the layout of the first page, changed the order of the menu on the
side, made a note about the delay in revising M3 (the MODAF Meta Model),
Page 31 of 66
and added a section for Configuration Control Policy. Revisions 1.1.003 and
1.1.004 revised the Navigation and allowed the user to print the document
without the navigation links (Ministry of Defence, 18 May 2009).
Version 1.2.000 was a major revision of MODAF and was released in June of
2008. It suggested some alternate names for some of the views, introduced
a new viewpoint: the Service Oriented Views, and allowed software and
artifacts as a resource type. This version also permitted the use of services
in some of the operational views, and added an interactive Frequently Asked
Questions section (Ministry of Defence, 18 May 2009). The new service
oriented viewpoint was largely based upon the NATO Architecture
Framework as discussed earlier in this report (Ministry of Defence, 27
November, 2008).
Revision 1.2.001 made some minor updates to M3, and some corresponding
view changes to SV-5 and SV-12. Revisions 1.2.002 and 1.2.003 made
some additional changes to M3. Revision 1.2.002 added software to SV-11
and added some requirements to SV-12, OV-1, and StV-1. Revision 1.2.003
added data and information elements to several views.
STRENGTHS OF MODAF
As mentioned earlier, MODAF is partially derived from the United States
Department of Defense Architecture Framework (DoDAF). MODAF includes
several components of DoDAF, including the operational viewpoint, systems
viewpoint, technical standards viewpoint, and the all views. The use of the
same views offers some compatibility between the two frameworks. MODAF
usually offers several options for representation in each view. For example,
an architect may elect to use a tabular representation or a specific modeling
tool such as UML.
MODAF is also advantageous to other defense frameworks because it offers
three additional viewpoints: the strategic, service oriented and acquisition
viewpoints. The strategic and acquisition views offer more in terms of
strategy and program management. They also show how capabilities are
shared and used in various operational activities. The service oriented view
supports the use of service oriented architecture (SOA). SOA is an approach
to integrating information technology services. The strategic, acquisition,
and service oriented viewpoints were discussed earlier in this document in
further detail.
Another strength of MODAF is that it is compliant and useful with the Unified
Modeling Language (UML). This allows MODAF to be more flexible and
useful with other architectures, like DoDAF. In March 2008, the UPDM
Group was re-formed by members of INCOSE and the OMG (Object
Page 32 of 66
Management Group) to create the Unified Profile for DoDAF and MODAF
(Hause, 2008).The group is working on ways to create a common
architecture language so that different architecture can be viewed together.
MODAF was one of the first architecture frameworks to work with UML.
WEAKNESSES OF MODAF
One of the main disadvantages of MODAF is its length. MODAF focuses on 7
different viewpoints, each of which has a number of different views. It
includes a great deal of material and requires a user to invest a considerable
amount of time to understand it. The framework itself specifies a number of
deliverables, some of which are required and others are optional. An
organization using MODAF would have to have a clear understanding of their
architecture before creating these deliverables. MODAF itself does not detail
a process for producing the required pieces of the framework, unlike a
framework such as TOGAF.
Before an enterprise can start to create viewpoints using MODAF, there are
several things they must sort out. First of all, the organization has to
understand the overall problem that they are attempting to solve. The
architects need to understand the overall strategy of the organization and
what they wish to ultimately achieve from the architecting process. It is also
critical that the process has stakeholders. Without stakeholders, the project
for creating an architecture is likely to fail, as there needs to be cooperation
throughout the enterprise, which can be facilitated by the stakeholders.
MODAF is not particularly flexible in that it is not easy to change the
structural representation of a view or viewpoint or an approach for a specific
view. Major changes need to go through an approval process. The MODAF
Steering Group must review and approve such changes (Ministry of Defence,
18 May 2009). Minor changes can either be reviewed and approved by the
Steering Group or by the MODAF Technical Group (Ministry of Defence, 18
May 2009). This can be challenging for organizations to do as they might
not necessarily fit the overall model of the Ministry of Defence. MODAF was
designed specifically to meet their needs and may not apply as well to
others.
