Supporting Organizational Qualities Through Architectural Patterns

Fritz Solms

Department of Computer Science, University of Pretoria, Pretoria, South Africa

Keywords:

Organizational Architecture, Architectural Patterns, Blackboard, Microkernel, Pipes and Filters, Organiza-

tional Quality.

Abstract:

The structure of an organization is known to affect an organization’s qualities. Commonly variations of the

hierarchical patterns are used to constrain the structure of organizations. Examples include functional and

divisional hierarchies, and organizations based on a matrix structure. In either case the depth of the hierarchies

can be varied. For software architectures a much wider set of architectural patterns is considered and patterns

are combined in such a way as to support desired system qualities. This paper suggests that many of these

patterns are applicable also for organizational architectures and proposes a conceptual framework for the

selection and application of architectural patterns in order to improve the quality attributes of an organization.

1 INTRODUCTION

A quality attribute of a system is the realization of

a non-functional requirement (NFR). Examples of

quality attributes include scalability, reliability, ﬂex-

ibility, integrability, performance, and cost. In soft-

ware architecture design it is common to select and

combine architectural patterns to specify an infras-

tructure within which it is easier to satisfy required

quality attributes (Bachmann et al., 2005; Harrison

and Avgeriou, 2007).

Functional and divisional hierarchies as well as

matrix structures are commonly used in organiza-

tional architecture design (Tran and Tian, 2013; Har-

ris and Raviv, 2002). These are based on the hierar-

chical pattern. Organizational architecture optimiza-

tion involves activities like optimizing, grouping into

clusters, specifying depth (tallness or ﬂatness of the

hierarchy), choosing and specifying the level of de-

centralization and standardization (Harris and Raviv,

2002).

This paper suggests that many of the architec-

tural patterns used in software architecture design are

equally applicable to organizational architecture de-

sign. In particular, the Microkernel pattern can facil-

itate service provider ﬂexibility and integrability, the

Blackboard pattern innovatability and the Pipes and

Filters pattern process ﬂexibility and low cost. It is

suggested that architectural patterns should be com-

bined to support required qualities across an organi-

zation.

2 RELATED WORK

Weber’s early work on organizational theory and the

bureaucratic model (Weber, 1958) included a study

of the impact of responsibility and accountability

localization within a hierarchical structure. Both,

functional and divisional hierarchies were considered.

The former are based on a segregation of activities re-

sulting in good skills separation. However, changes to

product scope or business processes require changes

across the organization which are not well managed

within a functional hierarchy. Divisional hierarchies

base the segregation on some other variable like prod-

uct or geography. Whilst changes to product scope

and business processes are better managed within di-

visional hierarchies, they commonly lead to duplica-

tion of resources and poor product integration.

Matrix structures (Mee, 1964; Galbraith, 1971)

were introduced to realize management and coordina-

tion across functional and divisional domains with in-

dividuals reporting to multiple managers. Beneﬁts in-

clude more efﬁcient resource utilization and wider in-

formation ﬂow resulting in increased innovation and

knowledge distribution. However, multiple reporting

lines may introduce conﬂicts and unhealthy competi-

tion for resources.

We have seen a trend from narrow and tall orga-

nizations toward ﬂat organizations with wider spans

of control (Docherty et al., 2001), i.e. to hierarchies

with less layers. Beneﬁts are reduced specialization,

loosened hierarchies of control, rapid and less centra

lized decision making, closer contact between man-

594

Solms, F.

Supporting Organizational Qualities Through Architectural Patterns.

In Proceedings of the 18th International Conference on Enterprise Information Systems (ICEIS 2016) - Volume 2, pages 594-599

ISBN: 978-989-758-187-8

agement and workers enabling the latter to gain a bet-

ter understanding of the strategic goals of the organi-

zation, improved innovation and knowledge distribu-

tion and lower cost. Disadvantages of ﬂat organiza-

tions include possible quality degradation in the con-

text of reduced control, difﬁculties around scaling to

large organizations and handling complex processes.

Also, ﬂat organizations put strain on management and

lack of precise deﬁnition of employee roles may lead

to uncertainty. (Dalton et al., 1980) have shown that

hierarchical depth, degree of specialization, span of

control, the level of centralization, administrative in-

tensity and the level of formalization and standard-

ization impact not only on frequency and duration of

labour strikes and the level of absenteeism, but also

on the turnover of the organization.

