- The role of interface in the relationships between various entities in a system environment -
‘Knowledge is a process of piling up facts. Wisdom lies in its simplification.’
- Martin H. Fischer
Dr. Fischer’s words can also be applied to the current trends and roles of IT. Isn’t this the answer to converting and refining information to give value to the digital big data that is exchanged among the countless number of connected entities? In this context, this is connection of management systems for person to person or machine to machine, person, environment or business through messages and signals for exchange, interaction and transmission by a simplified and convenient service.
In this process, IoT (Internet of Things) and IT have a very broad scope and play both ambiguous and defined roles. Today we will take a look into the IT involved in the relationship among various entities in an environmental system, focusing especially on interfaces.
Virtualization Built on Structured Order
By observing the characteristics, actions, intentions and the interactions that occur between the countless connected entities, we can collect a vast amount of data. Also, we can learn a lot about not only the order inherent in the entities but about the hidden order and attributes through the process of categorizing and systemizing the entities. This virtualization of reality is the essence of homomorphism mapping or modeling. In the virtual world, we can find each individual difference between entities through the modeling of patterns and parameter.
But in the computing environment, things are different because each entity has its own operating system and method of communication. So, it is very difficult to structure and develop a program to manage the order and information so that all of the entities can interact. Not only that, but the currently connected devices are being developed so that they can communicate in businesses and industries that extend across the entire infrastructure.
Let’s look at the connected car driving system that is receiving much attention recently as an example. The connected car driving system extracts various data from devices in the car such as mechanical diagnosis data, driving conditions, the state of the driver, driving habits, weather, road conditions, and traffic lights. It also provides services (collision prevention, inspection notifications) by modeling the interactions with all possible circumstances and communicating this data to a smart device. However, the system is currently limited because services are restricted by the relationships between certain car manufacturers and device brands.
The Smart Car, Where the Car and IT Meet!:
http://www.lgcnsblog.com/technology/the-smart-car-where-the-car-and-it-meet/
Finding the Similarities and Patterns between Different Objects!
We briefly discussed mapping before, but mapping and schematization of abstract objects and concepts began at the beginning of civilization. Modeling and simulation engineering was developed with the advent of the computer. We have even begun the modeling of people themselves. This sort of modeling began with the modeling of groups of individuals, then it moved to the modeling of more subdivided groups and now we are developing modeling of the relationships between each individual.
Sierpiński Pyramid and fractal self-similarity of ecosystems, groups and individuals. (Source: Senexrex, BoostZone)
As shown in the image above, each individual person, ecosystem, society, group, industry, business and machine looks very different but in fact, they are not. Within the order in the chaos of a system or phenomenon, a fractal structure and relationship can be found.
What sort of classifications can be done to further understand the relationships between objects?
1) Understanding Relationships through Clustering and Topological Structure
The purpose of a system or even a standard of monitor resolution has fixed boundaries and abstraction. Also, entities make relationships between each other. These relationships occur according to need and probability.
For example, let’s assume that there is a node on a network for each of ‘n’ number of entities. And in this case, if there is an extremely complex network of these entities made up of n(n-1)/2 number of connections, this is known as mesh topology. Of course, it is difficult for a completely connected mesh topology to exist. But in actuality, the microcosm known as the human brain is made up of approximately 100 million neurons and each neuron has approximately 1,000 synapses that are able to make a connection. There are also approximately 100 trillion connections within the human brain.
Tree form mesh topology (left) and a similar factory manufacturing system structure (right) (Source: http://www.renesas.eu/)
The fist connected societies did not have this sort of complexity. But in the virtual world, mesh topology is necessary for modeling, development and management. Mesh topology makes structuring, graphing, management and expansion easier. It becomes especially easy to express parent/offspring or parent/subordinate as well as class and inheritance through a detailed Sierpiński Pyramid or tree mesh topology. There are many cases for expressing data with nonlinear mesh topology such as manufacturing systems, factories or company organizational structures. Complex data structure design and investigation for things such as operating things such as computer topology, most OS file systems, search engines, XML DOC(Document Object Model), data bases and object oriented languages also become much easier.
It is also possible to materialize a network structure for connecting the equipment to automate the manufacturing system of a factory.
Integrated bus infrastructure conversion for conventional manufacturing network interface (Source: http://www.renesas.eu/)
The data is represented in a typical pyramid network structure. The system works by allowing the business system, the plant manager, development device (CNC, robots etc.) controllers (SCADA, Supervisory Control and Data Acquisition), equipment and tools such as embedded sensors and actuators to communicate through hierarchical control. However, the problem with this system is that each class of equipment uses a different protocol. So, because this process is complex and inefficient, a communication interface is required for each device and each class. Thus, actions are being taken to gradually simplify the system. A middleware known as USB (Enterprise Service Bus) is used to integrate a compatible message communication interface into a SOA (Service Oriented Architecture) for making nonlinear structures linear.
This type of middleware integrates functions such as message converters into a larger server interface makes it possible to create a simplified interface for each device with standardized regulations through a library provided in the architecture by the client or developers. Commercial service platforms such as JBossFuse, IBM WebSphere ESB, Oracle service BUS and WSO2 along with the Korean LG CNS Smart Green Platform provide Smart ESB. This sort of integrated middleware creates an important interface the onsite IT system and the cloud web service or factory that are the aggregate virtual resources of the industrial Internet system.
