Skip to content
Create an account for full access.

Systems of systems

Sometimes a system consisting of other systems is called a "system of systems". This is incorrect; it is simply a system, and the other systems within it are subsystems. It seems unnecessary to emphasize the fact that any system consists of systems. Nevertheless, the term "system of systems" (SoS) exists; it is associated with a special case of identifying systems based on their social characteristics, the characteristics of their creation, and the timing of their creation, rather than purely technical features of the nesting of systems within one another in the environment and their operation during usage.

A "system of systems" is a system that meets the criteria set by Maier:

  • Independent systems of creation/enabling its subsystems (the responsible party for the SoS cannot dictate the general development-modernization for the systems within the SoS)
  • Subsystems that can function independently of the existence of the target system (there are no authorities to direct the owners of the systems on how to operate)
  • Emergence/system effect from the integration into the system (the creator wishes to obtain functions/behavior from the target system, which cannot be achieved by working with the individual systems inside the system, and requires the collaborative work of these independent constituent systems within the SoS)
  • Evolutionary development (understanding what will happen in the system of systems at each subsequent step of the system creation program requires research and additional coordination, as there is no role that knows how the system of systems is structured at any given moment - the subsystems of the system of systems, as autonomous systems, change independently. This leads to a series of modernizations rather than a one-time project action. However, this does not differ much from "ordinary engineering" today, as even with "simple systems," they also undergo evolutionary development, although in this case, everything is known about the structure of the system organization, and the subsystems are not autonomous)
  • Geographic distribution of the systems within the SoS.

These criteria may vary in different engineering and management schools, but the general idea remains the same:

  • Normal "systems" imply centralized "ownership" of the system - the presence of project roles/stakeholders/"interested parties" authorized to make decisions regarding all parts of the system, empowered to manage everything within the boundaries of their system. This is the traditional case: a car with an engine and wheels, a railroad bridge, and a computer are typical "simple systems" that have their system engineers who entirely define their functions, structures, capabilities to work in different environments, formulate and execute plans for modernization and decommissioning. Each of these "simple systems" has one owner.
  • In a system of systems, each of the systems within it has its owner, and the subsystems can work, operate, or function autonomously without being part of the system of systems. This includes a city, a transportation network, the internet, any company - after all, individuals within these systems are self-proprietors and do not hurry to execute someone else's orders.

Thus, the difference between a "simple system" and a "system of systems" is defined not through a particular system structure or construction (defined during usage) but through the presence of independently created systems (defined during the creation phase) that describe and embody the individual constituents of the system, giving them the ability to be independently used. Therefore, metaphorically, we can represent a system of systems with a diagram showing that the system of systems consists of three subsystems ("constituent systems" of the system of systems), each of which is autonomous and, in turn, consists of its subsystems (showing two system levels), but the key point is that the figures holding briefcases indicate the agent-owners who must agree that autonomous systems can form a coherent system.

In a system of systems, the critical aspects are the owners of the system components and the parties authorized to allocate labor and capital resources (individuals, businesses); they are the ones who make a system of systems/SoS a unique case.

ISO/IEC/IEEE 21841:2019 identifies four types of system systems differing in their degree of autonomy:

  • Directed systems, where there is a designated architect (an agent acting as an architect, responsible for dividing the system into constructs and ways of organizing communication between the constructs) who can issue architectural decisions to the project teams of the constituent systems and a manager (an agent acting as a manager) who manages common resources.
  • Acknowledged systems, where there is a recognized architect, but they can only persuade the owners of the SoS constituent systems to adjust their systems according to the architecture developed by the architect. The architect has no governance authority, only mentoring regarding the architectural decisions they make.
  • Collaborative systems, where the owners of all constituent systems negotiate with each other extensively, but there are no organizational links (agents performing roles) for an architect, project manager, or equivalent designated organizational roles involved in creating and developing the SoS at the system level.
  • Virtual systems, where the owners of systems making up the SoS do not know anything about each other, and thus the project teams creating and developing the constituent systems do not explicitly influence each other.

Over a couple of decades of trial and error, the primary way of working with system of systems has been established as "the best practice": a joint, gradual, asynchronous evolution/continuous modernization of the autonomous systems within the system of systems - as attaining consistency and synchronization in changes within these autonomous constituent systems "within a single project from a single center" is extremely challenging.

The approval dates for modernization projects of individual constituent systems will differ, making it difficult to predict the emergence of functionality in the target system of systems. This leads to uncertainty regarding the funding allocated at different times for the modernization of system constituents. No one manages the overall reform project engineering and management- let alone a shared goal/strategy coordinator - as there is no unified vision for all, ultimately leading to ongoing modernizations instead of a one-time project. However, despite their differences, these methods emphasize collaboration, with an understanding that to guide the projects of creating any target systems, the creators are essential and should not be viewed separately from the systems. Therefore, our course will not focus specifically on systems of systems.

Communities (where everyone is occupied with their own tasks unlike an organization), society, and humanity (“agentality”) as a whole are formally also systems of systems. However, it is challenging to imagine a target system developed by these creators - except for the reproduction/replication of themselves and the sustainment of some autonomy from other systems - comparable to the discussion about biological systems. The engineering approach is only just starting to be considered here. The creators of such systems are understandably skeptical: history has seen enough "fathers of the nation", "leaders of the people", and other dictators who formally declared themselves the creators of significant communities and societies, but deeming social engineering projects successful initiatives is a critical question.

The primary argument against the engineering view of changes in communities and societies is that "people within these systems do not understand anything; leave it to evolution, just like in the case of human beings - do not engage in eugenics, there is no need to improve the human race because even if it is safe, it will generate inequality and other unpleasant consequences concerning justice and ethics". The main counter-argument is that "evolutionary trials in the majority of cases lead to errors, with results ending in extinction. Perhaps not letting us go extinct, transitioning to techno-evolution with smart mutations rather than routine ones, is a better approach. If we do not let sick and weak people die and instead treat-teach them, then why not do the same with communities and societies instead of letting them function as they are?".

Note the impersonal statements in the preceding paragraph: who will be the agent capable of engineering societal change, "curing-teaching"? A dictator with a tool for making changes, such as a "small army"? A sweet-talking politician? And if there are three such politicians, each saying something different, with each having a small army - should we leave all three? This itself is a form of evolution, and it is up to us to decide whether it will be Darwinian or techno-evolutionary. Unfortunately, there is no well-developed theory of collective agents that could predict the outcomes of evolution. Therefore, even systemic thinking capabilities in projects of social engineering (engineering systems beyond organizations, where the structure of labor and capital allocation is clear) are limited. However, they are adequate to prevent delving into utopias, like not building socialism based on the claim "all previous attempts were bad because it was not me doing it, but it will work for me, I am good and love the poor".