Effective & Efficient Access Control in Smart Farms: Opportunities,
Challenges & Potential Approaches
Ghadeer I. Yassin and Lakshmish Ramaswamy
Department of Computer Science, University of Georgia, Athens, GA 30602, U.S.A.
Keywords:
Access Controls, Smart Farming, Precision Agriculture, IoT.
Abstract:
The Internet of Things technologies has revolutionized the sector of farming and agriculture. It also helped
it to face the current environmental and societal challenges. IoT technologies are able to assist the farming
sector in many different applications including reducing wasted resources, real time monitoring of crops,
monitoring environment conditions, precision agriculture, farm data analytics and improving crops quality,
while decreasing the number of workers needed to complete farm related tasks. However, the nature of the
Smart farms are diverse in terms of the number, type and location of the installed smart devices, the variety
of the collected data, the number and type of workers who help in the farm and have access to farm related
data and equipment. At the same time, Farmers are very protective of their own data and sometimes refrain
from incorporating smart farming technologies to insure the safety of their data. Therefore, in this paper we
outline the security challenges in smart framing settings and explain the need for multi-user, multi- device
aware access controls in smart farms. We highlight different possible security scenarios that challenge the
adoption of IoT solutions in smart farms and discuss possible solutions.
1 INTRODUCTION
The world population is expected to increase to 11.2
billion by the end of the century according to the
United Nations (Roser, 2013). With the rapid growth
in world population, the food consumption worldwide
grows rapidly as well, it is estimated that food pro-
duction must increase by 70 % to feed this popula-
tion (Bruinsma et al., 2009). As a consequence farm-
ers are facing a growing demand to produce more
food regardless of the struggle they endure because
of the climate change, the scarcity of water, the ex-
treme weather conditions and more (aer, 2021). In
order to increase the yields of their farms to meet that
demand, farmers are turning into smart farming.
“Smart farming” combines traditional agricultural
methodologies with actuators, sensors, Internet of
Things (IoT), Information and communication tech-
nology (ICT) and drones to provide the farmers with
precise and accurate information that help in optimiz-
ing farming tasks with less labor force, increasing
crop yields, improving crops’ quality, decreasing the
waste in resources such as water, fertilizers and herbi-
cides while increasing farm profitability. Smart farm-
ing follows a cycle that includes observing data from
the surrounding environment then ensuring it follows
a set of predefined rules, identifies any deficiencies
and informs the users about them to take the appro-
priate action if needed.
Smart farms are considered a multi-device envi-
ronment in the sense that multiple smart devices can
be found scattered across the farm such as Drones,
Animal’s wearable devices, Soil moisture sensors and
more. It is also considered a multi-users environment
with a significant number of human workers cooper-
ating to monitor the environment data and make in-
formed decisions. In addition to that, There is hetero-
geneous data related to farms originating from multi-
ple sources such as sensor data, satellite imagery and
farm historical data. Each data type has a validity pe-
riod and a unique value associated to it based on the
information derived from this data and how this in-
formation can be used. This derived information can
have great impact on the livestock health and wellbe-
ing, the farm value and the owners themselves. There-
fore, effective access control mechanisms are needed
in smart farms to ensure the security of its data con-
sidering its diverse nature.
In this paper, we highlight the importance of effec-
tive access control mechanisms in smart farms. We
compare and contrast smart farm nature with other
smart domains as well as the security challenges that
Yassin, G. and Ramaswamy, L.
Effective Efficient Access Control in Smart Farms: Opportunities, Challenges Potential Approaches.
DOI: 10.5220/0010873000003120
In Proceedings of the 8th International Conference on Information Systems Security and Privacy (ICISSP 2022), pages 445-452
ISBN: 978-989-758-553-1; ISSN: 2184-4356
Copyright
c
2022 by SCITEPRESS – Science and Technology Publications, Lda. All rights reserved
445
face smart farms. We discuss which access con-
trol mechanisms could be suitable for the smart farm
system and discuss some potential solutions and ap-
proaches.
2 BACKGROUND INFORMATION
2.1 Smart Farm
A Smart Farm in its simplest form is a combination of
computing devices, machine learning (ML), artificial
intelligence (AI) and humans integrated into an infor-
mation driven system that is capable, knowledgeable
and a real time-decision maker, therefore, a Smart
System”. Smart farm’s sensors and equipment have
a predefined function related to the farm that can be
as simple as collecting temperature data or as com-
plicated as harvesting ripened crops and leaving un-
ripened ones behind.
