Enhancing Older Adults’ Well-Being Through QoS-Aware Edge-Cloud

eHealth Applications

Md Mahfuzur Rahman

1, 2 a

Department of Information and Computer Science, King Fahd University of Petroleum & Minerals (KFUPM),

Dhahran, 31261, Saudi Arabia

Interdisciplinary Research Center for Intelligent Secure Systems (IRC-ISS),

King Fahd University of Petroleum & Minerals (KFUPM), Dhahran, 31261, Saudi Arabia

{ mdmahfuzur.rahman}@kfupm.edu.sa

Keywords:

Older Adults’ Healthcare, Edge-Cloud Computing, Quality of Service.

Abstract:

Ensuring the well-being of older adults using eHealth applications has become increasingly important. Con-

tinuous health monitoring, emergency response, and various personalized applications can improve the well-

being of the older adults through intelligent and adaptive data processing. But, the effectiveness of such

systems highly depends on the provisioning of efﬁcient computing resources that meet Quality of Service

(QoS) requirements (of the applications), including low latency, faster computation, reliability, and security.

Leveraging Edge-Cloud Computing can offer a promising solution but an effective strategy is required to un-

derstand the Quality of Service (QoS) requirements of applications while ofﬂoading in Edge-Cloud Comput-

ing. Previous research lacks in focusing predicted behaviour of application workloads (and related QoS) while

scheduled on the Edge-Cloud platform. This research addresses this issue by proposing a scalable model facil-

itating resource allocation based on the speciﬁc future requirements of data processing tasks. In this research,

an efﬁcient heuristic algorithm is developed to maximize meeting the QoS constraints. The effectiveness of

the proposed approach is evaluated through simulations comparing its performance against existing methods,

thereby facilitating improved service delivery and user satisfaction.

1 INTRODUCTION

The rapid growth of the aging population presents sig-

niﬁcant challenges for healthcare systems worldwide.

As older adults face increased risks of chronic dis-

eases, cognitive decline, and mobility limitations, the

demand for continuous and efﬁcient healthcare ser-

vices is rising. Traditional healthcare models, which

rely heavily on in-person visits and hospitalization,

are often insufﬁcient in addressing the real-time

health monitoring and intervention needs of older in-

dividuals. In response to these challenges, eHealth

applications have emerged as a promising solution to

provide remote health monitoring, emergency assis-

tance, and personalized care.

Recent advancements in Edge-Cloud Computing

have further improved the capabilities of eHealth sys-

tems by enabling real-time data processing, intelli-

gent decision-making, and scalable analytics. Edge

computing brings computation closer to the data

source, reducing latency and improving responsive-

https://orcid.org/0000-0002-2871-9119

ness, while cloud computing provides centralized

storage, long-term analytics, and machine learning-

driven insights. However, ensuring Quality of Service

(QoS) in such systems remains a critical challenge, as

factors like network congestion, device heterogeneity,

data security, and real-time processing requirements

can impact the overall performance of health moni-

toring applications. In this paper, we propose a QoS-

aware Edge-Cloud eHealth monitoring framework de-

signed to enhance the well-being of older individu-

als. The framework integrates adaptive resource pro-

visioning to optimize computational efﬁciency, net-

work performance, and data security.

The proposed approach invents new and efﬁcient

heuristic solutions based on predicted resource alloca-

tion schemes. Consider, for example, an older people

or patient monitoring app (running on a smart phone)

that processes Internet of Things (IoT) data obtained

from various IoT sources such as wearable physiolog-

ical sensors, etc. The workload created with this mon-

itoring app can be handled locally by the smartphone

but since the smartphone is a power-limited device,

Rahman, M. M.

Enhancing Older Adults’ Well-Being Through QoS-Aware Edge-Cloud eHealth Applications.

DOI: 10.5220/0013505900003938

Paper published under CC license (CC BY-NC-ND 4.0)

In Proceedings of the 11th International Conference on Information and Communication Technologies for Ageing Well and e-Health (ICT4AWE 2025), pages 443-450

ISBN: 978-989-758-743-6; ISSN: 2184-4984

443

the computation can be also located in Edge-Cloud

environment depending on the QoS preferences. By

analysing the collected IoT data, the app can also de-

tect any abnormal behaviour of the patient and sub-

sequently more IoT devices (e.g. IP camera) can be

turned on and the obtained video data are then ana-

lyzed to monitor and understand the patient’s move-

ment more precisely and effectively. In that case, the

computation is preferred to be taken place in Edge/-

Cloud environment considering the predicted future

scalable workload. Essentially, it needs to be decided

that which part of the tasks should be ofﬂoaded to

Edge/Cloud so that smartphone resources can be still

utilized efﬁciently and also high computation require-

ment can be satisﬁed by migrating workload to Edge/-

Cloud. This migration to either Edge or Cloud also

depends on various QoS requirements (e.g. the min-

imum tolerable latency, privacy of data, etc.). There-

fore, in summary the research has the following key

goals:

• Design a dynamic computation ofﬂoading frame-

work for QoS-aware resource allocation for

eHealth data and services.

