CSC520

PRINCIPLES OF OPERATING SYSTEMS

TITLE
SMART OS: THE FUTURE WITH AI AND IOT

GROUP
CDCS2301B

GROUP MEMBERS

NAME STUDENT ID

HASINA NAFISA BINTI ZAHIRUDDIN 2025302863

NUR RAZAN MAISARAH BINTI 2025162973

ASKAIMI @ ABOT

NUR SYARIZANIE ELLYANIE BINTI 2025188983

SUHAIMI

LECTURER
MADAM LENNY YUSRINA BUJANG KHEDIF

SUBMISSION DATE

1
 TABLE OF CONTENTS

1.0 INTRODUCTION................................................................................................................... 3
2.0 THE IMPACT OF ARTIFICIAL INTELLIGENCE AND MACHINE LEARNING ON
OPERATING SYSTEMS..............................................................................................................4
2.1 Enhancement In Resource Management and Scheduling....................................................4
2.2 Flexible User Interfaces....................................................................................................... 4
2.3 Memory Management.........................................................................................................5
2.4 Energy efficiency..............................................................................................................................7
2.5............................................................................................................................................................ 7

3.0 INTEGRATION OF INTERNET OF THINGS (IOT) DEVICES WITH OPERATING

SYSTEMS.......................................................................................................................................8
3.1 Opportunities........................................................................................................................8
3.2 Challenges of integrating IoT in Operating Systems.........................................................12
3.3 Future Trends in operating systems................................................................................... 14
5.0 CONCLUSION...................................................................................................................... 18
6.0 REFERENCES.......................................................................................................................19

2
1.0 INTRODUCTION

3
2.0 THE IMPACT OF ARTIFICIAL INTELLIGENCE AND MACHINE LEARNING ON
OPERATING SYSTEMS
As the early operating system focuses on basic management especially in memory allocation,
input/output handling and process management, it might need to evolve to comply with the
demands of advancements in hardware and user requirements. By developing artificial
intelligence and machine learning in operating systems, it has contributed to the rapid growth of
modern computing. In this context, this section will explain the impact of AI and machine
learning in operating systems.

2.1 Enhancement In Resource Management and Scheduling

Artificial intelligence has strengthened resource management by utilising Q-learning and Deep
Q-Network (DQN) approaches in order to foretell requirements of a resource and modify CPU
allocations, plus reduce latency and increase throughput.
●​ Q-Learning focuses on task prioritisation by dynamically modifying it based on
system load.
●​ DQN improves decision making by learning from past allocations.
●​ Development of machine learning in operating systems such as Google’s Borg
system has enhanced resource scheduling particularly in data centers, making
certain of workload efficiency.
These different approaches of AI and machine learning have contributed to effectively
anticipating resource management by dynamically re-allocating I/O, memory and CPU resources
based on real-time analysis, as well as balancing server loads time efficiently.

2.2 Flexible User Interfaces

By employing recommended algorithms in user interfaces, it has delivered user friendly

interfaces according to user preferences. For instance, Google’s Assistant and Apple’s Siri are
examples of AI driven applications that offer user support through reinforcement learning and
dynamic user layout adjustments. These developments in operating systems have provided
intuitive and efficient interactions between users and its environment. In addition, it ensures
more user-centricity and improved responsiveness of an operating system which improves its
overall performance as well.

4
2.3 Memory Management

Artificial intelligence integration has significantly expanded the ability of conventional operating
systems algorithms to manage complex memory management tasks. As it brought in intelligent
algorithms and predictive approaches, it has changed memory management for current memory
management. These innovations have caused big impact such as:
1.​ AI-Enhanced Paging
In memory management, AI systems utilised Long Short-Term Memory (LTSM) and
Gated Recurrent Units (GRU). LTSM has effectively preserved persistent connections in
memory access patterns, which is beneficial for more accurate estimation of future page
requirements. Equally important, LTSM enables a large number of memory access
patterns to identify any complex structure that a basic heuristic-based approach could
miss and overlook. By taking advantage of both pass access trends, Belady’s MIN
algorithm introduced the most efficient page replacement choices as it is improved by
neural networks. These approaches have reduced page faults by up to 30% opposed to
traditional approaches (S Pandikumar et al., 2025).

