System-Of-Systems to the Rescue? Solving Unsolvable Problems
The term system-of-systems is increasingly used to describe systems operating under conditions of ambiguity, complexity, emergence, interdependence, and uncertainty. Although there is a good understanding of the kinds of systems that could be considered as systems-of-systems, a consensus on an exact definition of the term has yet to emerge --- bringing into question the nature of solutions to the problem space of system-of-systems. Terms, including large-scale systems, socio-technical systems, and cyber-physical systems are often used. Convergence in different nomenclature is: system-of-systems exhibit several specific characteristics, to various degrees. These characteristics are the subject of this article along with challenges associated with using system-of-systems as an approach to address modern worrisome issues.
There exists a plethora of definitions for system-of-systems (Keating and Katina 2011). In literature, term system-of-systems (SOS) is at times taken as an approach to address technical as well as non-technical issues ranging from ‘global earth observation systems’ (Fritz et al., 2008) to space exploration (DeLaurentis et al., 2006), and anything in-between, see for example, Gheorghe and Vamanu (2008) for an approach of data mining to address risk and vulnerability assessment at the world scale.
Despite a wide range of usability, the point at which SOS, as an approach, is necessary, is not always well articulated. This is so despite the promise of being primed as the viable approach to possibly resolving current problematic situations (Barot et al., 2012; Keating and Katina, 2012). In the meantime, the international community continues to look for solutions to the 21st century’s most vexing issues in energy and food security, healthcare costs, mass transportation risks, and terrorism, and corruption. Although, SOS and especially the engineering of SOS (SOSE), emerged as a possible approach to dealing with ambiguity, complexity, emergence, and uncertainty with considerable exploration of technical as well as human/social, organizational/managerial, and policy/political elements and their interactions (Barot et al., 2012; Keating and Katina, 2011), these issues are prevalent and some might even argue, increasing at an alarming rate.
SoS to the rescue?
Essentially, SOS/E has emerged as a discipline capable of holistically and systemically analysing complex systems of systems problems from different viewpoint. This viewpoint includes a combination and consideration of technology, policy, and economics. This combination might be necessary given the characteristics of SOS problem domains. Additionally, this problem domain offers insights into the need for SOS methodologies and technologies. These technologies might include robust methods, processes, tools and techniques supporting engineering of SOS.
Typically, these characteristics revolve around:
- Proliferation of intensive information technologies
- Divergent and often incompatible and politically driven worldviews
- Scarce and dynamically shifting resources that create uncertainty and instability
- Constant shifting conditions and emergent understanding of problem systems
- Rapid technological changes that outpace capabilities to support development, integration, maintenance, and evolution of systems
- Increasing demand for responsive actions and solutions to alleviate mission shortfalls
- Abdication of long-term thinking in response to immediate perceived needs
- Increasing complexities and uncertainties questioning traditional systematic approaches to deal with complex systems and their problems
- Increasing complexity as a result of interplay between technology, policy, and economic strategies
- Globalization and trans-boundary dependencies and interdependencies among complex systems with large implications on human life in the 21st century
This listing is representative of the operating environment in which critical systems whose continued and proper operability is vital for public well-being. Presently, critical systems include, but not limited to agriculture and food, water, healthcare, information and telecommunications, energy, transportation, banking and finance, and manufacturing (Kröger and Zio, 2011). However, and in current times, there is hardly any system that operates out of bounds of the suggested conditions.
Moreover, there is increasing recognition that such systems, including your business and your workplace, do not operate in isolation. This condition suggests that even if your system is safe and secure, it could be compromised through its interdependencies (Katina et al. 2014). Incredibly, and despite the importance of the topic of system-of-systems, there remains challenge of simply defining the term: system of systems. In a report to the US Secretary of Defense, the Stevens Institute of Technology (2006) noted that there were over 40 varying definitions. One could debate the definitions; we have left that for the intellectuals. In the present case, we offer a definition of Keating and his colleagues at Old Dominion University’s National Centers for System of Systems Engineering (2003, p. 40):
“The design, deployment, operation, and transformation of metasystems that must function as an integrated complex system to produce desirable results. These metasystems are themselves comprised of multiple autonomous embedded complex systems that can be diverse in technology, context, operation, geography, and conceptual frame.”