This leads one to model the design of an organi-

zational architecture more formally as an optimiza-

tion problem (Harris and Raviv, 2002; Tran and Tian,

2013). (Harris and Raviv, 2002) modeled the im-

pact of changes to an organization’s structure on its

basic activities of producing, designing and market-

ing products, its coordination activities and manage-

rial costs. They found the optimal structure follows

a life cycle as the company grows in size and com-

plexity. In particular, a company tends to evolve from

ﬂat structure to a highly centralized structure to a divi-

sional hierarchy to a functional hierarchy and ﬁnally

to either a matrix structure or a highly decentralized

ﬂat structure. They also found that ﬁrms that do not

face high resource constraints, highly regulated ﬁrms

and ﬁrms operating in stable environment tend to have

decentralized organizational structures.

Moving beyond hierarchical structures,

(Mintzberg, 1979) introduced the Structure of 5.

Here an operational core is managed by the middle-

line (middle management) reporting to the strategic

APEX of the organization with support from tech-

nostructure and support staff. Technostructure covers

areas which are not involved in operations but may

observe, design and optimize them, e.g. research and

development, business analysis, operational research

and strategic planning.

(Kolp et al., 2004) study the ease with which

NFRs can be addressed within different organiza-

tional styles. They include Structure of 5, Pyramid,

Matrix, Chain of Value and the Bidding Style. (G´omez

and Ortiz, 2013) extend the analysis of Kolp and Tung

Do to include further qualities including security and

availability.

(Berzisa, 2015) proposed Capability Driven De-

velopment (CDD) for integrating organizational and

information system development. They suggest

reusing organizational patterns, but these refer to op-

erational patterns and not structural patterns.

3 STRUCTURAL PATTERNS

In this section the hierarchical pattern widely used

in organizational architecture design is compared to

structural patterns used in software architecture de-

sign but currently not explicitly in organizational

architecture design. It is shown how patterns can be

combined to enhance different qualities in different

parts of an organization.

3.1 The Hierarchical Pattern

Functional and divisional hierarchies, ﬂat structures

(single level hierarchies) and matrix structures are all

based on the hierarchical pattern

Employee

ReportsTo

-superior

0..1

Manages

-subordinates

0..*

Figure 1: A UML class diagram representing the structure

of the hierarchical pattern. Employees have a communi-

cation channel to their superior as well as to their subor-

dinates. Exceptions are the top (CEO) and bottom of the

reporting hierarchy.

Figure 1 shows the reporting lines speciﬁed by the

simple hierarchical pattern. The pattern introduces

specialized roles requiring a person to manage a par-

ticular domain of responsibility at a particular level of

accountability as is more explicitly depicted in Figure

2. The natural human resourcing approach is thus job

based human resourcing.

domain of responsibility

level of

accountability

Figure 2: A UML object diagram depicting an instance

structure. It shows the domain of responsibility and level

of accountability assigned to employee X.

Accountability hierarchies improve reliability and

role based human resourcing around deﬁned domains

of responsibility at deﬁned levels of accountability re-

sult in standard skills requirements which simpliﬁes

scalability.

On the other hand, the reduced communication

channels and ﬁxed ring-fenced domains of responsi-

bility make it more difﬁcult to achieve high levels of

innovatability and learnability. Changing processes

within hierarchical organizations is often more difﬁ-

cult as processes typically cut across domains of re-

Supporting Organizational Qualities Through Architectural Patterns

595

sponsibility and business units. Finally, deep hierar-

chies incur a signiﬁcant cost-overhead.

The hierarchical pattern is particularly applicable

in areas where control and accountability hierarchies

are desirable. It’s use should be minimized in areas of

the organization where innovatability and ﬂexibility

are important.

3.2 The Blackboard Pattern

The blackboard architectural pattern (Buschmann

et al., 1996) has been applied to a wide range of

problems including speech recognition (Lesser et al.,

1975), distributed resource allocation planning (Han

et al., 2014), and decision support for stock trading

(Luo et al., 2002). It provides an infrastructure for

specialized processing units (experts) to collaborate

in developing a possibly partial or approximate solu-

tion to a difﬁcult problem for which no deterministic

solution strategies are known.

The pattern obtained its name from an analog

of mathematicians collaboratively solving a difﬁcult

problem using a blackboard as communication mech-

anism. An observable blackboard hosts both the prob-

lem speciﬁcation as well as the current state of the so-

lution. One mathematician may try one approach to a

solution and get stuck leading to another expert per-

haps continuing on the approach or alternativelyargu-

ing why the approach will not work and removing the

contributions made by the ﬁrst mathematician from

the blackboard. Note that the experts auto-orchestrate

a process amongst themselves.

