Computer Network Project CO3119 CC01 231
Computer Network Project CO3119 CC01 231
UNIVERSITY OF TECHNOLOGY
FACULTY OF COMPUTER SCIENCE AND ENGINEERING
Project
Contents
1 Task 1: Research the IoT infrastructure architecture framework in the world. 3
1.1 The simplified form . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.2 The meaning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.3 The alternative form . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
4 Task 4: Research and propose standards for security and privacy for IoT ap-
plications. 13
4.1 Existing standard: "European EN 303 645" . . . . . . . . . . . . . . . . . . . . . 13
4.2 Proposal standard, plus other aspects concerning the "Safe Railway Framework" 14
The architecture of the Internet of Things (IoT) can be depicted as a stratified composition
comprising of the subsequent strata: device, network, session, application, business, manage-
ment, and security. While the actual implementation for specific projects may vary, this is the
backbone/template of an IoT architecture.
The essential layers, roles, and functionalities are reorganized to further emphasize the secu-
rity aspect of the overall system. On the other hand, it redefines layers to include cutting-edge
technology (AI, machine learning, robust collecting devices with processing capability, and more).
In other words, this can be interpreted as a modernized representation of an IoT Architecture.
The applications and analytics element is responsible for the processing and presentation of
data obtained through the Internet of Things (IoT). This component encompasses various tools
such as analytics, artificial intelligence (AI), machine learning, and visualization capabilities.
Integration component. This is the component that ensures that the applications, tools, se-
curity and infrastructure integrate effectively with existing companywide ERP and other man-
agement systems.
The security and management aspect is an essential component of the Internet of Things. To
ensure the security of the entire system, it is imperative to safeguard the physical components
using firmware and embedded security providers. These providers include both traditional secu-
rity providers with the capability to support the Internet of Things and specialized IoT security
providers.
The infrastructure component includes various physical elements, specifically intelligent sen-
sors for data collection and actuators for regulating the surrounding conditions. Furthermore, it
includes the network that accommodates these sensors and actuators, commonly but not exclu-
sively manifested as a wireless network such as Wi-Fi, Bluetooth, 4G, or 5G. Other emerging
wireless alternatives include Long Range WAN and low-power WAN.
The Ho Chi Minh City National University (HCM VNU) Office’s research team has success-
fully implemented a project entitled "Research and Manufacture of IoT DataLogger for smart
traffic light systems." The project aims to build a smart traffic management solution by reno-
vating the existing traffic infrastructure. The team has demonstrated proficiency in microchip
technology and PLC-IoT devices. The manufactured Datalogger can be seamlessly integrated
into homegrown smart traffic processing and operating systems, as well as foreign existing coun-
terparts.
The PLC-IoT Datalogger, as a comprehensive set, is designed to integrate with the current
traffic lights. It serves as a terminal device that provides parameters to the intelligent traffic
management system. This system effectively processes and automatically adjusts the light signal
time based on prevailing conditions. The coordination of remote traffic controllers is facilitated
through a WIFI network. This solution offers the advantage of utilizing and upgrading traffic
light poles that employ outdated technology. Consequently, it establishes a foundation for the
development and enhancement of Ho Chi Minh City’s traffic systems from two perspectives:
providing information to traffic participants and enabling efficient management.
Essentially, the project was successfully deployed and tested at the HCM University Campus
in Di An, Binh Duong, and was further adopted by the Municipal Administration office in Ho
Chi Minh City for more research and implementation in 2020.
• There will be digitally stored data reports of the lighting operation to provide insights and
improvement.
• Automatically adjust/control the light system for convenience and reduce electricity usage.
Hình 4: Vilight Smart Lightning system takes the form of an electricity box
The project was successfully adopted, deployed, and transferred to the HCMC Department of
Transportation for future implementation throughout the HCMC area. On the other hand, the
cooperation products often prove their effectiveness and have been trusted by multiple regions
in Vietnam to install tailored versions of the system.
The system consists of an IoT Gateway device, an IoT Node device responsible for monitoring
the air statistics, and a suite of transmission security encryption hardware, as well as software
deployed on the cloud server. Additionally, it includes application software for air quality mon-
itoring and traffic serving, as well as the ability to store and process data collected from IoT
devices. The IoT Gateway devices offer support for real-time applications and enable protocols
for remote control and access, thereby facilitating direct communication between applications
running on the Gateway and the cloud server (Cloud Server). Furthermore, the IoT Node device
used for air environment monitoring employs various sensors to track a wide range of indicators
such as SO2, NO2, PM2.5, PM10, temperature, and humidity.
The devices are autonomously operated by software executing on a cloud server platform,
designed to automatically refresh data about images, and indices for monitoring air quality, as
well as incorporate notification capabilities to indicate status. This aids in evaluating pollution
and the environment within the designated region. The software incorporates various modules
that facilitate the management of network-connected devices, including but not limited to Gate-
ways responsible for security management, user administration, the establishment of routes and
connections, as well as the registration of new devices...
In the end, the research produced 5 IoT Gateway sets and 10 IoT Node sets plus 6 hardware
encryption devices. Most if not all of these were approved by the authority. On a side note,
the research also proposed an encryption protocol, which was patented. The system proved to
be highly effective, especially during the COVID period, significantly aiding the Saigon Hi-tech
Park in monitoring the whole area environment.
