Equality of Actors in Enterprise Architecture
Enterprise architecture (EA) should not have as its sole focus IT, but it should concern itself with the enterprise as a whole. That was the argument in my first article in a series about EA and its relation to IT. Of course, that does not mean that EA should not concern itself with IT. On the contrary, IT remains the most important focus area of enterprise architects only not with blindfolds blocking all other relevant aspects of an enterprise.
With ArchiMate from version 3 onward, one has the means to model the ‘physical’ parts of an enterprise. Yet, in my second article on the EA and IT topic, I argued that ArchiMate remains IT-centered using the standard Business-Application-Technology layering. In this article I came up with a new EA layering model where IT does not hold the center spot anymore: denoted by a Business-Actor-Infrastructure layering.
Although this new layering resembles the standard layering, the philosophy behind the layering is rather different. In the Business layer concepts are positioned that model an ‘enterprise’ with an abstract viewpoint. One refrains from specifics and looks at the working and structure of the enterprise as a whole in its environment. In general terms, one has ‘composite concepts’ in this layer. For example, the organization concept is positioned in the Business layer denoting the composition of all ‘the active structure elements’ the organization consists of. In the second layer, the Actor layer, concepts are positioned that ‘really’ execute behavior and manipulate ‘passive objects’. The ‘active structure elements’ that are distributed in ArchiMate over three layers (humans or groups of humans in the business layer, applications in the application layer and equipment in the technology layer) are positioned in this Actor layer. The third layer, Infrastructure, has mainly to do with concepts that support actors in executing their behavior. These range from IT infrastructure concepts to physical infrastructure concepts.
Standing out in this new layering is the fact that humans are positioned ‘on the same level’ as the non-human actors, instantiating a ‘different world view’. One can say that it is a shift from an anthropocentric world view to an ecological or environmental world view.
My philosophical foray
To be honest the philosophical reflection that ended the second article and which I repeated here, was just an afterthought while writing the article. As it goes in life, coincidentally I came across Luciano Floridi, a philosopher that in the last twenty years has worked on the philosophy of Information and IT. In a series of books Floridi investigates the role of Information and IT in society, and comes up with a new, fundamental philosophy which he argues is needed because of the massive introduction of ‘smart IT technologies’.
In this article, I will use Floridi’s work to put more flesh on my rather superficial philosophical reflection on doing enterprise architecture. In my opinion, a paradigmatic shift is needed to change the role of the architect, the architectural methods and architectural modelling. Why do we need to make such a shift in enterprise architecture and what are its consequences? And how is that related to treating a diversity of actors more equally by positioning them in the same architecture layer?
This article leans heavily on the work of Floridi, especially on his book ‘The fourth revolution’. Floridi does fundamental research in the nature of information and IT and its significance for humanity. One of his central ideas is that information technology develops from passive technology to active, autonomous technology and consequently impacts how humans see themselves. I will follow his line of reasoning and link it to the enterprise architecture field. But first, in the following paragraph I will briefly summarize the view on IT that still dominates the thinking in organizations and also within the EA field.
The positioning of IT in organizations
The ‘common’ outlook on IT in organizations can be explained as follows. The business part of an organization is where humans do the work making up the essence of the organization such as delivering services and creating products. Gradually, they are supported in their work by IT applications. IT applications run on the IT infrastructure which rapidly is being outsourced to a virtual cloud.
This rather simple perspective already runs short in organizations where physical equipment is to a considerable agree used to execute the organization’s business processes. For example, in hospitals, the military and energy producing companies, IT is also part of the physical equipment and seamlessly interacts with digital technology. Typically, in these kinds of organizations IT departments are responsible for infrastructural IT and administrative applications and the business departments are responsible for the (digitally enabled) physical equipment.
One of the successes of the agile movement is that business and IT start to join in teams, or even in new organizational departments, to reflect the fact that the organizational separation of business and IT is not productive anymore. The development and maintenance of IT applications is done hand in hand with innovating and doing business. Agile methods have been created to support this new way of working.
Thus, we are already shifting to a new perspective in doing business. On the other hand, the old metaphor of IT ‘supporting’ the business by delivering services is still prevalent. And, as I argued, in the EA field this perspective is dominating the field. There are no new perspectives that I know of for repositioning the EA field. The Technology (supports) Applications (supports) Business layering summarizes this position.
History and Information Technology
We normally designate our modern digital equipment with the term ‘Information Technology’. However, for millennia humanity has used information technologies to record and transmit information. To underline the significance of information technology, the difference between prehistory and history lies in the use of information technology in history. The ‘history era’ is also known as the ‘information age’. Floridi argues that we have recently entered the era of hyperhistory with the invention of the computer. The difference between hyperhistory and history is that in history ITs are only recording and transmitting information, where in hyperhistory computers have the capability to process information.