:Blackboard

(Knowledge Repository)

Expert Pool

Observer

Observable

:Controller

:Expert

Observable space

Figure 3: A UML object diagram depicting the structure of

the blackboard pattern.

In the blackboard pattern (see Figure 3) a con-

troller manages a pool of experts and feeds problems

into a blackboard which is observed by them. Oper-

ating as a self-organizing team, experts decide when

to process information or requests published on the

blackboard. The output of any processing is fed back

into the blackboard and may trigger further activities

by experts. The controller manages the scope of the

problem and decides when a solution or a sufﬁcient

approximation to it has been obtained. The natural

human resourcing for blackboard based architectures

is skills based human resourcing (Lawler and Led-

ford, 1992) and not role-based human resourcing.

The strengths of the blackboard pattern include

innovatability through auto-orchestrated processes as

well as learnability and knowledge distribution (Ju-

rado et al., 2012). The blackboard provides an in-

tegration infrastructure improving integrability (Ju-

rado et al., 2012) and auditability. Scalability can

be achieved by adding further experts for concur-

rent processing within a processing grid (Dong et al.,

2005). One can dynamically add and remove ex-

perts resulting new processes being immediately auto-

orchestrated across the new skills set. This increases

ﬂexibility. Security is enforced through access con-

trol on the blackboard (Ortega-Arjona and Fernan-

dez, 2008). Weaknesses include reduced predictabil-

ity and reliability as the success of auto-orchestrated

is not guaranteed. Communication overheads and

auto-orchestration may lead to performance concerns

(Schwartz, 1995). Due to redundancy and non-

stream-lined processes, the blackboard pattern is not

usually suited for organizations competing on cost.

Even though the blackboard pattern is currently

not used explicitly in organizational architecture de-

sign, one ﬁnds organizationalstructures which are im-

plicitly based on this pattern. In open source soft-

ware development organizations the version control

repository represents the blackboard, developers the

experts, and the project lead the controller. Other ex-

amples include Wikipedia, the bidding style discussed

by (Kolp et al., 2004) and even a board of directors.

One should consider the blackboard pattern when in-

novation,learnability,extreme ﬂexibility are required,

e.g. for R&D, marketing and software development

teams.

3.3 The Microkernel Pattern

The microkernel pattern was introduced for operating

systems to separate a minimal set of core functional-

ity for which ﬂexibility is sacriﬁced for efﬁciency and

reliability from pluggable extension code which re-

quires a higher level of ﬂexibility (Liedtke, 1995). It

has also been used for application servers (Goebel and

Nestler, 2004) and is the base pattern for Services Ori-

ented Architectures (SOA) with the Enterprise Ser-

vices Bus (ESB) representing the microkernel (Solms,

2012).

Figure 4 depicts the structure of the microkernel

pattern. The microkernel and internal servers are ef-

ﬁcient, reliable implementations which evolve only

slowly whilst external servers evolve more rapidly

to address evolving client requirements. The micro-

kernel provides a communication infrastructure de-

coupling the various servers. Clients access external

servers through an adapter.

ICEIS 2016 - 18th International Conference on Enterprise Information Systems

596

:Microkernel

:External

Server

:Adapter

:External

Server

:Adapter :Adapter

:Adapter:Adapter :Adapter

:Internal

Server

:Internal

Server

:Client:Client

Kernel

Space

User

Space

Figure 4: A UML object diagram depicting the structure of

the microkernel pattern.

The main purpose of using a microkernel pattern

is usually integrability (Solms, 2012) which, in turn,

improves reusablity and ﬂexibilitythrough decoupling

and pluggability. Auditability is achieved by logging

all messages traversing the microkernel.

On the other hand, decoupling via a microkernel

introduces performance and cost overheads (Solms,

2015; Sun et al., 2013) and the isolation of the vari-

ous responsibility domains (servers) reduces opportu-

nities for innovation and learning.

An example for implicit use of the microkernel

pattern with organizational architecture design is that

of an internal post-ofﬁce routing messages between

internal departments as well as to and from exter-

nal organizations to the appropriate internal business

units represents a microkernel. Also, corporate and

merchant banks can be seen to be implicitly based on

the microkernel pattern. Client faced business units

like credit derivatives and foreign exchange operate

like independent organizations. These are the ex-

ternal servers which need to be ﬂexible and innova-

tive, evolving their service offering to address chang-

ing client needs and environmental changes. Back-

ofﬁce services like transaction processing and reg-

ulatory reporting which need to provide their ser-

vices reliably and efﬁciently are the internal servers

which change infrequently. The communication be-

tween client faced external servers and the internal

servers providing back-ofﬁce services is commonly

facilitated through an Enterprise Services Bus (ESB).