COMPARISON OF MODAF TO OTHER FRAMEWORKS
DoDAF
As mentioned earlier, although MODAF is based upon the DoDAF framework,
MODAF is more based on MoD lifecycle processes, where views of the
Page 33 of 66
Description
MODAF
DODAF
Page 34 of 66
Page 35 of 66
The All Views in MODAF largely aligns with the Preliminary Phase
and Phase A: Architecture Vision in TOGAF
Page 36 of 66
Focus
Applicable MODAF
Models
Preliminary
AV-1, OV-1
Vision
AV-1, OV-1
Business Architecture
Information Systems
Architecture: Data
Architecture
OV-7, SV-11
Information Systems
Architecture:
Applications
Architecture
Technology Architecture
Opportunities &
Solutions
Migration Planning
SV-8
Implementation
Goverance
AV-1, OV-1
Change Management
SV-9, TV-2
Page 37 of 66
Page 38 of 66
Page 39 of 66
Page 40 of 66
DSTL
The Defence Science and Technology Laboratory (DSTL), which is an agency
of the UKs Ministry of Defence and its center of scientific excellence, houses
one of the largest groups of scientists and engineers in the UK. The mission
of DSTL is simple, which is to ensure that the UK Government and Armed
Forces have access to world-class scientific advice and from there can
deliver defensive research, specialist technical services, and the ability to
track global technological developments by establishing defensive policies
making and operations. In order to provide the best possible solutions to
the government, as well as bring exciting and innovative technologies to the
table; DSTL is required to build their architecture around MODAF (M2
Presswire, 2010).
BMT Hi-Q Sigma
BMT Hi-Q Sigma, a company which helps to deliver complex programmes
through the integration of programme management, and systems
engineering, uses a modern driven requirements approach by utilizing
MODAF (BMT Hi-Q Sigma, 2010). A recent example of this includes the
company being contracted in 2007 to produce a User Requirements
Document (URD) for the Apache Army Helicopter Mark 1 (AH). The intention
of this document was to produce a baseline capability and articulate what
was needed in order to meet the capability throughout its intend lifespan.
In order for the company to gather the future capability requirements, a
number of Operational Views from MODAF were produced, which allowed HiQ Sigma to scope the various aspects of the equipment required.
The following is a list of key benefits that MODAF provided to the company
for this particular project:
Page 41 of 66
Page 42 of 66
The System and Technical Views were used to represent the user
requirements and sample configurations that satisfied the operational
requirements represented in the Operational Views. This showed the
potential deployments and which existing systems could be used as a
system-of-systems in order to provide the required information
management and communications (Davis & Washington, 2009).
The Acquisition Views were easily understood in Gantt-style diagrams that
gave an immediate indication of programmed interactions and potential
scheduling conflicts. For the new system, the platforms that were going
out of service were immediately identified so that their replacement could
be identified and informed of the new requirements. Additionally, some
platforms were able to be scheduled for the fitting of the legacy
equipment and eventually be considered as replacements and as the
correct fit for the system (Davis & Washington, 2009).
In the end, the complete set of the six MODAF views were selected:
Page 43 of 66
Page 44 of 66
Page 45 of 66
The main purpose of this analysis is to study the artifacts, minimize their
coupling, and maximize the cohesion between them. Interface analysis is
also used for the organizational and strategy layer, and also the study of
homogeneity of the structure and artifacts (Bucher, Fischer, Kurpjuweit, &
Winter).
Heterogeneity analysiss- purpose is to locate the elements which could
increase efficiency of architecture homogeneity. Homogeneous structures
have low costs for maintenance, software, hardware and that all issues are
treated the same (Bucher, Fischer, Kurpjuweit, & Winter).
Complexity analysis and interface analysis go together- Its goal is to lower
the enterprise architecture complexity by studying the complexity measure
and dependency between the components. Compliance frameworks,
regulations by law, and market standards have gained tremendous
importance during the last couple of years (e.g., Solvency II, Basel II and
the Sarbanes Oxley Act) (Bucher, Fischer, Kurpjuweit, & Winter).
Compliance analysis- is analyzed if some policies such as process and data
ownership are correctly identified in the organizational level of abstraction
and if authorization and recovery mechanisms are successfully implemented
(Bucher, Fischer, Kurpjuweit, & Winter).
Cost analysis- calculates costs of enterprise architecture artifacts such as
new implementations or sales. It is very important to know IT costs and how
to distribute them to products, services, processes and other layers (Bucher,
Fischer, Kurpjuweit, & Winter).