Computers are able to store information, and this already makes a big difference with the laborious process of recording data until the sixties or seventies of the previous century. Moreover, the computer can process this information and make computations that beforehand were the prerogative of humans. Added to these abilities to record and process information, human computer interaction software (HCI is used to create, facilitate and improve communications between human users and computational systems).
We can see this usage of computing power as a rather straightforward extrapolation of the older IT technologies that we have used in ‘history’. One can say that Enterprise Architecture is still dominantly focused on this use of IT: applications that record information and support humans using computers.
Another development in IT is to use computing power to digitize analogue and mechanical tools and machines. Software starts to control and run all kinds of tools and machines. What we call embedded software is growing at a tremendous pace with the Internet of Things technology as the most recent development in this area.
Autonomous artificial actors
This aspect of ‘processing information’ is what sparks a new feature of digital actors, they become autonomous. To analyze this development, we take one step back and look at what Floridi calls the ‘in-betweenness’ of technology. The In-betweenness of technology means that technology is ‘in between’ a user and something which in the first place caused the invention of the technology. To give a simple example: a pair of sun glasses is technology in between the sun and your eyes protecting the wearer from bright sun light. In this sense it is ‘first order technology’: technology that is between human and nature that prompted the invention of sun glasses.
Subsequently, ‘second order technology’ is a tool that is used by a human and that operates on another piece of technology. A simple example is a screw driver that is used by a human and operates on a screw. An example of second order technology that had enormous impact is the engine. An engine is used by a human and it operates on mechanical equipment to provide energy and set the equipment in motion.
In third order technology, the technology itself is using technology that by itself also operates on technology, and so on. IT is in this sense third order technology. For example, in high volume trading, software agents do all the trading only with some supervision of users. Or think of self-driving cars and trains or smart homes that regulate the heating of a house. In third order IT the human as a user is steadily replaced by IT operated actors. We as humans are not anymore in the loop but we are on the loop or completely out of the loop (Floridi, p.30).
In standard EA layering the human workers are viewed as ‘users’ of applications and responsible for the actual work in the organization. This view is not valid anymore when artificial actors are doing the work with humans at best in a supervisory role. That is why my new EA layering model has one Actor layer containing all the different kind of actors: human, analogue / mechanical and digital.
The fourth revolution
Floridi sketches a historic development where humans must realize that they are less special than they always thought they were. We used to think that we are at the center of the universe. Then came Copernicus and proved that earth is not in the center of the universe, but just somewhere in the universe. In the second revolution Darwin showed that all species of life have evolved over years, humans are descendants from apes and that we are not conceptually different from animals in that sense.
At least humans were still thought to be masters off their thought. Then came Freud who uncovered the unconscious and showed that our rational thoughts were not rational at all. However, humans were the only ‘species’ able to reason (mostly rational).
The invention of the ‘reasoning computer’ is the fourth revolution that Floridi identifies for which he appoints Alan Turing as the ‘inventor’. Alan Turing described machines (computers) that process information, reason logically and with this ability have ‘intelligence’. Apart from humans, computers are also able to reason, causing a further relativization of the uniqueness of human beings. As Floridi phrases it
‘we are informational organisms (inforgs), mutually connected and embedded in an informational environment (the infosphere), which we share with other informational agents, both natural and artificial, that also process information logically and autonomously.’
Humans work and live in ‘a symbiotic relationship with their surroundings: animate and inanimate’ is how I described it in my previous article.
It is questionable if digital machines can be called intelligent (already), yet it is certain that they can ‘outsmart’ us and perform tasks in which they are far better than us humans.
Autonomous actors and its consequences for enterprise architects
The responsibility of the design and the creation of the artificial (digital) tools still rests with us humans. Enterprise architects are professionals with an important role in this design process. Alas, the enterprise architect profession still looks at IT as supportive data recording and processing technology and not as digitally autonomous actors. If enterprise architects have the ambition to design and guide the future evolution of organizations and its technology, this oversight has to be repaired.
In my two previous articles, I have pointed out two areas in need for improvement. First of all, enterprise architects should be concerned with the enterprise as a whole, not focusing solely on IT. Primarily because IT is such a pervasive technology that it is artificial to make a separation between IT and non-IT, and moreover because enterprises will become increasingly digital. The second change is to treat actors and their behavior in the same way (on the same layer): it is just not much of a difference what kind of actor is doing what kind of work, so in a layering view of enterprises one should paradigmatically treat them on the same layer.
An increasingly important design question will be how to control the behavior of autonomous actors with humans out of the loop. Moreover, it is important to ask how to create self-regulating software that enables systems to be self-learning while remaining predictable and how to govern the behavior of these systems.
Above all, this new position of a human not above other ‘actors’ but on the same level, also extends to the role of the designer (or architect) of the enterprise ecosystems. The perspective of the designers likewise changes from anthropocentric to ecocentric, and so one can say that this means a shift from ‘masters of the universe’ to ‘shepherds of the universe’. As masters of the universe designers focus primarily on effectivity and efficiency of enterprises, as shepherd of the universe care for actors and their environments will become more important.