Finally, clients are provided uniﬁed access channels

across external servers in the form of a web interface

and call center for human access and web-services

and messaging adapters for system access. This uni-

ﬁed interface represents the adapter of the microker-

nel pattern.

3.4 The Pipes and Filters Pattern

The pipes and ﬁlters pattern was originally used for

compiler design (Meunier, 1995). The pattern is

widely used for processing pipelines like signal or

video processing, Unix command pipelines and for

workﬂow systems (Scheibler et al., 2010).

Filters represent processing units which process

Pipe

Filter

Source

Sink

-outputPipe

-inputPipe

Figure 5: A UML class diagram depicting the structure of

the pipes and ﬁlters pattern.

inputs received from an input pipe and deliver out-

puts to an output pipe (see Figure 5). Pipes have the

responsibility of transporting artifacts between ﬁlters.

Processing units are decoupled through pipes and de-

ﬁned pipes and ﬁlters processes can be aggregated

into higher-level, reusable ﬁlters.

The main strength of the pipes and ﬁlters pattern

are process ﬂexibility including the ability to quickly

assemble new processes from the available processing

components (ﬁlters) (Scheibler et al., 2010). Weak-

nesses include low innovatability and learnability due

to the isolation of processing units and the lack of

overall control. The pattern does, however, incur

only low cost overheads as the structure is focused

on value-generating processing components.

The pipes and ﬁlters pattern is the base pattern for

the Chain of Value style discussed by (Kolp et al.,

2004). It is also commonly used to achieve ﬂexibility

in manufacturing organizations. The ﬁlters are ma-

chines performing processing steps on the product be-

ing built and pipes (e.g. conveyor belts) transport the

product between processing steps.

4 LEVELS OF GRANULARITY

When designing an infrastructure for an organization,

one is unlikely to use a single pattern. Instead one

may constrain the high-level structure for the system

as a whole (the ﬁrst level of granularity) with a partic-

ular pattern, whilst constraining the infrastructure of

lower level organizational components through other

patterns.

The infrastructure design of an organization can

be driven by a recursive decomposition into business

units requiring qualities and basing the selection of

the pattern constraining its infrastructure on the ease

with which the required qualities are realized within

the structure prescribed by the pattern as discussed

in section 3 and summarized in Table 1. Generally

one will want to select a pattern which supports the

most important quality attribute(s) for the respective

business unit whilst not negativelyimpacting on other

relevant qualities.

For example, an infrastructure based on the hier-

archical pattern could be chosen if reliability/control

are important and the microkernel or blackboard for

integrability and pipes and ﬁlters for process ﬂexibil-

Supporting Organizational Qualities Through Architectural Patterns

597

Table 1: Impact of architectural patterns on organi-

zational qualities. An up/down-arrow indicates that

the pattern makes it easier/more difﬁcult to realize

a quality attribute, whilst the tilde symbol indicates

that it has no signiﬁcant effect on a quality attribute.

Pattern

Quality Attribute

Quality

Innovatability

Learnability

Flexibility

Reliability

Integrability

Reusability

Scalability

Performance

Auditability

Microkernel

∼ ∼

↑

∼

↑ ↑

∼

↓ ↑

Blackboard

↑ ↑ ↑ ↓ ↑ ↑ ↑ ↓ ↑

Hierarchcal

↓ ↓ ↓ ↑ ↓ ↓ ↑

∼ ∼

Pipes & Filt

↓ ↓ ↑

∼ ∼

↑

∼ ∼ ∼

ity and low cost. The blackboard pattern is the only

pattern which provides infrastructural support for in-

novation.

As an illustrative example consider designing the

infrastructure for a hypothetical corporate and mer-

chant bank. Assume the organization has a large num-

ber of specialized trading units operating as virtually

independent business units with their own accounting

base (e.g. credit derivatives, foreign exchange and eq-

uities) which compete in the market by constructing

innovative hedging products for their clients. These

trading units consume back-end services like transac-

tion processing, regulatoryreporting and procurement

from back-ofﬁce units. These back-ofﬁce units need

to not only provide their services efﬁciently, but must

also be able to rapidly and at low cost change their

business processes to adapt to continuously changing

laws an regulations. The organization as a whole is

to be controlled by a relatively shallow management

structure.