Network Mesh Topology on the Internet of Things (Source: http://radar.oreilly.com/)
Network mesh topology structure of the Internet of Things as the extended version of each system is expected to take various forms. Bus topology for 1:1 bus connection and star topology (version 2 of tree form topology) are expected to be developed based on mesh topology. 1:1 bus connectivity structure is a direct communication structure for cloud to be used on each device or onsite. This is more suitable for complex combined device or organizational structures that mesh topology. This is a very natural structure and suitable for covering long-distance connections.
But modifications of the bus and tree structure are still needed in order to add routing nodes through clustering and classification as the complexity of systems increases exponentially. It also seems that Hadoop distributed file system or cloud server routes will support saving and communication computing on gateway nodes.
2) Understanding Surroundings and Relationships through Dynamic Entity Situation Recognition
Despite the schematization and standardization done through scientific technology, computing and IT, research is still being done to establish connections and relationships between different entities. Many things must be considered and this is still a very complex field that requires a lot of development. Also the independent ‘dynamic entities’ have become the center of attention as they are part of the assessment of relationships between different entities with a system and often change such as the location or status concerning physical entities. In a manufacturing system, these dynamic entities are products or robots or in the case of a service system, they are people. Because of the high level of complexity and diversity, abstraction for people, which are much more complex than the basic machine, is expected to be the most difficult.
However, with the application computing and AI (automata, feature extraction, machine learning, metaheuristic etc.) and statistical methods that can be found in nature or in living organisms, computer agents and device interfaces are being developed that bear likenesses to their creator of environment. Computer agents read the physical and psychological condition of the entities when they are taking an action or making a response based on information accumulated from entire population of entities. Then, through simulations, the agent is able to respond appropriately to things that occur and even predict them.
As renowned professor Arthur B. Markman said, there is a limit to the human memory and the emergence of a computer agent that will remember for us and assist us in our work will be welcomed. The computer agent will require a mediator as an interface and coordinator in a broad sense. This mediator is known as middleware.
Anticipatory computing engine and the relationship between the present and the future
Through this type of technological development, the future looks bright for IoT and cloud computing based on machines that can make decisions and communicate. For example, the early stages of the IoT are already providing us with great technologies such as keyless entry, automated temperature control systems and automated airport check-in systems. In a system where a person interacts with devices and an environment, the goal is to make device -> environment or device -> device connection and communication possible while promoting convenience for the user. This technology can also help people enjoy life with concise interfaces, anticipatory computing and analysis through the elimination reduction of human intervention along with automation and data.
This is the same for lean management in manufacturing where IT is implemented for demand forecasting when establishing a production plan along with the human cognitive process. The amount of a product that will be needed can also be predicted and the current production quantities can be adjusted. Unnecessary processes can also be eliminated with the IoT to provide an intangible service. The goal is to monitor the current user’s actions and status and to predict future users’ actions to improve the convenience of the system.
Shall we take a look at how this system works for automatic check-in at the airport?
1. The current process
1) Enter the airport → 2) Find the airline check-in counter → 3) Stand in line at the counter → 4) Wait restlessly → 5) Arrive at the desk, verify ID and check-in → 5) Check your luggage → 6) Run to the gate
2. The Service We Want
1) Enter the airport → 2) Verify you status with your smart device and the system will send a message to your device explaining that you have been checked in and how to check your luggage, your gate number and boarding time → 3) Check your luggage as instructed → 4) Make you way to the gate
The current airport check-in process and the check-in service we want are described above. But actually, the capability of IT is all lumped together in step 2-2.
Systemized Human Perception Processes in an Interface/Computer Agent. (Source: Perception and BDI reasoning-based Agent Model for Simulation of Human-Environment Complex System)
The image above walks us through an interface computing process that follows the human perception system as it takes place in the brain. Perception → inference (recognition → assurance → deliberation → intention) → decision → plan and action conclusion. Similar to this process IT takes the place of the check-in steps and sends a message via the interface to the user. And like the example above, services are already in use that can suggest contents based on the characteristics or attributes of an individual situation. A couple of examples of this are systems that suggest restaurants in the vicinity of a user or suggest music according to a user’s taste.
Various development and OS modeling that maintain our hyper connected society are focusing on the similarities in individuals, groups, industries and other members of systems and adjusting to them. This technology is also bringing order to the natural complexity of systems and the interactions between individuals in those systems. The technology overcomes limits and other variable with standardization, simplification and modulation through analysis and reduction of differences. It also looks as though eventually modeling of relationships between entities in this fractal environment can be done to produce templates that can be reused. We can see that with the combination of these templates and applications even more complex templates can be produced and an amalgamation of IT including the current diverse domains will exist.
We have now discussed the role of IT in the process of understanding the classification, awareness and relationships between entities in a system. In the next installment of this series we will look into the hidden complexities with a basic system and discuss the structure of interfaces.
Written by SeungYup Lee
Finished his doctorate at Pennsylvania State University and is now working as a researcher. Received his masters in industrial systems engineering at Yeonsae University his doctorate in industrial engineering at Pennsylvania State University. Is currently focusing on Formal model based hybrid SoS engineering, Diverse Simulation based methodology (agent, discrete event, system dynamics), manufacturing and supply chain systems, IT-based industry convergence and Cyber-Physical System .