Benefit analysis- calculates the contribution of different parts of the
companys units, systems or products to the whole organization. Every
organization has set goals and balanced scorecards, which are good
indications of how they perform (Bucher, Fischer, Kurpjuweit, & Winter).
The whole purpose of a system architecture is the link and collaboration
between its components; and that it works by using interaction protocols
among the components (Bucher, Fischer, Kurpjuweit, & Winter). The
interaction between the protocols should be error free, so there wont be any
deadlocks, logical inconsistencies or not controlled interactions. Another way
to control that all interactions go according the plan is by using Architecture
Description Language (ADL)
Model checking techniques are very important since the enterprise
architectures are widely distributed in organizations, and often too many
processes are running at the same time, interacting with each other but with
no standardized techniques to be analyzed. As a result of no coordination;
Page 46 of 66
they often could spend value time and resources (Bucher, Fischer,
Kurpjuweit, & Winter).
The Architecture Description Language provides well defined and distinct
syntax for processes, components, ports, interfaces, channels, and protocols
of interaction (Sircar & Kott, 2000). ADL is similar to other programming
languages that run code to check the system characteristics. This system
analysis tool can give important feedback on enterprise control. The strong
capabilities of the tool are to be able to eliminate errors in designs during
different situations such as when multiple components are interacting with
each other.
The figure below represents the ADL method:
Page 47 of 66
Quantitative
Analytical
Analysis
dimensions
Simlulation
Functional
(Lankhorst, 2005)
The analysis of architecture has to be done looking from different
perspectives. The first step is by comparing inputs and outputs; i.e.
functional versus quantitative (Lankhorst, 2005).
By performing the Functional Analysis, the architecture is looked upon from
the inside. This way data is generated on how a system conforms to the
particular architecture functions and what would be the causes of an
architecture change, and finally to check if the architecture is working or
being followed step by step. When a functional analysis is performed,
answers are not received to the questions on how fast will it be done or
Page 48 of 66
how much it will cost. These questions get a response when a quantitative
analysis is done, although it is very difficult to answer those kinds of
questions from an architectural model since the quantitative analysis of the
framework doesnt offer an option to regulate and correlate quantitative
results taken from the present detailed analysis methods.
The Functional and the Quantitative Analysis utilizes two different techniques
as well: the analytical and simulation technique. The simulation would show
the results of a models execution. The simulation is divided into Functional
and Animated simulation which looks at the performance and actions of the
system and as a result, helps the architects look and find ways to adapt the
changes, and also helps the stakeholders to have better knowledge on the
architecture. When running the functional simulation, architects and
stakeholders could be faced with interpretation problems; which could be
caused from the situations created from architecture abstraction.
Formal and Analytical analysis methods defer from the Simulation by
providing distinctive and reproducible results. By running the Analytical
techniques, architects and stakeholders can achieve better results than the
quantitative simulation during the quantitative analysis, which helps the
architects detect the bottlenecks of the framework, its performance, and also
helps to make better decisions if many options are available during the what
if analysis. When utilizing the analysis techniques, it will also answer the
question if the existing techniques should be used or if it will be necessary
create new ones. If architects choose to create new techniques, there will be
two available options: buy it or build it? From the Buy it question, offspring
two other questions: Which available technique should be chosen?, and
How to apply it?, and from the other perspective from the Build it
question, two more questions will be asked: How will the development be
implemented? and For which of the issues will the technique be built for?
(Lankhorst, 2005).
Quantitative Analysis
The enterprise architectures concern is to have a detailed description of how
all the important fundamentals of the enterprise operate together, such as
business products, processes, and software applications. From the
quantitative point of view, these fundamentals are related to each other in
multiple ways. Often, there is a misunderstanding that quantitative analysis
would get too complicated and give high level details when ran on the
architectural level. In the other side, performance engineers think that
quantitative (non-functional) aspects are equally important as functional
aspects to all stages of the system design and enterprise architectures, and
not just as a back-up idea. In enterprise architecture modeling, the
Page 49 of 66
quantitative sides are put on a global level, where the weight falls on the
structure which in enterprise architectures is used as a mechanism to
regulate the quantitative properties of organizations and their systems
(Lankhorst, 2005). When doing a quantitative analysis, there are several
issues to look at:
1. Quantitative analysis is used as an optimization tool for systems or
different processes by having a definite answer in case different design
alternatives are available.