Figure 1 Shows the different entities in smart farms
and the interaction that takes place among them. Dif-
ferent sensors collect huge amounts of farm-related
data and are connected to gateway devices that per-
form preprocessing steps such as filtering and analyz-
ing the collected data and transfer only the necessary
data to the cloud services where it can be further pro-
cessed and stored then presented to the user through
suitable interfaces.
2.2 Smart Systems
IoT technology and its applications are currently used
in multiple smart systems such as smart homes that
contain multiple IoT devices connected together and
to the internet through a communication protocol and
are controlled, accessed and monitored remotely us-
ing a mobile application or voice commands issued to
a smart assistant. In addition, smart health care refers
to completely autonomous and connected healthcare
solutions provided for patients along with feedback
from doctors. Smart Cities, on the other hand, refers
to any city that utilizes modern technologies to con-
vert conventional city’s entities into an autonomous
entity and consists of a collection of smart services
such as city lighting, emergency services, traffic man-
agement, water management etc.
2.3 Data Sources & Analytics
Smart farms are majorly driven by data and its ana-
lytics. Sensors detect changes and collect data about
their environment. While actuators introduce change
Figure 1: Different Smart Farm Entities’ Interaction.
to its environment based on the sensor’s collected
data. This is a real time process that is mostly auto-
mated. There are major domains in smart farms that
use the heterogeneous data collected by sensors and
apply them in different applications that are capable
of providing the farmers with the most accurate an-
alytics to help them improve farm productivity and
decrease human interventions as shown in Figure 2.
Precision Farming: Achieved through multiple
monitoring and controlling applications including
continuous weather monitoring, Crop’s health mon-
itoring and Soil analysis which is one of the most im-
portant applications of precision farming that helps to
easily decide the best type of crops to plant in the soil,
the best timing to start sowing etc.
Precision Livestock: Achieved by providing live-
stock with different monitoring sensors and veterinary
wearable devices to continuously detect, analyze and
transmit data related to the animal’s health and loca-
tion.
Smart Greenhouse: Enables planting different types
of plants anytime of the year regardless of the weather
conditions by precisely monitoring and controlling its
climate.
2.4 Access Control Policies
Access control refers to processing every request
to the system’s data and resources, then deciding
whether it should be denied or granted. Access con-
trols identify users through verifying their login cre-
dentials and then grant them the appropriate access
level that is associated with their authenticated cre-
dentials.
There are different access control policies that can
be used in smart systems based on its requirements
including the “RBAC” model that manages the access
to the different resources based on hierarchical rights
and permissions assigned to various roles. ABAC”
model, on the other hand, shifts from the list of roles
to a list of attributes and contextual factors by eval-
uating requests against the requester’s attributes, the
required action, and the context. “MAC” model lim-
ICISSP 2022 - 8th International Conference on Information Systems Security and Privacy
446
Figure 2: Main Smart Farms’ Domains, Applications, Data & Data sources.
its access to system resources based on its sensitivity
and the user’s authorization to access a resource with
that specified sensitivity. “TBAC” enables placing re-
strictions on the access to resources and its data on the
basis of particular time of the day or particular days
of a week. Finally, “LBAC” restricts the access to re-
sources based on the location of the user so that the
users can access the resource only if they are physi-
cally present in the predefined location.
3 MOTIVATION
Many farming systems use insufficient access con-
trols and new smart farming related projects do not
consider using them. (De Araujo Zanella et al., 2020).
On the other hand, smart farms have a unique nature
that requires specialized access control solutions.
3.1 Comparing Smart Farms with
Other Smart Systems
There is a wide range of differences and similari-
ties between Smart farms and other Smart systems
as shown in Table 1. These differences indicate
that security-related countermeasures applied to other
Smart systems are not suitable for Smart Farms.
Stakeholders & Security Awareness: In smart
farms, The stakeholders range is very wide and the
system is very dynamic as the owners usually hire
different workers based on the farming season or the
scalability of their farm. The stakeholders have dif-
ferent levels of security awareness. Owners are more
aware about security since their main goal is to pro-
tect their data as it directly affects their profitability
if it were to be stolen or leaked to a rival. However,
other stakeholders usually have much less awareness
about security and might not be concerned about it
since it does not directly affect their profit.