The most important quality requirement for the

small trading business units is innovatability. Hence

one may consider either basing their infrastructure on

the blackboard architectural pattern or use that pattern

within those business units as incubator for innova-

tion.

The back-ofﬁce services could provide their

streamlined services within a pipes and ﬁlters infras-

tructure which allows for rapid and efﬁcient process

changes driven by regulatory requirements. Select-

ing the microkernel pattern as integration infrastruc-

ture between the loosely coupled client facing and

back-ofﬁce business units enhances ﬂexibility and the

ability to easily add further business units. The back-

ofﬁce units represent the internal servers of the micro-

kernel pattern, the trading units the external servers

and a uniﬁed client interface the adapter. The mi-

crokernel is responsible for routing requests from the

client adapter to the appropriate trading units and be-

tween the trading and back-ofﬁce units. The concrete

manifestation might be an Enterprise Services Bus

(ESB) within a service-oriented software architecture.

For the control infrastructure for each of the busi-

ness units one could consider a single layer hierarchy

which in turn reports to a relatively ﬂat hierarchy for

overall control and accountability of the organization.

The above combination of pattern is chosen to facil-

itate innovation, ﬂexibility, integrability and account-

ability and control where these respective qualities are

important.

5 CONCLUSIONS AND FUTURE

WORK

Organizational architecture design has predominantly

focused on structures based on the hierarchical pat-

tern including functional and divisional hierarchies,

matrix structures and ﬂat hierarchies. Hierarchical

structures are strong in control and accountability,

but achieving qualities like innovatability, ﬂexibil-

ity, learning, and knowledge distribution is difﬁcult

within such structures.

This paper argues that one should also consider

other architectural patterns like blackboard, hierarchi-

cal and pipes and ﬁlters to achieve qualities like ﬂex-

ibility, innovatability, and integrability. One should

select and combine patterns in such a way that they

facilitate the required qualities across organizational

components. Furthermore, viewing enterprise archi-

tecture as an aggregation of organizational and soft-

ware architecture makes it attractive to use the same

concepts across both these sub-domains. Such an ap-

proach may make it easier to achieve alignment of

software architecture and organizational architecture

– it is across both of these that an organization needs

to realize its desired quality attributes.

Future work will include include empirical stud-

ies investigating the correlation between the use of ar-

chitectural patterns and organizational qualities. The

work of Harris and Raviv (Harris and Raviv, 2002)

can be extended to provide a formal optimization

framework for selecting and combining architectural

patterns.

Whist the selection of architectural patterns pro-

vides an infrastructure within which it is easier to re-

alize certain qualities, it is the selection of architec-

tural tactics like load balancing, resource reuse, ac-

tive and passive redundancy which concretely realize

these qualities (Solms, 2012). Future work will aim

to include the selection of architectural patterns and

tactics within a comprehensive method for organiza-

tional architecture design.

ICEIS 2016 - 18th International Conference on Enterprise Information Systems

598

REFERENCES

Bachmann, F., Bass, L., Klein, M., and Shelton, C. (2005).

Designing Software Architectures to Achieve Quality

Attribute Requirements. IEE Proceedings - Software,

152(4):153–165.

Berzisa, S. e. a. (2015). Capability driven development: An

approach to designing digital enterprises. Business &

Information Systems Engineering, 57(1):15–25.

Buschmann, F., Meunier, R., Rohnert, H., Sommerlad, P.,

and Stal, M. (1996). Pattern-Oriented Software Archi-

tecture, Volume 1: A System of Patterns. Wiley, Chich-

ester, UK.

Dalton, D. R., Todor, W. T., Spendolini, M. J., Fielding,

G. J., and Porter, L. W. (1980). Organizational struc-

ture and performans: A critical review. Academy of

Management Review, 5(1):49–64.

Docherty, J. P., Surles, R. C., and Donovan, C. M. (2001).

Textbook of Administrative Psychiatry, chapter Orga-

nizational Theory, pages 33–43. ASM Zahidul Islam,

second edition edition.

Dong, J., Chen, S., and Jeng, J.-J. (2005). Event-based

blackboard architecture for multi-agent systems. In

Information Technology: Coding and Computing,

2005. ITCC 2005. International Conference on, vol-

ume 2, pages 379–384 Vol. 2.

Galbraith, J. R. (1971). Matrix organization designs: How

to combine functional and project forms. Business

Horizons, 14(1):29 – 40.