2. Quantitative analysis could be used as a tool that could help architects
take the right measures in order to be able to prevent negative effects
of changes.
Page 50 of 66
customer has multiple requests this represents the time that it takes
for a request to be completed; or in the case of an information system,
the time it takes to finish a batch of a process.
Product view - In this case the stakeholders are: product managers
and operational managers. Processing time represents the time it
takes to complete a product, excluding the wait, or delay time. In a
computer system it is represented by the time when the CPU is busy.
System view - In this case the stakeholders are: owner and managers.
Throughput is represented by the number of requests that are
completed by the system in a time unit. For example: the number of
scanned pages in an hour. In system view, the term processing
capacity represents the maximum attainable throughput depending on
the number of available resources and their capacity.
Resource view - In this case the stakeholders are: resource managers
or capacity planners. The representation of time in percentage
calculated by using the time the resource is busy compared to total
operation time is known as Utilization. Utilization is a great tool that
shows the effectiveness in time of resources. By analyzing the time of
the resource it takes to be completed, bottlenecks could be found as
well. In an information system architecture, the network load time is
represented as utilization (Lankhorst, 2005).
Sometimes performance measures of different views could be interrelated,
and sometimes can be in conflict with each other when trying to increase the
performance of the belonging system. For instance, an example of high
throughput which as a result could lead to higher resource utilization. This
situation could be helpful to a manager because it increases the response
time, but it is not good for a users time. As a result, system performance
optimization is a very important and key issue (Lankhorst, 2005).
Quantitative Analysis - Performance Analysis Techniques for
Architectures
Although many applications are used to create models of enterprise
architectures, many of them, if not most of them do not analyze
architectures from the quantitate point of view. Multiple performance
analysis techniques have been planned for computing, manufacturing, and
telecommunication systems; and are therefore very efficient with a relatively
highly accurate estimate. Quantitative analysis can be used on all aspects of
enterprise architecture, beginning from the technical infrastructure layer, to
software applications, ending with the business processes supported by
these applications (Lankhorst, 2005). Additionally, enterprise architecture
Page 51 of 66
layers should be related to each other, and this happens also from the
quantitative side. For example, lower layers influence the higher layers and
higher layers influence the lower layers.
Infrastructure layer infrastructure domain is strongly affected from
the performance evaluation of computer and communication systems. The
hardware specifications in a system are described from the queuing models;
in the meanwhile the applications workload are controlled from the abstract
stochastic arrival process (Lankhorst, 2005).
Application layer the discipline of performance engineering of
software applications are considered at a global level, from which the
analysis of performance characteristics of some architectural styles and also
the performance issues are based from the SAAM method, which is a fivestep method analyzing software for architectures. Architecture description
language (ADL) suggests the queuing of models from software architecture.
Compositionality, which is another important architecture issue, describes
that the performance of a system could be expressed as the sum of the
performance of each of its components. Stochastic extension of process
algebras could be used for the compositional performance analysis, but they
are very computationally intensive. Process-algebra-based approaches allow
compositional specifications of performance models, but the analysis results
are not always compositional (Lankhorst, 2005).
Business layer there are many business and general process
modeling tools, such as Arena and ExSpect, that are used for quantitative
analysis using discrete-event simulation. The simulation is very helpful to
interpret the data, but if it is inserted at a very detailed level, it might be
very difficult to read for inexperienced users. Different companies offer
different analytical methods that help with completion times and critical path
analysis of processes, simulation based performance analysis, and very
complicated computational analytical solutions (Lankhorst, 2005).
Quantitative Analysis - Quantitative Modelling
ArchiMate language is a tool used for quantitative modeling of serviceoriented enterprise architectures. As mentioned above, the performance
views are the same for any modeling language, and as a result can be
converted to the ArchiMate language. ArchiMate uses the following
quantification prospects: triggering, access, realization, and used by.
Triggering relations are easily proven and very important in determining the
behavior of dynamic systems. Some quantitative methods that work for
business process-oriented modeling formalism cannot be applied to
ArchiMate because of two reasons:
Page 52 of 66
(Lankhorst, 2005)
Reliable input data is the most important step of quantitative analysis. The
data can be retrieved from different sources. In an already existent system,
measurement could be very accurate, but retrieving the data correctly is
fundamental. It is not always clear on what needs to be measured, what
Page 53 of 66
must have a defined number or number of measurements, and last but not
least whether the measurements should be retrieved from different
situations so that there is a variety of data. What needs to be noted is that
while the system or organization still has work in progress, measurements
shouldnt be taken. For instance, if data is needed it should be retrieved
from the specifications of the components that are used or from another way
which is to get the data from similar architectures (Lankhorst, 2005).