2.​ Smart Page Replacement

With the help of support vector machines (SVM) in operating systems, it has further
developed on LRU and FIFO algorithms through deep analysis of its program behavior
characteristics. By taking into consideration, SVM enables page categorisation according
to its potential future access more effectively. This method is able to attain 40% greater
memory use contrasted to conventional approaches (S Pandikumar et al., 2025).
Moreover, combination use of traditional algorithms alongside Deep Q-Learning
enhanced page replacement optimisation through strategy adjustments and analysis from
system behaviour. It frequently updates the mappings based on the obtained outcomes,
maintaining a Q-table that links states (the current memory configuration) to which pages
it will replace.

5
3.​ Intelligent Fragmentation Management
By dynamically optimising memory allocation, Genetic Algorithms and Particle Swarm
Optimisation (PSO) are utilised to minimise memory fragmentation.
●​ PSO algorithms are developed to integrate numerous allocation agents to identify
the most efficient approaches.
●​ Genetic algorithms focus evolutionary theories to produce ideal memory
allocation strategies.
Contrasted to conventional memory allocation, these algorithms can reduce
fragmentation by as much as 50%. It is more recommended that genetic algorithms
expand a group of allocation strategies functioning and grow them through mutation,
crossover and selection operations.

Other than that. Decision tree algorithms also contributed to examining allocation of
memory patterns to balance block sizes for internal fragmentation. Meanwhile, K-means
clustering provides identification of memory block patterns and control reorganisation
operations, by grouping identical allocation requests together.

4.​ Memory Compaction Optimisation

A3C which is also known as Asynchronous Advantage Actor-Critic are essential to
optimise timing of memory compaction. This approach focuses on predicting the optimal
compaction schedules to maximise a reward function according to its system reliability
measures by utilising system state information as input features. Moreover, A3C has the
capabilities to learn from numerous system instances and strengthens the effectiveness of
its learned policies, consequently making it successful.

6
2.4 Energy efficiency

After implementing an artificial intelligence approach such as Long Short-Term Memory

(LTSM) in energy-aware scheduling, it enables estimation of workloads and minimises power
consumption. By doing this, this approach provides proactive power management through
learning from the past workload patterns. With the goal to reduce energy consumption during
idle times, AI-driven Dynamic Voltage and Frequency Scaling (DVFS) continuously modifies
voltage and frequency in response to workload. It is further improved with Reinforcement
Learning, which will continuously learn the ideal parameters. These approaches helped boost the
operating efficiency especially in massive data centers frameworks.

2.5 Security and Threat Detection

Most operating systems that incorporate artificial intelligence utilised deep learning models to
identify the information of network packets, and statistically detection of anomalies in order to
keep track of resource utilisation patterns including its behavioural analysis to keep updated on
system call sequences. As it will frequently absorb new data, the system automatically extracts
features that might indicate apparent attack patterns. This AI algorithm is crucial in operating
systems to strengthen its security from external or internal threats.

7
3.0 INTEGRATION OF INTERNET OF THINGS (IOT) DEVICES WITH OPERATING
SYSTEMS

3.1 Opportunities
The benefit, effectiveness, and scalability of connected systems can be greatly improved by the
many opportunities that are created by the integration of Internet of Things (IoT) technologies
with operating systems. These potentials are especially beneficial for supporting the creation of
autonomous, intelligent, and interoperable devices in a variety of industries. Vital opportunities
consist of:
1.​ Real-Time Data Processing
The ability to handle data in real time has been significantly enhanced by the integration
of IoT devices with operating systems. In circumstances where data must be handled
under strict time limitations in order to ensure efficiency, functionality or safety in
real-time systems are crucial.

Time-sensitive tasks are managed and predictable behaviour is ensured by modern

operating systems, especially real-time operating systems (RTOS) like FreeRTOS,
Zephyr or RTLinux. These operating systems use priority-based preemptive algorithms to
schedule processes, reducing latency and enabling consistent response times which are
essential in:
a.​ Industrial automation:
Machines that have sensors keeping track of factors like pressure, vibration and
temperature. The operating system must interpret this information and initiate
control measures, such as turning off a motor in milliseconds, to avoid
malfunctions.
b.​ Healthcare monitoring:
Patient vitals like heart rate and blood sugar levels are monitored by wearable
medical equipment. In order to notify medical professionals of irregularities and
possibly save lives, the operating system must process sensor information in real
time.

8
 c.​ Smart cities:
Real-time data can be utilised by IoT-enabled traffic management systems to
modify traffic signals, reducing congestion and speeding up emergency response
times.
​ Operating systems that help with IoT real-time data processing also monitor:
a.​ Interrupt handling: Responding immediately to signals from sensors.
b.​ Concurrency: Executing several tasks, including data collection, filtering and
actuation more effectively.
c.​ Memory and power optimization: Essential for battery-operated equipment that
must run constantly.