This definition should suffice in the present discourse. Keating’s research suggests looking at SOS as complex autonomous systems that must be integrated to fulfil a mission that cannot be achieved by an individual system. Those interested in this topic would do well to know institutions offering resources related to system of system problem landscape. These entities are listed in Table 1.
Dominant themes of Systems of Systems
How do you know that it’s time to apply SOS concepts for a problem domain? Characterizing a problematic domain may be a good starting point. Maier (1996) posits looking at five things: 1) operational independence of complex systems, 2) managerial independence, 3) evolutionary development, 4) emergent behaviour, and 5) geographical distribution as enabling characteristics of SOS approach. Table 2 provides a summary of these characteristics along with an elaboration (Keating and Katina, 2011). Consistent with the US Department of Defense (2008), Jamshidi (2009) and Keating et al. (2003), suggest a need to identify situations in which SOS approach is necessary. This identification, is an essential step toward in enabling the proper engineering efforts to take place, especially in today’s resource constrained society.
Along these thought, the following is an articulation of dominant themes that could serve as triggering points suggesting SOS approach might be necessary to enable design, (re)design, and strategic system development to produce desirable results:
Ambiguous problem boundary
In organizations, knowing the boundary enables management and analysts to know the limits of responsibilities. In technical and ‘simple’ systems, it might be relatively easy to identify system boundaries. However, when an organization operates in a SOS landscape, the boundary of the system is fluid and negotiable. In fact, stakeholders might have competing and diverging perspectives on scope of operations and other elements of system-of-systems, including system boundary. Therefore, a shifting and ambiguous formulation of system boundary could indicate a need for a system-of-systems paradigm.
Emergent system behaviours
Emergence is a principle in classical systems theory that generally suggests that system properties (behaviours) develop from an interaction of system elements. Unexpected behaviour is considered an obstacle for traditional systems engineering. However, SOS/E appear to embrace emergence and in fact it is concerned with the integration and interaction of complex systems to create new behaviours and structures. For SOS, holistic approach is order of the day and encompasses sociotechnical, managerial, political, and policy dimensions. There is no telling what would happen once all these elements interact --- then again, once you find yourself in this situation, this should serve as an indicator for a need for system-of-systems approach.
Integrated multiple systems
One of key differences between systems engineering and system of systems engineering has to do with type of systems in a system of systems. First, SOSE is suited for complex systems composed of many entities (these entities are complex in their own right), a high level of interactivity that is further characterized by dynamic nonlinear relationship, fluctuating structure and complex relations between parts, as well as open to influences from the environment. In the SOS paradigm, a metasystem works with autonomous embedded complex systems as a one integrated system.
The concept of complementarity suggests that multiple different perspectives of a system will reveal truths regarding the system that are neither entirely independent nor entirely compatible. Given the nature of SOS, constituent of multiple systems with different management teams, it is conceivable to see how one’s perspectives can be considered correct, from a particular system vantage point, and yet incorrect, from alternative system vantage point. In SOS paradigm, there is a need to consider implications for possible diverging perspectives and/or conflicting ideologies regarding, for instance, the nature of situations and possible solutions.
Taking context to be factors, conditions and circumstances constraining and/or enabling system solutions, one can see the utility of context in traditional systems engineering as well as system-of-systems engineering. In systems engineering, context enables the scoping of a system by providing actors who can influence the system even though they are outside of the system. Given the nature of system of systems, it is expected that context will vary from system to system (consider managerial independence). In fact, context can have more implications for the solution space and may even become more critical than the technical aspects of a solution. A proper consideration of context in SOS operating conditions can offer insights into transforming independent systems into a coherent whole system-of-systems.
The environment in SOS is considered highly dynamic, uncertain, and changeable over time. These factors make it difficult to have a complete knowledge and understanding performance and other issues in system-of-systems landscape. It is no wonder it has been suggested that SOS environment consists of a wide variety of variables that can affect the state of the system of interest (Adams and Keating, 2009). This would then suggest that understanding the different variables in SOS environment becomes an essential element of synthesis and analysis.
The selected themes, are simply presented here as a means that should alert system owners, operators, and analysts of the need for SOS paradigm, approach, and methodologies. It this list complete? The list is by no means an exhaustive list of relevant indicators for SOS approach; should these themes arise and we still insist on using traditional approaches, then we should not be surprised when the proposed solutions do not work, or even worsen the situation.
