CC BY
44. Smetanin V.I.Protection of the environment from production and consumption waste: a tutorial / V.I.Smetanin. - Moscow: Kolos, 2000. P.232
5. Minh Tue, N., Katsura, K., Suzuki, G., Tuyen, L., Takasuga, T., Takahashi, S.,Tanabe, S. (2014). Dioxin. Related compounds in breast milk of women from Vietnamese e waste recycling sites: Levels, toxic equivalents and relevance of non dietary exposure. Ecotoxi-cology and Envirionmental Safety, 106, 220-225.
6. Perkins, D., Brune Drisse, M.-N., Nxele, T., & Sly, P. (2014). WEEE: A Global Hazard. Icahn School of Medicine at Mount Sinai., 80, 286-295.
EXPLORING ADAPTIVE RESOURCE MANAGEMENT IN NETWORK ENVIRONMENT: CONSIDERATIONS ON INTERNET OF THINGS USE CASE
Usmanova N.
D.Sc., Professor, Telecommunication technologies Dpt., Tashkent university of information technologies
Yunusova D.
Master of Telecommunication Engineering, Tashkent university of information technologies
7. Loleit S.I. Development of environmentally friendly technologies for the complex extraction of precious and non-ferrous metals from electronic scrap: 05.13.01 / M, 2009 - 41 p.
8. Strizhko L.S., Fokin O.A., Shigin E.S. Analytical control and certification of electronic scrap containing precious metals, National Research Technological University "MISiS" 2009 https://www.waste.ru/modules/sec-tion/item.php?itemid=255.
Abstract
In terms of ever widening the scope of digital economy it is important to find out the ways for effective functioning the components of complex Internet of Things (IoT) network environment, allowing the applications to be duly provided whereas IoT is based on the idea that things will be available at any time, in any place and for anyone, concatenated into a single system, thus creating new opportunities and challenges for the various application domains. The issue of how to attract the adaptive resource management in IoT is considered in this paper. Capabilities of IoT network environment are briefly described, and demonstration of adaptive resource management is provided based on use-case simulation.
Keywords: Internet of Things, network environment, adaptive resource management
Introduction
The ever-increasing rate of change in the society due to the digital economy and major development trends of the modern world via the intensive introduction of information technologies transform significantly all spheres of human life, permeating almost every aspect including how people interact, how the economic landscape is being shaped, making business processes improved, influencing to decision making and the skills needed to get a good job. An emerging digital economy has the potential to generate new scientific research and breakthroughs, fueling jobs and economic growth. The level of penetration and the degree of implementation of the digital economy is defined among main development priorities, both in a single country, and worldwide, caused by the fact that the competitiveness of the country's economy in the future can be determined by the level of digitization of all and every activity and processes. The digital economy reflects the transition from the third industrial revolution to the fourth industrial revolution, Industry 4.0, wherein the concept of the Internet of Things (IoT) technology came onto the arena representing a set of interconnected physical objects -things, which are equipped with built-in technologies for interacting with each other or with the external environment [1].
Digital economy relates to the economy that is based on digital technologies, including digital communications networks, computers, software, and other related information technologies. The digital economy includes the following key components: technological infrastructure - hardware, software, and communication networks; digital processes - processes that ensure the successful conduct of a business; e-commerce - the sale of goods via the Internet. These components along with the Internet of things, shape the 'smart environment' in broad meaning, allowing the applications to be duly provided whereas IoT is based on the idea that things will be available at any time, in any place and for anyone, concatenated into a single system, thus creating new opportunities and challenges for the various application domains [2,3]. By linking smart devices, conventional consumer elements, and physical ownership over the Internet, the Internet of things erases the boundaries between Internet technologies and products, which do not fall into that category and thus achieve significant social, technological, and economic benefits [4].
The improvement in the management capabilities and available bandwidth offered by the complex environment of smart and ubiquitous infrastructure and networking is accelerating the development of new kinds of applications, interfaces, and services. The ultimate goal of such networks is to automatically adapt their
services and resources in accordance with changing environmental conditions and user needs. Such 'adaptive' capabilities imply the usage of sophisticated technologies in order to integrate every object of the environment that would be enabled with important computational power and storage capabilities. The challenge behind such implementation is to simplify the managerial task by automating the decision-making process, and enabling the users to seamlessly find their way in such pervasive environments. This paper considers the use case to cope with such a challenging task of adaptive resource management in IoT enriched environment. Although it is not pretending to describe the theoretical background in this sophisticated area of research, the appropriate considerations related to the performance of objects are given based on use case simulation.