Similar to Smart Farms, The stakeholders range in
Smart Cities and Smart Healthcare is wide. How-
ever, they are usually provided with technical security
training on how to protect their data. Those Systems
are less dynamic than smart farms as the stakeholders
are hired by the government and their change is much
less frequent when compared to smart farms.
Data Security: Smart Farming security systems are
in their infancy. Very little efforts are being taken to
ensure the security of farm data. In addition, farms are
mostly self-financed by their owners who do not have
enough finance to implement enough security mea-
sures. Smart Homes are also self-financed systems
but their security measures might not be as costly as
Smart Farms and therefore its owner can easily sup-
port it. Other systems such as smart healthcare and
smart cities are supported by many organizations and
a lot of investment is being offered to them.
3.2 Data Validity
Farm data originates from various sources such as
sensor’s data, satellite data or historical data. Each
type of the data has its own validity period; when-
ever passed, the data is deemed useless. The validity
periods range from long validity such as the yearly
planted crops in the field to seasonal validity such
as season’s weather data to short term validity such
as soil moisture data that are acted upon immedi-
ately. Therefore, the data should be stored and pro-
tected during its validity period and erased otherwise.
For example, whenever a livestock animal is sold or
sent to a slaughterhouse, its data becomes unneces-
sary to store and therefore should be deleted immedi-
Effective Efficient Access Control in Smart Farms: Opportunities, Challenges Potential Approaches
447
ately while soil data should be deleted once it is ana-
lyzed and provided to the farmer etc.
3.3 Farm Ownership
Farms can be owned, rented and operated by differ-
ent types of owners (Bigelow et al., 2016), (Bigelow,
2019) such as families, non-operator landlords, oper-
ator landlords, cooperatives etc.
Family Farms: In family farms, the decision mak-
ers are typically the owners of the equity and the as-
sets who usually assign predefined tasks to the other
family members. However, a lot of conflicts happen
between the family members. Therefore, the family
holds a broad meeting to discuss members’ roles and
responsibilities and negotiate their demands. There-
fore, it would be beneficial to propose access control
mechanisms that support the negotiations of roles and
demands and ensure they are enforced.
Rental Farms: The operator landlords of rented
farms usually have conflicts with the tenants due to
their involvement in farm decisions as both of them
combine their own resources to produce the required
farm commodities. Typically, they negotiate their
terms and create a farm lease contract describing their
agreement that covers the amount of fertilizers that
will be used, the time when a specific equipment will
be used, the time when the field will be harvested etc.
However, the conflicts happen due to the unawareness
of the landlord about certain circumstances that affect
the execution of the agreement. Therefore, granting
access of the farm data to landlords based on a prede-
fined role and agreement will keep them more aware
of circumstances and thus eliminate such conflicts.
Partnership Farms: Partners and cooperatives have
conflicts over the share of each partner; some farm
assets such as livestock or crops might belong to
one partner only. Those assets and their related data
shouldn’t be accessible by the other partners as they
are out of the scope of the partnership. Therefore, ac-
cess controls should be flexible enough to suit assets’
data accessibility based on such cases.
3.4 Participants’ Profiles
Figure 3 assumes the different in-farm and outside-
farm access attempts to different smart devices and
sensors. The figure shows 8 different individuals in-
teracting with the smart farm based on their assigned
tasks. Both workers A and B are responsible for field
sowing and irrigation. Therefore, they have access to
plant data and soil data. Worker C is the only worker
permitted to monitor the autonomous tractor. There-
fore, its accessibility is granted only to him and he is
restricted from accessing any other equipment or data
in the farm system. On the other hand, Worker D is
granted accessibility to the drone when he is present
inside the field and is denied accessibility otherwise.
Other individuals such as the veterinarian and worker
G are involved in occasional animal health monitoring
and therefore do not need continuous access to the an-
imal data.
Temporary Labor: It is common to hire temporary
labor at the time of harvesting or sowing. The main
problem that faces the farm owner with temporary la-
bor are the trust issues. Those workers are mostly
coming to work in the farm for the first time and are
not trusted as other permanent workers and they might
abuse farm related data by uncovering them to a ri-
val or by disrupting the system by modifying sensor’s
parameters causing it to perform unnecessary actions
that leads to profit losses. The accessibility granted to
temporary workers should be limited and end by the
time their hiring period ends to ensure they will not
be able to abuse the system in any form once they are
no longer hired.