Quantitative Analysis - Quantitative Analysis Technique
The quantitative analysis technique is composed of three steps:
1. Model normalization
2. Top-down workload calculation
3. Bottom-up performance calculation
By using different formulas during these three steps, it would be easier to
study the effects of input parameter on performance (Lankhorst, 2005).
Functional Analysis
Existing techniques in formal methods could be used for functional analysis.
The formal methods are mathematically precise because they use different
systems and are tested in various situations. The following aspects: static,
structural and dynamic, or behavioral are distinguished in functional analysis
of frameworks. The signature is the basis when trying to analyze the static
structure of architecture. The symbolic representation of the structural
elements and their relationships abstract from other architecture aspects
such as rationale, visualization and pragmatics. The most important feature
is the separation of concerns which help in the improvement of the
complexity of the architecture. The signature could be built in XML for a
variety of reasons, such as storage and communication, and at the same
time could be used as an independent piece, such as graphics for
visualization. The formal semantics of a symbolic model in a given
architecture could be described from a signature, but when interpreting it,
the details are quite obvious. A signature can have a large number of
interpretations because of the various available architectures that give
different meanings to the signature (Lankhorst, 2005). The techniques used
for static and dynamic analysis that will be described later on; will give a
better idea on how enterprise architectures should be interpreted on an
individual concept and by various relationships. These techniques are very
helpful for enterprise architects as it checks the correctness of their
architecture, increases clearness, and widens the architecture descriptions
with important information.
Page 54 of 66
Frame-based systems,
Semantic networks,
KL-ONE languages,
Object-oriented representation,
Semantic data models,
Type systems.
(Lankhorst, 2005)
Page 55 of 66
(Lankhorst, 2005)
The intermediary has just a commercial interest. They sell the product to
customers and makes sure that the paperwork is filled in a correct way, until
the customers agrees by signing the contract. After this point, the
intermediary works with the customer or insurance company in case of
premium issues. For the insurance company, it would be helpful to eliminate
the intermediaries, but before doing that, the architects need to investigate
quickly and precisely on what the impact of this kind of change will be on the
company. As a start, it is required to analyze the relationships between the
different views and logical theory. The basis is a single architecture model,
which is connected to a signature used in the logical analysis. In the
signature there are roles such as collaborations, and there are domain
dependent types such as an insurance company and its customers. When
doing the structural analysis, architects need to study every single available
relationship that could be in the architectural design, and make sure if the
change will affect any of the relations. For example, if a service provided by
an application changes, all the users that need to use that specific service
will be affected (Lankhorst, 2005).
As mentioned above, every relationship in the architecture model is
represented by a relation in the logic. The translation includes constraints
between the sorts and the relations, in order to make the correspondence
precise. For example:
x: Customer(x) -> Role(x)
xy: Participate(x,y) -> Role(x) /\ Collaboration(y)
Page 56 of 66
(Lankhorst, 2005)
The first relation shows that every Customer could be a Role; in the second
case the each Role can be part of Collaborations, and that Collaborations
could be Roles. This is just a simple example that shows the rules that are
written, which shows the effects of changes that are much more complex.
Those rules are ran by rules engine that automate the impact analysis.
Going back to the above example, if the insurance company will remove the
intermediaries, it will have to check all the relations where the intermediary
is involved since they will be affected by the break up. Because of these
changes, many business processes will be working on this because of the
interactions that are part of the collaborations, for example the Close
Contract process is part of them, which uses multiple applications, some of
which will be affected by the change.
Below shows the Impact (in red) on collaborations.
(Lankhorst, 2005)
This figure below shows the impact (in red) on Close Contract business.
Page 57 of 66
(Lankhorst, 2005)
This figure below shows the impact (in red) on applications.