2.​ Edge Computing

Instead of depending entirely on cloud infrastructure, edge computing is a new paradigm
that processes data at or close to the point of generation, usually on the IoT device itself.
Faster data processing, increased dependability, and improved privacy are made possible
by the integration of operating systems with edge-based IoT devices.

In order to manage these decentralised processes, edge computing-specific operating

systems like Ubuntu Core, AWS IoT Greengrass and lightweight RTOS platforms are
essential. Systems can lower latency, offload cloud processing and continue to function
even with irregular connectivity by carrying out operations like data filtering, anomaly
detection and local decision-making on the edge. Examples of use cases are as such:
a.​ Smart agriculture:
Sensors installed in the field can instantly modify watering by using locally
processed data on soil moisture.
b.​ Autonomous vehicles:
It would be inefficient to rely on cloud computing for critical choices like
avoiding obstacles which are made on-board.

9
 c.​ Remote infrastructure:
For locations with unstable networks, such as offshore rigs, edge devices with
real-time operating systems guarantee constant surveillance and management.
The OS in such systems must support:
a.​ Task execution with minimal latency.
b.​ Containerization or sandboxing like Docker or WASM at the edge.
c.​ Local synchronisation and caching of data.

3.​ Scalability
From a few smart home devices to millions in a smart city grid, IoT ecosystems can take
many different forms. As a result, scalable systems that can easily handle this fluctuation
must be supported by the operating system. Current IoT-focused operating systems such
as Mbed OS and RIOT are designed to enable:
a.​ Dynamic device discovery.
b.​ Automated configuration.
c.​ Resource allocation across a growing number of nodes.
The capacity to manage various hardware architectures, applications, and communication
protocols such as Bluetooth, Zigbee, LoRa, and Wi-Fi without requiring major OS
changes is another aspect of scalability. Examples of scenarios in which it is relevant
include:
a.​ Smart factories using predictive maintenance over thousands of sensors and
machinery.
b.​ Urban infrastructure where a common operating system platform is required for
systems like waste management, lighting and surveillance to scale and collaborate
together.
In order to handle this, operating systems need to provide multi-tenant environments in
fog/edge nodes, strong network stacks, and modular kernel designs. The image below
demonstrates the graph visualization in the context of digital transformation, which
includes IoT integration. It displays the growing importance and interconnectedness of
IoT, AI and OS development in recent academic and industrial research.

10
 Graph visualization of the evolution of digital transformation themes

4.​ Interoperability
The capacity of various IoT systems and devices to cooperate and share data without any
problems, despite communication protocol, manufacturer, or operating system, is known
as interoperability. This interoperability is made possible by operating systems that
support open standards, including as MQTT, CoAP, OPC UA, and LwM2M, which allow
devices to communicate using a "common language." For instance, a Zigbee-based
thermostat can communicate with a Wi-Fi-based smart plug and a cellular-based security
camera if the operating system is built to support several stacks.

OS differences make interoperability more difficult, thus these differences are abstracted
via middleware layers like IoTivity and AllJoyn or platforms like Contiki-NG and
TinyOS. Some of the advantages include:
a.​ Easier system integration in mixed environments.
b.​ Faster deployment cycles.
c.​ Reduced vendor lock-in.
​

11
​ Listed below are a couple of challenges the OS needs to overcome:
a.​ Protocol translation.
b.​ Resource harmonization like common memory or addressing models.
c.​ Semantic data interpretation.

3.2 Challenges of integrating IoT in Operating Systems

There are several operational and technological difficulties with integrating Internet of Things
(IoT) technology with operating systems. The sheer level of complexity, resource limitations and
variety of IoT ecosystems are the root causes of these challenges. Key challenges include:
1.​ Security and Privacy
IoT devices usually have minimal computational capability, which limits the use of
standard security measures, and they are frequently placed in open, uncertain areas with
minor physical security. The operating system is essential for protecting data and system
integrity through factors like:
a.​ Secure boot and firmware integrity checks.
b.​ Access control and user authentication.
c.​ Data encryption at rest and in transit.
However, due to hardware limitations many IoT-related lightweight operating systems
such as TinyOS or RIOT find it difficult to integrate thorough security layers. For
example, There could be privacy violations if a smart home camera gets hacked since
OS-level encryption fails to be in effect.