Labor Working on Multiple Farms: Some workers
work in two or more different farms simultaneously.
Through those workers, a farm rivalry or a competi-
tor might gain access or insight to the farm data and
could use it to develop competing crops to increase
their profitability. The data about the owners’ farm
practices can also be used against them for regula-
tory enforcement purposes. Therefore, farm owners
should decide which and when their data is accessed
by workers. For example, owners may choose to auto-
matically notify the worker to turn on/off an irrigation
system without receiving soil moisture data from soil
sensors. They also may want to limit workers from
accessing data outside the daily working hours or to
prevent them from receiving full farm’s history data.
Other Participants: The range of farm participants
is not limited only to farm workers and owners but it
also includes government officials, technicians, aca-
demic researchers, veterinarians etc. Therefore, The
farm owners should be able to control accessibility to
their own resources based on what they believe is pro-
tecting its security.
3.5 Livestock
Security of livestock’s data is often neglected as it is
not classified as personal data and hence very little at-
tention is paid to its security. However, Livestock data
is a very precious part of farm data.
Livestock Location Data: Livestock location data is
continuously collected by GPS-powered wearable de-
vices and drones. This data is used to track the ani-
ICISSP 2022 - 8th International Conference on Information Systems Security and Privacy
448
Table 1: Differences and Similarities between Different Smart Systems.
Smart Farms Smart Cities Smart Healthcare Smart Homes
Stakeholders’ Range Wide Wide Wide Small
Security Awareness Lacked High High Medium
System Dynamism High Low Low Non
Devices Power Low High High High
Device location Outdoor Hybrid Indoor Indoor
Maintenance Frequency Low High High Low
Security Measures Weak Strong Strong Medium
Connectivity Unstable Stable Stable Stable
Financed By Self Organizations Organizations Self
mals when it is left to feed in open areas. However,
If adversaries or some untrustworthy workers were in-
correctly permitted access to only one animal location
data, they can plan to steal the animal and compro-
mise the safety of the whole herd.
Livestock Health Data: It is the most crucial part of
the livestock collected data and greatly influences its
market prices. Therefore, purchasers can be granted
access to the animal’s data in order to make an in-
formed decision. General live-stock’s data such as
previous health conditions might be used to negoti-
ate the offered prices with the owner. In addition, ac-
cessibility to the full herd data can lead to inferring
information about herd size and value. Therefore, the
accessibility should be very limited and only tempo-
rary during the purchasing process.
Veterinarians & Assistants: Some veterinarians
work with different farms in the same area and can
easily anticipate health issues breaking out in that area
through monitoring and combining livestock health
data from different farms. They can take advantage of
such knowledge to propagate exaggerated news about
infectious spread and influence the livestock market
price and consequently make profit themselves by in-
vesting in livestock. Beside that, the assistants and
nurses working in the veterinarian’s office access live-
stock health data to prepare health summaries or to as-
sign periodic timing for vaccination. In this case, they
should be granted a lower level of accessibility to the
data or to be granted accessibility only upon request.
Critical Data: When treating livestock with med-
ications, it is very important to use the appropri-
ate dosage which is generally determined in rela-
tion to the animal’s weight. Overdosing can cause
death while underdosing may lead to drug resistance.
Therefore, such data should be secured to prevent ad-
versaries from harming the livestock .
Livestock Environment: The environment where the
animals are kept is monitored with different types of
sensors to prevent animal sickness and death. If a ri-
valry gained accessibility to these sensors, they could
Figure 3: Farm Operations & Accessibility.
make adverse conditions inside the barns to affect the
yield of the animals or make them sick.
3.6 Smart Equipment
Smart farming depends heavily on a large number of
smart equipment with various security related aspects
that need controlled access control policies.
Leasing, Lending & Borrowing: Smart farm equip-
ment are very expensive. Therefore, farmers turn to
leasing them. However, leased equipment can col-
lect sensitive crop yield data, fertilization’s applica-
tion rates, field conditions etc. This data can infer
the financial position of the farm and also affect the
selling prices of crops. Therefore, leased equipment
should be subject to access control negotiations be-
tween the farmer and the leasing company. Some
farmers also tend to lend and borrow equipment from
neighboring farms. Therefore, both farm’s data can
be combined to infer information about crop health
in the area and animal disease outbreaks. Therefore
there should be specific access control mechanisms to
ensure that the lender and the borrower of the equip-
ment do not obtain data about each other’s farms. This
can be achieved through policies that ensure data col-
lection is restricted by the location of the equipment
as well as the time of the use.