Emerging dominant SoS perspectives
There appears to be three main categorizations of SOS perspectives: government, enterprise, and academia (Keating and Katina, 2011):
SOS Applications: A government perspective
In this perspective, a dominant theme is bringing together different technologies, people, and organizations to effectively deal with emerging issues such as terrorism and cyber-threats and the likes. Of course, this perspective is within the context of military operations. In this perspective, four themes seem to take precedence:
Technology emphasis – the use of technology as the overarching aspect for problem domain for SOS developers. SOS role in military operations is seen as supporting command and control through the integrated use of technology to achieve SOS missions.
Interoperability emphasis – government agencies tend to recognize that providing for public safety involves multiple people, technologies, and organizations. To fulfil public safety needs, government is concerned with ensuring that ‘all’ technical (sub)systems in a SOS are interoperable. This calls for effective data and information exchange across multiple ‘independent’ subsystems to provide integration for command and control.
An extension of systems engineering – there is a mature linkage between military functions and systems engineering. This has creating a situation in which SOS is seen as a subset and an extension of systems engineering, where the main issues include: consistent technical top-down, iterative process of requirements engineering analysis, functional analysis and allocation, design synthesis and verification, systems analysis and control, and the ‘architecture’ paradigm.
Systems acquisitions – in systems engineering, interoperability of systems is necessary especially given the development process of complex technical systems. These are often acquired through the acquisition process. Since there is already a strong link between the field of acquisition and government systems, SOS is seen as just another approach to system acquisition.
SOS Applications: An enterprise perspective
The international community as well as local communities recognize the need for consideration of the technical as well as non-technical aspects of a complex situations. Recognizing that problems do not exist in isolation, the Trans-Atlantic Research and Education Agenda in Systems of Systems (T-AREA-SOS) report recommends the use of SOSE in dealing with emerging interdepend societal challenges on a large scale (Barot et al., 2012). This is not to suggest that SOSE is an extension of systems engineering, since the enterprise perspective tends to drop the term ‘engineering’ from SOS (Keating and Katina, 2011). Such entities tend to place emphasis on seeing holistically --- incorporating local and international scale especially as an enterprise has systems operating as interdependent systems. Central tenets for the enterprise SOS perspective involve:
Expanding beyond technical considerations – there is greater emphasis on expanding SOSE beyond the ‘technology-only’ focus to include a diverse element of strategies, social, policies, and pathologies in system components, all of which might contribute to understanding and the solution-space.
De-emphasis of engineering – it is not by accident that the enterprise perspective appears to de-emphasize engineering in SOSE. SOS presents a broader reach and also permits inclusion of non-engineering applications such as enterprises and the different types of problems beyond technical type of problems along with solutions commonly associated with engineering.
Dominance of architecture – the notion of architecture at the SOS level is central to the enterprise perspective. The notion of system architecture is a dominant theme in systems engineering and has certainly found roots in emerging research in SOSE (Valerdi et al., 2007), particularly in the enterprise architecture paradigm.
SOS Applications: An academia perspective
Academia offers several articulations beyond technical - traditional engineering, and architecture (Keating and Katina, 2011). The emphasis is on developing philosophical, epistemological, and ontological underpinnings for applications of SOS including transportable methods, tools, and techniques. Central tenets for the academic SOS perspective include:
Distinguishing SOS from other disciplines – in academia, there are on-going efforts to develop groundings for SOSE involving theoretical and conceptual differentiations. In a typical academic fashion, there exists confusion as well as calculated efforts to distinguish SOS/E from other fields. This could have impact on, for instance, how one might approach risk and vulnerability in system-of-systems.
Phenomena exploration is central - in contrast to government and enterprise SOS/E perspectives, which show little patience for academically grounding of SOSE practice and leveraging enterprise interests, academia promotes theoretical work with an emphasis on investigation, inquiry, and seeking to understand the SOS phenomena, before driving to ‘solutions-first’ mode.