3.1 Context
In the year 2022, the Vietnam Railway Authority (VNRA) reported 213 railroad accidents, and
166 was the number of casualties. At a glance, the number is not significant, however, this only
counts traffic accidents which usually happen between the railroad and other transportation. I.e.
the pedestrian doesn’t follow the traffic signal and then proceeds to get struck by the incoming
train. However, whenever there is an accident, the train (and the surrounding traffic) often be
halted for further investigation. This would cause hours, sometimes even half a day in order to
continue the traffic. At the same time, nearly every train on the entire line got halted, thus
delaying not only the train that caused the accident but also cost others time delay. And in
business, any delay means the loss of (a high amount of) money. This doesn’t count other incidents
such as livestock/obstructions on the railway, the degradation of the line, and damaged tracks
due to natural causes, ... Every single incident, albeit small or major, involves human lives or
does not often cause disorder and time delay to either the local schedule or on the entire line
from North to South Vietnam. In order words, time losses, and further down the line, money
losses are prominent.
The current train monitoring system does have multiple proposed projects, such as an au-
tomatic warning system, by-the-railroad visual surveillance, and a systematic train operation
management system, ... However, there still are the needs of humans to fully operate the whole
system. Staff are needed to stay on guard at each major checkpoint along the railroad, or there
are countless "survey teams" to routinely patrol their assigned area to find any abnormalities.
The proposal aims to mitigate the workforce needed to stay on guard/stay on patrol while pro-
viding the same, if not more precise surveillance of the entire railroad system, either at traffic
junctions or carefully observing any degradation of the system.
The core functionalities of the system are to provide real-time surveillance, and remote control
utilities (lights, barriers, ...). The surveillance system will be divided into 2 separate sections:
one for direct control (via a command center) and the other one for public service: via a tool
(web interface for example), the public gains access to the necessary data of the railway: trains’
locations, approximate time arrival, stations’ status, ... It also allows user to contribute to the
system: people can provide extra information about a situation for further insights.
The system is as follows:
• Core: This is the utter most important part of the system, which plays a role in finalizing
the collected data and is represented in 2 ways: an "internal" command center where direct
control can be made towards vital components of the system.
• Backbone: this serves as the link between the Edge and the Core. In summary: it is the
communicator within the system.
• Edge: numerous devices come in the form of imagery collectors to survey key areas; pressure
sensors to monitor the railway conditions at critical points on the system (i.e. Weak bridges,
potential landslide areas, ...); those devices and the like can be considered "Passive data
collectors"; there shall be drones or small robots that serve as movable surveillance actors
to provide rapid responding observation to any unplanned situation (accidents, railroad
system damages, ...); the last "edge" data collector comes in the form of an application,
where local people can report abnormalities to the system, thus trigger a system response
to counteract the incident (users report a major incident at location A, the system will
dispatch appropriate countermeasures to this location).
• Security: any surveillance system is a double-edged sword, the gathered information can
bring useful applications and also potentially attract dangerous exploitation. For example
in the current context: normally the "exact" or real-time location of certain trains can’t
be easily accessed. However, with a ready access system, anyone can have the pinpoint
location of any train in the system, thus opening numerous vulnerabilities such as robbing,
kidnapping, and terrorizing, ... Therefore, the system should be made secured at any point
of interaction, either "internally" (within the system communication, data transfer) or
"externally" (the data can be accessed from the outside world.
4.2 Proposal standard, plus other aspects concerning the "Safe Rail-
way Framework"
Existing standards, such as the one mentioned above, cover a wide range of applications and
implementations. This proposal extends one clause of the standard, in order to prevent an un-
expected angle of attack: internally.
Clause number 7 states: Ensure software integrity. This means the system must have a me-
chanic to ensure only authorized actions are allowed and detect and prevent illegal ones. However,
there is still one actor that plays the uttermost vital role in operating any system: the human.
Humans are not software, and should be mentioned in another category, such as human resource,
or operator training, ... But, we do believe that the software (development) should seriously
consider the following scenario: the operators themselves are the ones who sabotage the system
in any form: data leak, create backdoor, ...
History proved that whenever there is a centralized power (i.e. a top brass of an organization)
there can be misuse of power. One common way to combat this is to decentralize the power, and
for major actions: it need approval from multiple actors in order to proceed. Many organization
works well with this mechanic: the United Nation Security Council, nuclear missile launch control
needs multiple keys from different personnel to approve the launch, ...
Things to do: for every major action (delete the database, change vital passwords, add new
administrators, ...), the software must have a mechanic to gather the approval from appropriate
personnel to be done. Either by designing a specific role-based system, or outright preventing a
single person from holding too much power. On the other hand, the operator’s training should
also account for this matter, thus providing suitable knowledge and skills.
Specifically for the "Safe Railway Network": the command application that provides the
ability to dispatch responding countermeasures to any crises: it should limit the number of
"units" one operator can deploy at the same time. For example: hypothetically, users informed
that there was an accident at Bien Hoa. The operator (without any form of limiter) can dispatch
multiple if not all the drones available in the area to the accident site, thus exhausting the entire
(local) system and preventing any flexible response if there were another crisis in the same area.
Traditionally, cameras are statically installed, therefore, one reliable way to have an entire
surveillance system is to have multiple cameras that are connected by wire into a "central"
storage system. From this storage, the demanded data shall be transferred to somewhere else,
like a display system to review the footage, or some processing unit to make analytics operations.
"Reliable" but not secured or immune to the weakness of centralized systems: if any major node
in a system (in this case let’s say the central hub) is disabled, the entire system will consequently
be disabled. Furthermore, basic systems often leave out the processing operations to specific
subsystems, say there will be a data processing subsystem. This eliminates the scalability of the
system: whenever there is too much input information or too many processing requests, the data
processing subsystem can be easily overrun and thus cause instability.