Equipment Repairs: Some companies such as John
Deere had made it illegal to repair their farm tractor
(Horton and Kirchmeier, 2020). And in case of the
Effective Efficient Access Control in Smart Farms: Opportunities, Challenges Potential Approaches
449
tractor malfunction, the farmer will be forced to re-
pair the tractor at the company which will possibly
need access to the tractor sensors and data to identify
the source of the malfunction. In this case, the farmer
should be dynamically notified of any attempted ac-
cess to his equipment’s data through access control
requests and to have power to accept or deny such re-
quests.
Specialized Equipment: Some equipment such as
drones can collect a huge amount of data about the
food production process, location of critical infras-
tructure, footage of workers and livestock etc. Those
drones might be incorporated from other traditional
IoT sectors such as the military sector. This sector’s
drones are capable of monitoring and identifying their
users and tracking their performance and can compro-
mise the security of the whole smart farm system.
Critical Data: The pH levels are very critical to the
health of the crops. only specific levels allow the nu-
trients to be more available and effectively absorbed
by the plants. Any tamper of these levels can lead to
deficiencies in plant nutrients or plant toxitization if
it was not corrected by farmers. It can be corrected
by adjusting the ammonium supply levels. If adver-
saries get insight to this critical type of data, they can
plan a physical attack that aims to temper the pH level
with required levels of ammonium that ensure that the
crops will be damaged.
4 RESEARCH QUESTIONS
There are multiple access control policies that can be
applied to smart settings in general as discussed in
section 2.4. However, Identifying which of these poli-
cies are relevant to smart farms and can be practically
applied to it might be tricky. Therefore, we discuss
some of the questions that are critical for determining
the best access control mechanism that suits farms.
1. Are the Policies based on Devices or Capabili-
ties?
Current access control policies implementations in
other smart systems are mostly device-based. That
means that the access is either granted or denied per
the device itself. However, this implementation is not
flexible enough to capture the access control desires
that the farmers may think of to protect their data.
Therefore, the policies shouldn’t be simplified for de-
vice control only but rather it should be expanded to
involve its capability as well. This is driven by our
observation that the majority of smart farm devices
combine multiple capabilities with different sensitiv-
ities in a single device. For example, Drones are ca-
pable of tracking livestock and spraying fields. The
sensitivity of livestock tracking might be higher than
spraying fields from the farmers point of view. While
other capabilities such as erasing sensor’s data is very
critical and might not be granted to any of the stake-
holders. Therefore, a capability-centric model should
be designed where the capabilities that can be per-
formed over each smart device can be defined.
2. Which Contextual Factors Affect the Policies?
Some contextual factors might be essential to con-
sider with access control policies such as:
Location of the Devices: Ensures that certain de-
vices’ capabilities are used with devices in certain lo-
cations.
Location of the Stakeholders: Ensure that stake-
holders are restricted from using certain capabilities
unless they are connected to the farm network.
Time of Use: Is important especially for tasks that
should be timed in advance.
Presence of Others: Is important for some capabili-
ties that require supervision from others.
3. Should Policies Vary based on the Roles?
Since there are many roles assigned to different indi-
viduals inside the farm, those roles must play an im-
portant factor when deciding the suitable access con-
trol policies. For instance, the owner’s role differs
from a regular worker’s role and therefore the individ-
ual with the owner role should be able to express ac-
cess controls over the individual with a regular worker
role. Also, a returning temporary worker might be
more trusted than a new temporary worker and there-
fore can be assigned a role with more flexible access
controls in the farm system.
4. What happens when Access Controls Fail?
There are consequences for system malfunctions re-
lated to access controls including:
False Acceptance: This type of malfunction refers
to granting accessibility to some individual who was
originally intended to be denied it. The severity of
this malfunction depends on who was allowed incor-
rect access control. For example, a mild false accep-
tance might be caused by allowing co-owners access
to a capability they don’t need. A more severe case
would happen if a temporary worker was falsely al-
lowed access to a critical capability such as deleting
sensor’s data. The severest form would be granting
access to a 3rd party individual such as a farm rival.