(Lankhorst, 2005)
Functional Analysis Dynamic Analysis
Functional analysis techniques used for process algebras and data flow
networks are important for dynamic analysis of architectures. Sometimes
there are two or more roles acting on the same time, overwriting or
destroying one-anothers work, as a result would require one to create a
protocol to stop this from happening in the future. Functional behavior
analysis can be looked upon as formal methods in which a qualitative
analysis will check to find errors, have an improved consistency, and focus
on the logic of models.
Page 58 of 66
(Lankhorst, 2005)
In the structural concepts we have products, roles and objects; in the
structural relationships we utilize association; the behavioral concepts use
processes; and behavioral relationships use triggering. The structural and
behavioral concepts are linked using the assignment and access relations
(Lankhorst, 2005). Each employee has to follow the steps below, as a
procedure necessary to order the products:
1. Orders must be registered in the Order Registry before placing it. The
Order Registry is used only for administrative purposes. Since orders
could be made of many products, the Order Registry will be used to
check the order.
Page 59 of 66
2. As soon as the employee sends the order form, the supplier are
required to gather the products and distribute them to ArchiSell as
soon as possible.
3. When the supplier delivers the order, the employee will have to make
sure if there is an order matching this delivery. After confirming, the
order will be accepted.
4. Lastly, the employee will check the Product Registry to see who
ordered the product, and as a result will be the owner of the product.
(Lankhorst, 2005)
Signature
The signature of the business process architecture gives examples of the
signature that could be:
1.
2.
3.
4.
5.
6.
Role
Object
Employee
Product
Order Registy,
And Product Registry
(Lankhorst, 2005).
Page 60 of 66
Page 61 of 66
o codomain name=Order_Registry
Place_order_for_product
o domain name=Employee
o codomain name=Employee
Accept_product
o domain name=Employee
o domain name=Order_Registry
o codomain name=Employee
Register_product_acceptance
o domain name=Employee
o domain name=Product_Registry
o codomain name=Employee
o codomain name=Product_Registry
(Lankhorst, 2005)
Pseudo-language could be used to describe the interpretation of processes,
but matrices of input/outputs could be used as well.
Data Flow Networks
Data flow networks are used as a technique for the behavior of complex
systems, and it contains processes in which communicate data over lines.
Processes do the transformation of data inside a given system, and where a
line as First In First Out connects a maximum of two processes together.
When data is sent over the line from a process, it will arrive in an x time, in
the same order it was initially transmitted. Business functions could be
represented from data flow diagrams. This technique gives a general view
of the business and ends with a detailed analysis of all the functional areas
of interest, which are up to the requirements of the required level. For
instance, a top-down expansion could be used to analyze the target in a
precise way showing all the activities through diagrams. In the ArchiSell
example, from a data flow point of view, every process is described as an
independent data-consuming/producing entity, where these entities have
input and output ports. The data flow interpretation has a goal to study the
data flow inside a process and not the actors that perform the process
(Lankhorst, 2005).
The figure below helps interpret the data flow network as the following:
Not to include role information about roles and individuals within the
role sort. Data flow diagram shouldnt contain any details of actors or
process steps.
Stores will be represented as registries. They will store the
information which could be retrieved later.
Page 62 of 66
The input and output ports will have a value associated with them
representing the number that have to be sent or received
(Lankhorst, 2005).
Below shows ArchiSell as a data flow network:
(Lankhorst, 2005)
The following values are communicated:
list of products that have to be ordered;
list of products that have to be ordered;
order registry record;
list of products that have to be ordered;
supplier order;
list of products received;
order registry record;
list of products accepted;
list of products accepted;
product registry record;
product registry record;
order registry record;
order registry record
(Lankhorst, 2005).
The data flow diagram defines each data flow step, where the functions
transform the input values into output values. This function could be
described using pseudo-language and can derive an active simulation of the
Page 63 of 66
process architecture. The data flow diagram will also look at the actors and
how to assign them correctly to processes (Lankhorst, 2005).
Summarizing The Analysis of Enterprise Architecture
The relation between enterprise architecture models and architecture
analysis are very important, with two main classes of methods as the focal
point, which are the quantitative and functional analysis. Most performance
analysis are examined in a given domain inside the architecture. A new
approach was introduced for the distribution of workload and performance
measures through the layered framework model which could be used as an
analysis framework. The top-down and bottom-up technique are used to
analyze the performance of a document management system for storage
and retrieval of damage reports. Queuing formulas are used for the answer
times from lower to higher layers in which could increase the local service
times.