• Evaluate the suitability of proposed approach

through simulations with a simulation tool.

Section 2 describes the related work and section 3 de-

tails the design of the dynamic computation ofﬂoad-

ing framework for QoS-aware resource allocation for

eHealth data and services. Section 4 shows the evalu-

ation results, and ﬁnally, Section 5 concludes the pa-

per.

2 RELATED WORK

Edge computing has been widely adopted in health-

care systems to enable real-time health monitoring

and reduce latency in critical medical applications.

Studies such as (Islam et al., 2024) and (Ahmed et al.,

2023) propose edge-based architectures for process-

ing vital signs, motion detection, and chronic disease

monitoring in older individuals. The cloud compo-

nent provides additional storage, long-term analysis,

and integration with medical institutions. However,

these studies primarily focus on edge-based computa-

tion without addressing dynamic resource allocation

challenges for QoS enhancement.

Meeting QoS requirements in data processing is

crucial. Without fulﬁlling strict QoS requirements,

the results may not be presented to the users in a sat-

isfactory manner. Various researchers (Herrera et al.,

2020; K

ulzer et al., 2021) have addressed QoS is-

sues in data processing scenarios. Hoseinyfarahabady

et al. (Hoseinyfarahabady et al., 2019) discusses a

cloud scenario, where the disk I/O bandwidth is iden-

tiﬁed as the potential bottleneck causing QoS viola-

tion. Thus, they proposed an instance placement al-

gorithm with full consideration of disk I/O balancing.

The authors in (Hoseiny et al., 2021) formulated task

scheduling problem as an NP-hard problem and in-

troduced two task scheduling algorithms to allocate

IoT workload on edge-cloud environment. The algo-

rithms minimize the computation cost, communica-

tion cost, and delay violation and the performance im-

provement made by those algorithms were compared

with the genetic-based algorithm.

Recent research in (Mukhopadhyay et al., 2024)

has explored hybrid Edge-Cloud frameworks for re-

mote health monitoring, where computing tasks are

intelligently ofﬂoaded to either edge or cloud nodes

based on network conditions and resource availability.

These works highlight the importance of low-latency

decision-making but fail to address the impact of net-

work congestion, device heterogeneity, and real-time

QoS constraints, which are critical for older adults’

healthcare applications. Works in (Louvros et al.,

2023) and (Chi et al., 2020) introduce QoS-aware

scheduling algorithms that prioritize emergency cases

and optimize resource allocation across distributed

healthcare networks. However, they rely on static pro-

visioning techniques that do not dynamically adjust to

real-time network conditions, leading to potential de-

lays in older adults’ healthcare applications.

Moreover, research in (Peng et al., 2023) proposes

Artiﬁcial Intelligence (AI)-driven resource schedul-

ing that leverages machine learning to predict net-

work congestion and adjust resource allocation dy-

namically. While this enhances service reliability,

the high computational overhead of AI models may

pose scalability issues when deployed on resource-

constrained edge devices. Righi et al. (da Rosa Righi

et al., 2020) used Autoregressive Integrated Moving

Average (ARIMA) and Weighted Moving Average to

predict IoT load behavior and to anticipate future scal-

ing in and out. Etemadi et al. (Etemadi et al., 2020)

used time series prediction model to predict the IoT

workload and Xu et al. (Xu et al., 2020) used ensem-

ble learning algorithm to make accurate predictions

on IoT workload. Our proposed framework builds

upon workload prediction approaches by introducing

adaptive edge-cloud resource provisioning that bal-

ances latency, computational efﬁciency, and network

performance for real-time older person’s health mon-

itoring.

The study in (Rema and Sikdar, 2021) focuses on

addressing overcrowding in hospital emergency de-

partments (ED) by employing quantitative methods

IS4WB_SC 2025 - Special Session on Innovative Strategies to Enhance Older Adults’ Well-being and Social Connections

444

for resource planning and deployment. By analyz-

ing patient ﬂow and forecasting demand, hospital ad-

ministrators can make informed decisions. The study

examines 7748 ED arrivals from a Bengaluru hospi-

tal, analyzing patient ﬂow during each working shift.