False Rejection: This type of malfunction refers to
denying accessibility to some individual who was
originally intended to be granted it. The severity
of this type of malfunction might be less problem-
atic than the false acceptance malfunction. However,
it might disrupt essential work schedules and cause
losses of profits.
Continuous Accessibility: This failure comes in the
ICISSP 2022 - 8th International Conference on Information Systems Security and Privacy
450
form of granting continuous accessibility to individ-
uals who were supposed to access the system for a
specific period of time. Therefore, a policy expiration
option should be provided to allow only temporary
access for data and devices for as long as needed and
to revoke that access automatically once that period
ends in order to avoid any undesired accessibility.
5. How to Design Default Policies?
The process of assigning policies can be simplified
through default policies such as:
Default Accept: That can be provided in a very cau-
tious manner to the most trusted roles over some pre-
specified capabilities. It can also be provided over mi-
nor capabilities or data that are considered risk-free.
Default Deny: Should be implemented over critical
capabilities such as erasing sensitive data and any ad-
ministrative capabilities such as adding and removing
users and devices to the system.
5 DISCUSSION & POTENTIAL
APPROACHES
Reflecting on the research questions, Different Access
control policies can be used in smart farm systems.
Role-Based Access Control ”RBAC”: This model
can be used to ensure that stakeholders have access
only to the resources they need to complete their jobs.
The accessibility can be restricted based on the indi-
vidual’s assigned role as roles can determine the per-
missions granted to each individual and ensure that
no individual can perform higher level operations on
farm resources. Hierarchy of the roles allows specific
roles to inherit permissions from other roles. This
will benefit the higher level roles such as Owners-role
to share the responsibility of adding new users to the
system, assigning access policies for them and adding
or removing smart devices from the system etc.
The different roles that can be implemented in Smart
farm system includes: Owner, veterinarian, veteri-
narian assistant, temporary field worker, regular field
worker, purchaser etc. As hierarchy suggests, the
role of temporary field worker can inherit some ac-
cess controls and permissions from the regular field
worker role etc.
However, Hierarchical roles might be a major cause
of conflicts between stakeholders when they have
conflicting access control demands over system re-
sources. Therefore, resolving this kind of conflict
should be achieved automatically. The smart system
should be aware and able to identify conflicts and re-
solve them based on the priority of each role in the
hierarchical level.
As per the contextual factors that affect access control
policies both LBAC and TBAC models can capture
the majority of them.
Location-Based Access Control ”LBAC”: This
model can be used to capture the stakeholders’ loca-
tion and determine their accessibility rights based on
that location. For example, a drone operator can only
operate it when he is connected to the farm network.
Time-Based Access Control ”TBAC”: This model
can be used to ensure that different system resources
can be accessed only based on predefined times. For
example, a livestock purchasers’ access controls to
livestock data can be timed in advance to end by the
time they make their decision or in another case a
regular worker shouldn’t access crops data after the
working hours.
Attribute-Based Access Control ”ABAC”: This
model can place more limits on the access based on
User’s attributes, Environment’s attributes and Re-
source’s attributes. For Instance, it can be used hi-
erarchically with ”RBAC” which can be used first to
determine who has access to a resource then ”ABAC”
determine what they do with the resource based on
different attributes.
Figure 4: Creation and Enforcement of Access Control
Policies.
As different devices have many capabilities, the
capabilities must be incorporated in the process of as-
signing access control policies. Capabilities reflect
more fine-grained access control policies than the per-
device access controls.
Based on our observation of the smart farm nature,
we argue that the smart farm access control system
should be multi-user, multi-device aware where farm-
ers can express their system requirements through
policies. The system must contain main components
as follows:
1. Friendly User Interface: The system should have
a unified friendly UI for adding/removing users and
devices and assigning policies to them.
2. Expressive Policy Language: Whenever the se-
Effective Efficient Access Control in Smart Farms: Opportunities, Challenges Potential Approaches
451
curity level becomes more complicated, the usabil-
ity usually drops. However, any complicated process
might discourage farmers from using such systems,
especially that the security awareness in farms is not
high. Therefore, The system should have simple ex-
pressive policy language such as roles, time, loca-
tion, device, capability etc. In addition, It could be
beneficial to use policy automation to fill the gap be-
tween farmer intuitive policies and the matching de-
tailed technical configurations and rules.