2.​ Resource Constraints

A number of Internet of Things (IoT) devices are limited in amounts of memory,
processing power, and energy. To prevent going over the device's capabilities, integrating
a fully functional operating system or even a lightweight real-time operating system
(RTOS) requires careful optimization. The following are common limitations for IoT
devices:
a.​ Processing power.
b.​ Memory which is often less than 256KB RAM.
c.​ Energy where it is battery-dependent.

12
 Operating systems need to be incredibly efficient and lightweight while still having the
ability to multitask, control sensors and connect to networks. Performance and feature
support are frequently traded off for developers.

3.​ Standardization
There are many different suppliers, protocols and operating systems in the fragmented
IoT ecosystem. Interoperability, maintenance and integration are made more difficult by
the lack of one common standard. Scalability is restricted by the need for OS developers
to either support various stacks or include platform-specific customizations. For instance,
a smart parking sensor system that uses exclusive operating system protocols may not be
able to connect to the larger traffic control platform of a city.

This diagram compares an AI-driven OS Agent interface that analyses the user's
objective and calls the required services with a standard user who interacts directly with
system components. It is ideal in demonstrating how OS interfaces have changed as we
progress towards intelligent and autonomous systems which are essential to edge
computing and the Internet of Things (IoT).

13
 4.​ Maintenance and Updates
IoT devices usually operate in isolated or difficult to reach locations. It becomes
logistically difficult to roll out OS patches or upgrades, particularly when taking these
factors into consideration:
a.​ Over-the-air (OTA) update capabilities.
b.​ Update security to ensure only authorized firmware is installed.
c.​ Minimal downtime during updates.

3.3 Future Trends in operating systems

Operating system integration is expected to incorporate a number of significant patterns as IoT

develops further, reflecting the increasing complexity, intelligence and widespread use of
connected devices. The next generation of IoT solutions is greatly influenced by these trends,
which are powered by developments in network, hardware, and software technology. Some of
the most necessary emerging future trends are as follows:
1.​ AI-Enabled Operating Systems
In order to process data locally and make intelligent decisions, embedded AI/ML models
will be a more common feature of future IoT operating systems. These OSs powered by
AI can:
a.​ Predict device failures.
b.​ Adapt resource allocation dynamically.
c.​ Enhance cybersecurity through anomaly detection.
​ Example:
A smart grid node's AI-enabled operating system may identify abnormal power
consumption trends, raising the possibility of electrical problems or energy theft. As seen
in the OS Agent Interface picture below, the article titled "Enhancing Operating System
Performance with AI: Optimized Scheduling and Resource Management" supports this
idea by suggesting integrating AI agents into the OS architecture.

14
 Relation between Operating System and AI
2.​ Integration with Blockchain
Intelligent and safe systems are being made possible by the convergence of blockchain,
machine learning and the Internet of Things (IoT). Through integrating blockchain
technology into IoT operating systems, decentralized data management has become
possible, improving security and reliability. Data integrity, authentication and safe
transactions between IoT devices are issues that this integration solves. Some of the key
benefits are:
a.​ Decentralized Security: Data created by IoT devices is impenetrable thanks to
blockchain's immutable ledger, which increases trust in data integrity.
b.​ Smart Contracts: IoT devices can operate independently thanks to the automated
execution of predefined actions when certain circumstances are fulfilled.
c.​ Enhanced Privacy: Blockchain allows for private and secure data transactions,
giving users more control over their data.
Example:
Integrating blockchain with IoT in smart cities can enhance urban planning and services
by securing data from multiple sensors, such as traffic and pollution, and allowing
transparent sharing among stakeholders.

15
 3.​ Context-Aware and Adaptive Systems
In order to allow devices to modify their operations in response to human behaviour and
environmental conditions, future IoT operating systems are developing to be
context-aware. These systems can make decisions in real time by utilizing AI and ML
which improves user experience and efficiency. Some of the key features include:
a.​ Sensor Fusion: Integrating information from several sensors to obtain a thorough
understanding of the situation.
b.​ Adaptive Learning: Systems gradually improve performance by learning from
user interactions and environmental changes.
c.​ Predictive Analytics: Dynamically adapting operations by anticipating user needs
and system requirements.
​ Example:
Context-aware systems in smart homes can modify security, heating and lighting settings
according to individuals habits and preferences which improves comfort and reduces
consumption of energy.