Meanwhile, the functional analysis techniques help us understand how to
better read architectures, reduce mistakes, and enrich architecture
descriptions with more important information in easier steps and more
accurately. In addition, the examination of the difference between static,
structural, dynamic, behavioral aspects, compared symbolic and semantic
models were looked at. In which, the most important part of symbolic
model is the signature which gives details about structural elements and
relationships. The semantic models integrate static and dynamic elements.
This framework uses the integration of different techniques such as static
analysis, process algebras and data-flow networks. In the end, the results
of quantitative and functional analysis make it clear to architects on what
issues need to be addressed in the making of enterprise architectures.
Page 64 of 66
References
Artisan Software Tools: UK's Defence Science and Technology Laboratory
chooses Artisan Studio. (12 January). M2 Presswire. Retrieved March 4,
2010, from ProQuest Computing. (Document ID: 1937530421).
Bailey, Ian. (12 April 2006). Ministry of Defence Classification and Reference
Data Support for MODAF.
Bailey, Ian. (12 November 2007). Representing Information Exchange
Requirements using the MOD Architecture Framework (MODAF).
BMT Hi-Q Sigma. (2007). Project Case Study- Attack Helicopter Mid Life
Update URD. Bath, BA2 3DZ.
BMT Hi-Q Sigma. (2010). About Us. Retrieved March 1, 2010, from BMT HiQ Sigma: http://www.hiqsigma.com/?/1791
Bucher, T., Fischer, R., Kurpjuweit, S., & Winter, R. Enterprise Architecture
Analysis and Application An Exploratory Study. St.Gallen: Institute of
Information Management, University of St. Gallen.
Davis, G., & Washington, K. (2009, Winter). Developing a System Using the
MOD Architecture Framework -a Case Study. Defence Codex(5), 78-83.
Design XML schemas using UML. (2003, February 1). Retrieved February 21,
2010, from IBM:
http://www.ibm.com/developerworks/xml/library/x-umlschem/index.html
Gorman. P. <Patrick.Gorman429@mod.uk> (2010, February 06). About
Modaf [tjakova@quinnipiac.edu]. (2010, February 11).
Hause, Matthew. (October 2008). An Overview of UPDM- A Unified Profile for
DoDaf/MODAF. http://www.mil-embedded.com/articles/id/?3653
IRC Software Glossary. (2009). Retrieved February 15, 2010, from Natural
Resource Ecology Laboratory:
http://www.nrel.colostate.edu/projects/irc/public/Documents/Software/Glos
sary.htm
Jones, J., Fatma, D. F., Blevins, T., & Siegers, R. (2007, May). The Synergy
of Architecture Frameworks: What Goes Around Comes Around. Retrieved
March 3, 2010, from Information Management:
Page 65 of 66
http://www.information-management.com/infodirect/20070518/10826021.html
Lankhorst, M. (2005). Enterprise Architecture at Work - Modelling,
Communication, and Analysis. Enschede: Springer.
Lockheed Martin UK. (2010). Architecture Framework Modelling. Retrieved
March 1, 2010, from Lockheed Martin UK: http://www.lmisgs.co.uk/defence/architecture_framework_modelling.htm
Ministry of Defence. (4 September 2004). MOD Architectural Framework
(MODAF) Tri-Fold
Ministry of Defence. (31 August 2005). MOD Architecture Framework
Version 1.0.
Ministry of Defence. (31 August 2005). Project Management Reference
Guide.
Ministry of Defence. (19 November 2008). The MODAF Technical Standards
Views (TV) Viewpoint.
Ministry of Defence. (20 November 2008). The MODAF Technical Standards
Views (TV) Viewpoint.
Ministry of Defence. (21 November 2008). The MODAF Acquisition (AcV)
Viewpoint.
Ministry of Defence. (24 November 2008). The MODAF Strategic Views
(StV) Viewpoint.
Ministry of Defence. (27 November 2008). The MODAF Service Oriented
View (SOV) Viewpoint.
Ministry of Defence. (3 February 2009). Ontologies and Their Uses MODAF
1.0.
Ministry of Defence. (12 February 2009). The MODAF Operational Viewpoint
(OV).
Ministry of Defence. (17 February 2009). The MODAF System Viewpoint
(SV).
Ministry of Defence. (10 March 2009). MODAF Meta Model V 1.0.
Page 66 of 66