3. Resolving Conflicts: The system should be able to
identify and resolve conflicts of access demands.
4. Access Policies: Policies should at least consist
of the triplet: <Individual’s Role, Device Capability,
Contextual Factors>.
5. Policy Creation, Enforcement & Execution: All
specified access control policies should be enforced
on the system. And access control requests should be
evaluated against the enforced policies.
Smart devices are usually managed through a hub
device. The hub device is used to facilitate the com-
munication between devices and the cloud. Once the
farmer specifies the required access controls, it should
be sent through the hub device to the server on the
cloud. That server then should generate the required
policies and ensure it is enforced over all system re-
sources. Whenever an access control to a system re-
source is requested, it should be automatically for-
warded to the server. The server should check the
request against the created policies. If the request is
valid it should be accepted on the specified resource,
otherwise it should be denied as shown in Figure 4.
6 RELATED WORK
Many researches have started looking into the secu-
rity and privacy issues in different IoT domains such
as (Fan et al., 2019) who designed access controls
considering fog computing for providing data confi-
dentiality, variability and attribute based encryption.
However, there have been a limited work explor-
ing specifically the security in smart farming and pre-
cision agriculture despite that the U.S. Department
of Homeland Security issued a report (Aida Boghos-
sian, 2018) identifying the different threats to preci-
sion agriculture and emphasizing the need for more
research regarding this area. Most researches were
focused on blockchain solutions such as (Kamilaris
et al., 2019) who studied the challenges and implica-
tions of using blockchain technology projects in the
agriculture sector and (Ferrag et al., 2020) who pro-
vided the consensus algorithms for the solutions that
are based on blockchain and how can they be adapted
for smart farming. We omit further blockchain work
as its not related to our main focus.
7 PAPER SUMMARY
Smart farming is an emerging sector of IoT applica-
tions. It faces many challenges ranging from adoption
of its technologies to the security issues that stems
from applying it inside the farms. Therefor, an exten-
sive research is required in this area. In this paper we
explored different security scenarios that stems form
the diverse nature of smart farms. We identified the
structure of smart farms and the different stakeholders
who are involved in the system. We explored relevant
access control policies that can be particularly applied
to smart farms.
REFERENCES
2021 Smart farming: How iot-driven precision agriculture
helps feed the globe.
Aida Boghossian, S. L. (2018). Threats to precision agricul-
ture,u.s. dept. homeland secur., washington, dc, USA.
Bigelow, D. (2019). Farmland ownership and tenure.
Agricultural Resources and Environmental Indicators,
2019, 87:84.
Bigelow, D., Borchers, A., and Hubbs, T. (2016). Us farm-
land ownership, tenure, and transfer. Technical report.
Bruinsma, J. et al. (2009). The resource outlook to 2050:
by how much do land, water and crop yields need to
increase by 2050. In Expert meeting on how to feed
the world, volume 2050, pages 24–26.
De Araujo Zanella, A. R., da Silva, E., and Albini, L. C. P.
(2020). Security challenges to smart agriculture: Cur-
rent state, key issues, and future directions. Array,
page 100048.
Fan, K., Xu, H., Gao, L., Li, H., and Yang, Y. (2019). Ef-
ficient and privacy preserving access control scheme
for fog-enabled iot. Future Generation Computer Sys-
tems, 99:134–142.
Ferrag, M. A., Shu, L., Yang, X., Derhab, A., and Maglaras,
L. (2020). Security and privacy for green iot-based
agriculture: Review, blockchain solutions, and chal-
lenges. IEEE access, 8:32031–32053.
Horton, T. J. and Kirchmeier, D. (2020). John deere’s at-
tempted monopolization of equipment repair, and the
digital agricultural data market-who will stand up for
american farmers? CPI Antitrust Chronicle, Jan.
Kamilaris, A., Fonts, A., and Prenafeta-Bold
´
ν, F. X. (2019).
The rise of blockchain technology in agriculture and
food supply chains. Trends in Food Science & Tech-
nology, 91:640–652.
Roser, M.(2013). Future population growth. Our World
in Data https://ourworldindata.org/future-population-
growth.
ICISSP 2022 - 8th International Conference on Information Systems Security and Privacy
452