4.​ Quantum-Ready and Post-Classical Support

The requirement for operating systems to be quantum-ready continues to rise as quantum
computing develops, particularly in the Internet of Things area. In order to protect against
potential quantum threats, quantum-ready operating systems are developed to
communicate with quantum processors and integrate post-quantum cryptography
methods. Among the crucial factors to take into account are:
a.​ Post-Quantum Cryptography: To secure IoT connections, cryptographic methods
that are immune to quantum attacks are being implemented.
b.​ Hybrid Architectures: Creating systems that operate effectively with resources
from both traditional and quantum computing.
c.​ Future-Proofing: Ensuring that when quantum capabilities become widely used,
existing IoT deployments will be updated to incorporate them.

16
​ Example:
Adopting operating systems that are quantum-ready ensures long-term security and
compatibility with new quantum technologies in critical areas like energy and the finance
industry.

4.0 CASE STUDIES AND EXAMPLES

17
5.0 CONCLUSION
The integration of Artificial Intelligence and machine learning as well as Internet of Things into
operating systems has significantly influenced the overall functioning of modern operating
systems which is also known as smart OS. These applications have integrated new features that
enable users to generate more complex requests.
Despite all those positive outlook, there are still significant obstacles to overcome in
generating a smart operating system. It will need to go through a deep consideration to ensure an
effective AI and IoT operations while protecting it against vulnerabilities in computer systems
and unauthorized access. Furthermore, depth in research also needs to be conducted for
simplified systems that can function in limited settings.
​ As years go by, designers of future operating systems are likely to come up with new
advances, therefore they should carefully assess the information to foresee the necessity of a
system extension in order to fulfill user demands while designing operating systems that are
intended to enable such a facility.
​ As data-driven workloads could increase across multiple computing platforms, smart OS
designs are essential to grow the ability handling specialised hardware and enhance data flow
especially in AI and IoT applications.

18
6.0 REFERENCES

1.​ Arif, H., & Khan, M. H. (2024). Intelligent defense: Redefining OS security with AI.
ResearchGate.
https://www.researchgate.net/publication/389317274_Intelligent_Defense_Redefining_O
S_Security_with_AI
2.​ Barbosa, J. B., Costa, H. G., & de Souza, J. M. (2021). Digital transformation: A
bibliometric review and research agenda. Revista de Administração Mackenzie, 22(1),
1–30.
https://www.scielo.br/j/ram/a/mGpm3mhb5vZ5VLPbmmfYBwt/?format=pdf&lang=en
3.​ Gubbi, J., Buyya, R., Marusic, S., & Palaniswami, M. (2013). Internet of Things (IoT): A
vision, architectural elements, and future directions. Future Generation Computer
Systems, 29(7), 1645–1660.
https://citeseerx.ist.psu.edu/document?repid=rep1&type=pdf&doi=6ddc20248f15e7326d
1b48f455152eee98d993d6
4.​ Kamble, T., & Bhat, G. (2022). Integration of machine learning into operating systems:
A survey. ResearchGate.
https://www.researchgate.net/publication/362518003_Integration_of_Machine_Learning_
into_Operating_Systems_A_Survey
5.​ Rao, A. R., & Srinivas, S. (2020). The Integration of IoT, Machine Learning, and
Blockchain: A Convergence for Secure and Intelligent Systems. ResearchGate.
https://www.researchgate.net/publication/389498830_The_Integration_of_IoT_Machine_
Learning_and_Blockchain_A_Convergence_for_Secure_and_Intelligent_Systems
6.​ Shaikh, M. A. (2023). Post-Quantum Cryptography and Operating System Adaptability
for Secure IoT Systems. IAEME Publications.
https://iaeme.com/MasterAdmin/Journal_uploads/IJAIML/VOLUME_4_ISSUE_1/IJAI
ML_04_01_013.pdf
7.​ Shankar, V., & Prakash, S. (2024). Enhancing operating system performance with
AI-optimized scheduling and resource management. ResearchGate.
https://www.researchgate.net/publication/391880725_Enhancing_Operating_System_Per
formance_with_AI_Optimized_Scheduling_and_Resource_Management
8.​ Sharma, A., & Sinha, P. (2024). Advancing Specialized Operating Systems for the
Internet of Things: Insights and Future Trends. ResearchGate.
https://www.researchgate.net/publication/389098506_Advancing_Specialized_Operating
_Systems_for_the_Internet_of_Things_Insights_and_Future_Trends
9.​

19
​

20