Grounding in systems theory - the academic perspective of SOS/E appears to appreciate grounding itself in systems theory. There is an emphasis on accepting and promoting laws, principles, and theorems of systems science as a basis for a deeper understanding phenomena and development of the field of SOS/E
There are clear differences in government and military, enterprise, and academia perspectives. However, they all recognize the value that SOS and SOSE can offer in dealing with emerging issues in society.
SoS at work: connected health systems
In an example of application of the concept of SOS, let us consider a connected healthcare system. A connected healthcare system can encompass a wide variety of technologies that can improve delivery of healthcare in terms of both, quality and cost. An example of this technology is patient portal technology that allow for integrated interaction and communication between a patient and their healthcare providers including hospitals, individual doctors, nurses, insurers, and offer access to government programs.
The philosophy behind such an approach is that the collaborative efforts of such many systems enables better patient care. Systems involved might range from physiological monitoring systems (e.g., Body Sensor Networks), general environment sensing systems (e.g., Wireless Sensor Networks), aiding systems (e.g., robotic surgery systems), networked medical devices in a hospital setting, and linked health informatics (e.g., electronic health record), as well as medical devices in general that are part of the delivered healthcare.
Applying SOS/E in such a situation requires coordination and integration of the multiple systems to create an integrated system and connected healthcare system with a capability beyond any single healthcare systems. Clearly, one of the aims of an SOS approach to healthcare is improving healthcare delivery. This is not new. What is new, however, and supposedly is the efficiency at which healthcare can be delivered. Of course, you have to ask, among the many questions you might have, how is efficiency measured in SOS healthcare, which healthcare systems have taken on this approach, and what are the results?
Unfortunately, we don’t have the answers. However, we suggest that such an approach include aspects of in-home healthcare, where devices control and monitor the dosage of drugs, remotely, check and warn of hazards including accidental falls and appliance shut-off, or establish behavioural patterns and symptoms: well does, your smartwatch do that? Similarly, such a system could include virtual doctor visits to reduce the cost of hospital visit especially to rural areas.
Creating SOS healthcare has its challenges. Technologically, such systems would need to collaborate. In most cases, such systems might have been developed independently by different developers and used by different operators, not to mention possible different operation systems. In a case of SOS healthcare, each connected system would need not only to communicate so as to understand each other, say via a common interface, but also to be able to improve the health of the patient, by offering new and previously unavailable capability to healthcare professional. This suggests a need for safety, security and information governance, and assurance in SOS healthcare system. The last thing you want is for someone to hack your ‘mechanical___’.
Beyond these challenges in SOS healthcare application, there are challenges associated with the domain itself. These include but not limited to:
- Evolving systems – it is understood that SOS constantly evolve and must adapt to their environment; in turn, concept of operations for SOS changes, new elements are added and new capabilities are needed. Well, what does this mean for someone interested in analysis and synthesis? How does one, for example, address hazards, risk and vulnerability?
- Emergent behaviour – it is understood that the overall behaviour of a SOS cannot be described purely in terms of the behaviour of individual systems. Given that there could be competing objectives of systems in a SOS arrangement, it is conservable that systems involved might not benefit equally. Consider, routing an airborne radar over a high-risk area will affect the system --- increase the risk and could affect the safety of the system. Yet, this activity will beneficial other elements of the SOS. For instance, troops and aircrafts receive a more detailed intelligence package. Certainly, this is could be a challenge, since the presented situation does not appear to be a win-win. Then, it is necessary to consider the implications of ‘emergence’ and ‘benefits’ in SOS applications.
- Functional decomposition – it is understood that functions do not reside in one locality but instead are dispersed across various systems. Individual systems can have their own goals, missions, and functions. In a SOS endeavour, the coming together of individual systems creates capabilities to address issues beyond capability of individual systems. This suggests a need to distinguish capabilities required at the system and the system of systems level, the consideration of effects this could have on the individual system capability to meet its ‘regular’ goals, missions, and functions. Even, more critically, the approach and mechanisms that enable the creation of highly performing system of systems.
SOS FUTURE…I have seen the future, but I have a non-disclosure!
You now have some idea as to what SOS entails and there know of the fact that applications are vast in government, enterprises, and academia. A general consensus is that SOS and SOSE can serve as an alternative to the traditional approach to dealing with modern society’s most vexing situations.