Resources in IoT Network Environment: why adaptive management matters?
The term "Internet of Things" usually refers to scenarios where objects, sensors and other items of daily life, not usually considered computers, are equipped with network connectivity and computing functions, so that these devices can generate, use and exchange data with minimal human intervention. However, there is no single and universal definition. While there are many definitions of the Internet of Things, one of the simplest ways to describe it is 'connecting things to the Internet', whereas things are physical objects in which technologies such as sensors and actuators are embedded, as well as software and network connections. Connecting things to the Internet allows them to interact and exchange data with other devices and systems, and they can also be controlled remotely. IoT devices are self-contained data collection points that can provide instant or archived up-to-date information. They can interact with other systems in different ways depending on their function. Some may periodically send data, some may receive commands, and others may trigger actions and alerts. This versatility makes the IoT an essential element for automating and optimizing business processes [5].
According to the ITU-T Y.2060 - Overview of the Internet of things recommendation [6], IoT reference model is composed of four layers as well as management capabilities and security capabilities which are associated with the four layers, namely: application layer, service support and application support layer, network layer, and device layer. Each layer represents its own capabilities, e.g. network layer consists of the two types of capabilities: networking capabilities (providing control functions of network connectivity, mobility management or authentication, authorization and accounting), transport capabilities (providing connectivity for the transport of IoT service, application specific data information, as well as the transport of IoT-related control and management information).
In general, it is supposed to collect information from remote devices, control these devices, exchange data between them, redistribute tasks, analyze the information received, and plan the operation of an object taking into account the received data. To track such
processes, certain communication, analytic and security tools are required, which, when combined into a common platform, form a single system of interaction, not only with objects, but also with other users.
It is obvious, that environment to form and support all of those communication and interchange capabilities is complex and pervasive. The main reason is due to the requirements to deal with technologies and have specific 'managerial' skills; such technologies need to integrate every object while optimizing the control of its powerful local computational and storage capabilities. Among such skills the ability to support any kind of communication through wireless technologies, possibility to reconfigure when 'tailoring' the parameters is required, etc. [7,8]. In this problematic context different classes of problems to be solved on the basis of IoT: the first is to remotely monitor and manage a set of interconnected network devices, each of which can interact with infrastructure and physical environments (e.g. temperature and humidity sensors control a network of devices that control the climate system of a smart building); another task is to use data received from smart devices with the ability to connect and probe, for intelligent analysis in order to identify trends and relationships that can generate useful information (e.g. analysis of daily air temperature fluctuations will predict the need to turn on or turn off the heating system). If the technological integration of computational autonomic nodes poses several problems, the seamless integration and cooperation of situated-autonomic communications nodes represents an interesting and novel research challenge. It will not be sufficient to dynamically integrate the nodes from the communication point of view, but it is also necessary to provide them with the means to share their knowledge of what is going on in IoT network environment.
Thus, there are three major aspects to consider when dealing with models and algorithms for adaptive resource management in IoT network environment:
1) How to choose and create models and algorithms that will support adaptive behavior.
2) How to represent models that are necessary for adaptive resource management in IoT network environment to achieve an adaptive behavior.
3) How adaptive objects communicate together, organize among themselves in a possibly large context.
There are different approaches developed and experimented in related research, e.g. [9-11], with describing the analytical background for evaluating the behavior of components, however, a few works is devoted to the issue of adaptive resource management in IoT network environment. Herewith the attempt is given by the authors of this paper to demonstrate the characteristics of components described within state-flow (SF) model while attracting adaptation in the complex environment (this is mainly inspired by the considerations on modeling IoT-based self-adaptive software given in [9], with classification of IoT devices as sensor-devices and act-devices used for finite-state machine modeling: sensor-device senses the changes in environment, act-device can change the environment).
Use - case implementation and results
To study the characteristics of the IoT the model is build in the Matlab-Simulink-Stateflow environment
(Figure 1), which includes the following blocks: Pulse Generator; Chart; Display.
Fig.l. Model in the Matlab-Simulink-Stateflow environment
Pulse Generator block generates clock pulses that set the speed of execution of operations in the transition diagram Chart. The Display shows the simulation results.