Time series modeling techniques, particularly expo-

nential smoothing proposed by Hyndman, were used

to generate short-term forecasts. Model validation

and residual analysis were conducted to ensure accu-

racy. Prediction intervals with a 90% conﬁdence level

were obtained on a shift-wise basis, allowing for efﬁ-

cient resource reallocation and demand estimation by

hospital management. (OpenAI, 2024).

3 PROPOSED SYSTEM

Edge-Cloud system architecture combines the dis-

tributed computing capabilities with the scalability

and storage potential of cloud infrastructure, creat-

ing a versatile framework for IoT applications. At

its core, this architecture leverages edge devices or

fog nodes situated closer to the data source for real-

time processing and decision-making, reducing la-

tency and bandwidth usage. These edge/fog nodes act

as intermediaries between IoT devices and the cen-

tralized cloud servers, ﬁltering and analyzing data lo-

cally before transmitting relevant information to the

cloud for further processing or storage. This hierar-

chical approach optimizes resource utilization and en-

hances the responsiveness and efﬁciency of IoT sys-

tems, making them suitable for a wide range of ap-

plications, from smart cities to industrial automation.

Additionally, the edge-cloud architecture offers ﬂexi-

bility in deployment, enabling seamless integration of

new devices and services while ensuring data secu-

rity and privacy through robust encryption and access

control mechanisms. Figure 1 shows an architecture

related to Edge-Cloud system architecture considered

in this research.

3.1 Remote Healthcare Monitoring

Application

An eHealth application represents a digital solution

that harnesses technology to deliver healthcare ser-

vices and information remotely. These applications

encompass a wide range of functionalities, including

telemedicine consultations, electronic health record

management, remote monitoring of vital signs, medi-

cation management, and health education resources,

etc. A Remote Healthcare Monitoring Application

is a digital solution designed to facilitate the remote

monitoring of patients’ health and medical conditions.

Figure 1: Edge-Cloud Architecture.

This innovative platform leverages technology to en-

able healthcare providers to monitor patients’ vital

signs, symptoms, and medication adherence from a

distance, thereby enhancing patient care and manage-

ment. Through the use of various medical devices

such as wearable sensors, smart monitors, and mobile

applications, patients can conveniently transmit real-

time health data to healthcare professionals, allowing

for timely interventions and adjustments to treatment

plans. Remote Healthcare Monitoring Applications

offer numerous beneﬁts, including improved access

to care for patients in remote or underserved areas,

early detection of health issues, reduced hospital ad-

missions, and enhanced patient engagement and em-

powerment through active participation in their own

healthcare journey. Overall, these applications play a

vital role in advancing telemedicine and revolution-

izing the delivery of healthcare services by bridging

the gap between patients and providers regardless of

geographical barriers.

Remote Healthcare Monitoring Applications uti-

lize IoT devices to gather real-time health data from

patients remotely. These IoT devices, such as wear-

able sensors and smart monitors, continuously collect

various health metrics like heart rate, blood pressure,

and activity levels. This data is then transmitted se-

curely to the monitoring application, where it is an-

alyzed and interpreted. The workload in this con-

text refers to the processing and analysis of the vast

amount of health data generated by these IoT devices.

In this research, a remote healthcare monitoring ap-

plication is considered and the releated workload is

efﬁciently analyzed to ensure timely monitoring and

intervention for patients. Additionally, the historical

time-series information of workloads (related to data

processing, analysis, and interpretation) are consid-

ered to make future workload prediction for providing

effective remote healthcare monitoring services.

Enhancing Older Adults’ Well-Being Through QoS-Aware Edge-Cloud eHealth Applications

445

3.2 Dataset

Workload datasets provide valuable insights into the

performance and efﬁciency of the application, allow-

ing healthcare providers to make informed decisions

regarding resource allocation, capacity planning, and

optimization strategies. By analyzing workload data,

providers can identify patterns, trends, and areas for

improvement in the delivery of remote healthcare ser-

vices. Understanding the workload patterns and de-

mands helps in optimizing the allocation of resources

such as computing power, storage, and network band-

width within the application hosting infrastructure.

This ensures that the application can effectively han-

dle ﬂuctuations in workload without compromising

performance or patient care.

The dataset, used in this research, was gathered

from IoT devices installed in apartments occupied by

elderly individuals living alone and was subsequently

uploaded to the SSiO platform (Swedish Society for

Industrial Organization, 2025). It was collected in

compliance with General Data Protection Regulation

(GDPR) regulations, ensuring that individual identi-

ties cannot be discerned from the data. The data was

collected on an apartment-by-apartment basis, with

timestamps indicating when IoT events occurred. The

dataset spans from 2019 to 2021. This realistic in-

coming trafﬁc from the SSiO IoT healthcare applica-

tion system (Swedish Society for Industrial Organiza-

tion, 2025) is studied, developed and modeled in this

research. Figure 2 displays day-long samples for 4

different days extracted from the considered dataset.