To effectively deal with emerging situations, there is a need me ‘things’ to use. Arguably, these ‘things’ include SOS technologies and tools. To this list, one might suggest methodologies, methods, processes, tools and techniques supporting SOS/E. The next logical question is, where are these ‘things’?
At this point, it is safe to say that these ‘things’ are essential in the analysis, designing, and evaluation of systems operating in the present landscape, and if we desire to change how such systems are performing, it is essential to develop such ‘things.’
Albert Einstein is often attributed to having said: we cannot solve our problems with the same thinking we used when we created them. To this we suggest that perhaps, the means being suggested as ways to solve problems, including SOS/E, must have solid conception foundations that is different from the thinking that created the problems in the first place; if this is the case, then it is possible to develop ‘things’ that would support this conceptual foundation and such ‘things’ would indeed be different from those we might be currently using.
Adams, K. M., & Keating, C. B. (2009). SoSE methodology rev 0.2 (No. NCSOSE Technical Report 009-2009). Norfolk, VA: National Centers for System of Systems Engineering.
Barot, V., Henson, S., Henshaw, M., Siemieniuch, C., Sinclair, M., Lim, S.-L., … DeLaurentis, D. (2012). Trans-atlantic research and education agenda in systems of systems (T-AREA-SoS) (State of the Art on Systems of Systems Management and Engineering No. TAREA-PU-WP2-R-LU-9) (p. 141). Leicestershire, UK: Loughborough University. Retrieved from https://www.tareasos.eu/docs/pb/T-AREA-SoS_FINAL_SOAReport_V2.0.pdf
DeLaurentis, D. A., Sindiy, O. V., & Stein, W. B. (2006). Developing sustainable space exploration via a system-of-systems approach. In The American Institute of Aeronautics and Astronautics. San Jose, CA. Retrieved from http://arc.aiaa.org/doi/pdfplus/10.2514/6.2006-7248
Fritz, S., Scholes, R. J., Obersteiner, M., Bouma, J., & Reyers, B. (2008). A conceptual framework for assessing the benefits of a global earth observation system of systems. IEEE Systems Journal, 2(3), 338–348.
Gheorghe, A. V., & Vamanu, D. V. (2008). Mining intelligence data in the benefit of critical infrastructures security: Vulnerability modelling, simulation and assessment, system of systems engineering. International Journal of System of Systems Engineering, 1(1), 189–221.
Jamshidi, M. (2009). System of systems engineering: Innovations for the 21st century. Hoboken, N.J.: John Wiley & Sons Inc.
Katina, P. F., Pinto, C. A., Bradley, J. M., & Hester, P. T. (2014). Interdependency-induced risk with applications to healthcare. International Journal of Critical Infrastructure Protection, 7(1), 12–26.
Keating, C. B., & Katina, P. F. (2011). Systems of systems engineering: Prospects and challenges for the emerging field. International Journal of System of Systems Engineering, 2(2/3), 234–256. https://doi.org/10.1504/IJSSE.2011.040556
Keating, C. B., & Katina, P. F. (2012). Prevalence of pathologies in systems of systems. International Journal of System of Systems Engineering, 3(3/4), 243–267.
Keating, C. B., Rogers, R., Unal, R., Dryer, D., Sousa-Poza, A. A., Safford, R., … Rabadi, G. (2003). System of systems engineering. Engineering Management Journal, 15(3), 35–44.
Kröger, W., & Zio, E. (2011). Vulnerable systems. London, UK: Springer-Verlag.
Maier, M. W. (1996). Architecting principles for systems-of-systems. In 6th Annual INCOSE Symposium (pp. 567–574). Boston, MA: INCOSE.
Stevens Institute of Technology. (2006). Report on System of Systems Engineering: Submitted to the Secretary of defense (p. 53). Hoboken, N.J.: Stevens Institute of technology. Retrieved from http://www.boardmansauser.com/downloads/SoSSEreporttoDoD.pdf
US Department of Defense. (2008). Systems engineering guide for systems of systems. Office of the Deputy Under Secretary of Defense for Acquisition, Technology and Logistics. Retrieved from http://www.acq.osd.mil/se/docs/SE-Guide-for-SoS.pdf
Valerdi, R., Ross, A. M., & Rhodes, D. H. (2007). A framework for evolving system of systems engineering. Crosstalk, 20(10), 28–30.