Chart block is intended for creating Stateflow event modeling diagrams (SF diagrams). The Stateflow event-driven simulation package is based on finite state machine theory. It allows to represent the functioning of the system based on a chain of rules that define the correspondence of events and actions performed in response to these events.
An SF chart is a graphical chart created by the graphical user interface tools in the Stateflow add-on package. SF-diagram is used for visual representation of the modeled system operation. This is achieved by analyzing all stages of its work, indicating the active and passive blocks at a given time and transitions between them based on the results of the analysis of certain conditions. At the same time, the blocks differ in the color and thickness of the lines with which they are represented.
The main object of an SF diagram is a state. The properties for active and passive mode of the state dynamically replace each other depending on the events taking place. Each state has its own parent and can have children (lower-level states). If the state is unique, then its parent is the SF diagram itself, also called the root diagram.
The diagram consists of states nested within each other. The ground state is called exclusive. If a single state is being built, then it will be exclusive. Each state can be a parent, that is, it can have its own children. If another state is included in any state, then the first one will become a parent.
States can be neutral or engaged. There are two types of states: parallel, existing simultaneously- concurrent (AND) and mutually exclusive (OR).
Each state has a memory attribute, or its own history, which provides a determination of the future transition to another state based on information about the past state of the system. The memory feature is the function with the highest execution priority.
Transitions are another graphical object of SF diagrams. Transitions represent the relationship of one object to another and are usually represented by red arrows. The occurrence of an event changes the status of its associated states and can trigger an action or transition associated with it. In this case, events are propagated from top to bottom (from parent to child). Each event must be defined using one or another condition, written as a logical expression. The main property of events is the visibility property of the event. The Action Language is based on the C syntax and contains arithmetic and logical operators and functions, some special functions and user functions.
Following functions are denoted:
- change (data_name) or chg (data_name) - generate a local event when the value of the data specified in the argument changes;
- in (state_name) —condition function that evaluates to true if the state specified as an argument is active;
- send (event_name, state_name) - sends the event specification to the state specification (direct event transmission);
- matlab (evalString, arg1, arg2, ...) or ml () - a procedure that performs calculations written in the evalString string in the notation of MATLAB functions (ml notation), with the arguments listed after it;
- matlab.MATLAB_workspace_data or ml - a procedure that performs calculations using ml notation.
To study the characteristics of IoT, taking into account the initial data in the Chart block, it is necessary to create the SF-diagram shown in Figure 2.
Fig. 2. SF diagram for N=2
In the st1 state, the initial data is set: i- is the address of the sensor device;
q- is the probability of an unsatisfactory state of the controlled object; k- is the total number of experiments.
After establishing the initial data, the transition to the state st2 is carried out. In this state, the sensor devices are polled in a cyclical manner. The address of the next sensor device is formed and the request is transmitted. The request transmission time is taken equal to 1 ms. At the beginning, the request is passed to the first sensor device (state st3).
The measurement time of the sensor device was taken to be 0.3 ms. In the state st3, it generates a random number in the range from 0 to 1 (p). If the readings of the sensory device are satisfactory (p> = q), then the response "normal" is transmitted to the state st2. If the readings of the sensor device 1 are unsatisfactory (p <q), then the response is "abnormal" to the state st7. The response time is assumed to be 1 ms. In the st7
state, the control device, depending on the indication of the sensor device, generates and transmits the corresponding control command to the executive device 1 (state st4). The control command generation time is 0.2 ms, and the control command transmission time is 1 ms. In the st4 state, the executive device, in accordance with the control command, performs an action to normalize the state of the object. After performing the action (0.3 ms), the response "norm" (1 ms) is transmitted to the st2 state. Next, the polling of the next sensor device begins.
The results of calculating the average delay time for servicing sensor devices are shown in Table 1. The number of sensor devices is 2 (N = 2).
Average service delay time for N = 2
Table 1.
q 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
T,mc 4.8 5.2 5.6 5.9 6.5 7.2 7.8 8.3 8.8
Figure 3 shows the transition diagram for N = 3. The calculation results are shown in Table 2.
Average service delay time for N = 3
Table 2.
q 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
T,mc 7.3 7.5 8.1 9.2 9.9 10.6 11.8 12.6 13.2
Fig 3. SF diagram for N=3 Dependency graphs are shown in Figure 4.