Goebel, S. and Nestler, M. (2004). Composite component

support for ejb. In Proceedings of the Winter Interna-

tional Synposium on Information and Communication

Technologies, WISICT ’04, pages 1–6. Trinity Col-

lege Dublin.

G´omez, O. and Ortiz, J. (2013). Modeling the organiza-

tional style structure in ﬁve. Procedia Technology,

7:384 – 390.

Han, X., Mandal, S., Pattipati, K., Kleinman, D., and

Mishra, M. (2014). An optimization-based distributed

planning algorithm: A blackboard-based collaborative

framework. Systems, Man, and Cybernetics: Systems,

IEEE Transactions on, 44(6):673–686.

Harris, M. and Raviv, A. (2002). Organization design. Man-

age. Sci., 48(7):852–865.

Harrison, N. B. and Avgeriou, P. (2007). Leveraging

architecture patterns to satisfy quality attributes. In

Oquendo, F., editor, Software Architecture, First Eu-

ropean Conference, ECSA 2007, Aranjuez, Spain,

September 24-26, 2007, Proceedings, volume 4758 of

Lecture Notes in Computer Science, pages 263–270.

Springer.

Jurado, F., Redondo, M. A., and Ortega, M. (2012). Black-

board architecture to integrate components and agents

in heterogeneous distributed elearning systems: An

application for learning to program. Journal of Sys-

tems and Software, 85(7):1621 – 1636. Software

Ecosystems.

Kolp, M., Do, T. T., and Faulkner, S. (2004). Analysis styles

for requirements engineering: An organizational per-

spective. In Chang, S., editor, Handbook Of Software

Engineering And Knowledge Engineering, volume 3,

chapter 1. World Scientiﬁc.

Lawler, E. and Ledford, G. (1992). A skill-based approach

to human resource management. European Manage-

ment Journal, 10(4):383–391.

Lesser, V., Fennell, R., Erman, L., and Reddy, D. (1975).

Organization of the hearsay ii speech understanding

system. Acoustics, Speech and Signal Processing,

IEEE Transactions on, 23(1):11–24.

Liedtke, J. (1995). On micro-kernel construction. SIGOPS

Oper. Syst. Rev., 29(5):237–250.

Luo, Y., Liu, K., and Davis, D. (2002). A multi-agent deci-

sion support system for stock trading. Network, IEEE,

16(1):20–27.

Mee, J. F. (1964). Matrix organization. Business Horizons,

7(2):70 – 72.

Meunier, R. (1995). The pipes and ﬁlters architecture. In

Pattern Languages of Program Design, pages 427–

440. ACM Press/Addison-Wesley Publishing Co.,

New York, NY, USA.

Mintzberg, H. (1979). The structuring of organizations: a

synthesis of the research. Theory of management pol-

icy series. Prentice-Hall.

Ortega-Arjona, J. L. and Fernandez, E. B. (2008). The se-

cure blackboard pattern. In Proceedings of the 15th

Conference on Pattern Languages of Programs, PLoP

’08, pages 22:1–22:5, New York, NY, USA. ACM.

Scheibler, T., Leymann, F., and Roller, D. (2010). Execut-

ing pipes-and-ﬁlters with workﬂows. In Internet and

Web Applications and Services (ICIW), 2010 Fifth In-

ternational Conference on, pages 143–148.

Schwartz, D. (1995). Blackboard systems. In Cooperating

Heterogeneous Systems, volume 299 of The Springer

International Series in Engineering and Computer

Science, pages 27–42. Springer US.

Solms, F. (2012). What is software architecture? In Pro-

ceedings of the South African Institute for Computer

Scientists and Information Technologists Conference,

SAICSIT ’12, pages 363–373, New York, NY, USA.

ACM.

Solms, F. (2015). A Systematic Method for Architecture

Recovery. In 10’th International Conference on Eval-

uation of Novel Approaches to Software Engineering,

pages 215–222, Barcelona, Spain.

Sun, H., Wang, X., Yan, M., Tang, Y., and Liu, X. (2013).

Towards a scalable paas for service oriented software.

In Parallel and Distributed Systems (ICPADS), 2013

International Conference on, pages 522–527.

Tran, Q. and Tian, Y. (2013). Organizational structure:

Inﬂuencing factors and impact on a ﬁrm. Ameri-

can Journal of Industrial and Business Management,

(3):229–236.

Weber, M. (1958). From Max Weber: Essays in Sociology.

Oxford University Press, USA.

Supporting Organizational Qualities Through Architectural Patterns

599