Fig. 4. Dependency graphs for N=2 and N=3
Based on the analysis of the results obtained, such a regularity has been revealed that, for a given q, an increase in the sensor device by one leads to an increase in the average delay in servicing sensor devices by 1.5 times.
Figure 5, based on this regularity, shows a graph of the dependence of the average service delay time on the number of sensor devices at q = 0.1.
Fig. 5. Average service delay time on the number of sensor devices
It follows from Figure 5 that with the considered initial data at N> 10, it is necessary to increase the number of interrogation devices for sensor devices or to increase the data transfer rate in the communication channels.
Conclusion
The complex and ubiquitous IoT network environment deals with various requirements to accomplish different objectives in a changing conditions, while states of network nodes and environment itself must be dynamically satisfied, expressing the adaptation feature. To address this issue and to study the characteristics of IoT, a simple model was built in the Matlab-Sim-ulink-Stateflow environment, which includes the Pulse Generator blocks; Chart; and Display. The functions within SF-diagram, states and transitions are described, along with data types that have a visibility property. To create and modify data, the Add ^ Data menu items of the SF-diagram editor were used, and the general structure of the state and the meaning of these definitions are described. The Action Language based on the C syntax contains arithmetic and logical operators and functions, some special functions and user functions. The simulation results obtained for the average delay time for the service of sensor devices, and the transition diagrams are presented. This kind of study in a proper modeling more complex infrastructures may give appropriate patterns to be revealed based on the analysis of the results obtained. Authors consider this task for further research with providing more detailed analytical background and simulating the IoT environment for adaptive resource management.
REFERENCES:
1. Rayes A., Samer S. The Road to Digitization. Volume 49. Internet of things—From hype to reality. Springer Nature Switzerland AG 2017, 2019 (eBook) https://doi.org/10.1007/978-3-319-99516-8
2. Going Digital: Shaping Policies, Improving Lives. OECD The Going Digital project overview [electronic resource] //https://doi.org/10.1787/ 9789264312012-en (retrieved: 24.10.2021)
3. Digital Economy Report 2021. Cross-border data flows and development: For whom the data flow/United Nations publication, UNCTAD/DER/2021, Geneva, 2021
4. Heath D., Micaleff L., What is digital economy? Unicorns, transformation and the internet of things [electronic resource] //https://www2.deloitte.com/mt/en/pages/technology/a rticles/ mt-what-is-digital-economy.html (retrieved: 20.10.2021)
5. Efimov M.M., Kirichek R.V. The Internet of Things: Prospects for Adaptive Systems //Telecom IT. 2020. Vol. 8. Iss. 1. pp. 55-66 (in Russian). DOI 10.31854/2307-1303-2020-8-1-55-66.
6. ITU-T Y.2060 - Overview of the Internet of things//https://www.itu.int/rec/T-REC-Y.2060-201206-I
7. Brogliato B., Lozano R., Maschke B., Egeland O. (2020) Adaptive Control. In: Dissipative Systems Analysis and Control. Communications and Control Engineering. Springer, Cham. https://doi.org/10.1007/978-3-030-19420-8_8
8. Nazim Agoulmine , Sasitharan Balasubramaniam , Dmitri Botvitch , John Strassner , Elyes Lehtihet and William Donnelly '"Challenges for Autonomic Network Management" //https://citeseerx.ist.psu.edu/viewdoc/download?doi= 10.1.1.101.8844&rep=rep 1 &type=pdf (retrieved: 12.05.2021)
9. Euijong Lee, Young-Duk Seo, Young-Gab Kim, Self-Adaptive Framework Based on MAPE Loop for Internet of Things/in Sensors (Basel). 2019 Jul; 19(13): 2996. Published online 2019 Jul 7. doi: 10.3390/s19132996
10. Nazim Agoulmine, Autonomic Network Management Principles: From Concepts to Applications/2011 Elsevier; https://doi.org/10.1016/C2009-0-62958-5
11. Mezghani E., Exposito E., Drira K. A model-driven methodology for the design of autonomic and cognitive IoT-based systems: Application to healthcare. IEEE Trans. Emerg. Top. Comput. Intell. 2017;1:224-234. doi: 10.1109/TETCI.2017.2699218