3.3 Prediction Model

Preparing a workload prediction model for a health-

care monitoring application involves several key steps

to ensure its accuracy and effectiveness. Initially,

it requires gathering and preprocessing relevant data

sources, including patient demographics, medical his-

tory, vital signs, and historical admission records.

Next, suitable time-series models are selected based

on the nature of the data and the prediction task.

These models are trained using historical data to learn

patterns and relationships between variables that in-

ﬂuence workload ﬂuctuations in healthcare settings.

Additionally, the model may incorporate external fac-

tors such as seasonal variations, public health trends,

and demographic shifts to improve its predictive ac-

curacy. After training, the model is validated using

separate datasets to assess its performance and gen-

eralization capability. Finally, the SARIMA model

is deployed within the healthcare monitoring applica-

tion, where it continuously analyzes real-time data to

Figure 2: Example Workloads.

Figure 3: Workload Prediction using SARIMA model.

predict future workload demands (see Figure 3). Reg-

ular monitoring and evaluation of the model’s perfor-

mance ensure its reliability and relevance over time,

allowing healthcare providers to proactively manage

resources and optimize patient care delivery.

3.4 QoS Aware Task Placement (QTP)

The developed task placement algorithm (Algo-

rithm 1) follows a systematic process for task schedul-

ing at a ﬁxed interval µ. Initially, incoming task

requests are accumulated into a batch, termed new-

IS4WB_SC 2025 - Special Session on Innovative Strategies to Enhance Older Adults’ Well-being and Social Connections

446

TaskList, and assessed for their Quality of Service

(QoS) requirements [Algorithm 1, Line 1-3]. Tasks

are categorized based on their QoS sensitivity, in-

cluding considerations for latency, security, and scal-

ability. If a task’s certain QoS requirement cross’s

a threshold α then it is marked as QoSSensitive to

be scheduled properly. If a task is not QoSSenstive

then it will be marked as well [Lines 3-9]. These

tasks are then merged with existing tasks in taskList

for further processing [Line 11]. Workload predic-

tion is conducted for each task until a future time t+µ

(where t is the current time and t+µ is the next re-

evaluation time), distinguishing between Compute-

Sensitive and ComputeNonSensitive tasks based on

workload thresholds β [Lines 13-16].

Algorithm 1

1 Repeatedly scan newTaskList for new task at a

fixed interval µ

3 Characterize each task depending on QoS

requirements

5 Foreach task

in newTaskList:

6 IF task

.QoSRequirements > α

7 mark task

as QoSSensitive

8 ELSE mark task

as QoSNonSensitive

9 End For

11 taskList ← merge(newTaskList, existingTaskList)

13 Foreach task

in taskList:

14 IF predict(task

.ComputeRequirements, µ) > β

15 mark task

as ComputeSensitive

16 ELSE mark task

as ComputeNonSensitive

18 T

← MakeSortedSubList(QoSNonSensitive,

ComputeSensitive,taskList)

20 T

← MakeSortedSubList(QoSSensitive,

ComputeNonSensitive,taskList)

22 T

← MakeSortedSubList(QoSSensitive,

ComputeSensitive,taskList)

24 T

← MakeSortedSubList(QoSNonSensitive,

ComputeNonSensitive,taskList)

26 T

← MergeSubTaskListsSequentially(T

, T

)

28 Foreach task t

in T

29 If t

in existingTaskList and t

on Fog

Platform:

30 offloadingCost(t

, FogToCloud) < γ :

31 success ← schedule t

on Cloud

Platform

32 if success equals false :

33 add t

at the end of T

34 ElIf t

in newTaskList:

35 success ← schedule t

on Cloud Platform

36 if success equals false :

37 add t

at the end of T

39 Foreach task t

in T

40 success ← schedule t

on Fog Platform

41 if success equals false :

42 success ← schedule t

on Cloud Platform

43 if success equals false :

44 schedule t

on local machine

46 Consider taskList as existingTaskList for next

interval

Tasks are further organized into four sublists (T

, T

) according to their QoS and Compute sen-

sitivity, with T

representing a special sublist formed

by merging T

, T

, and T

[Lines 18-26]. Allocation

decisions prioritize QoSNonSensitive but Compute-

Sensitive tasks for Cloud platform which also consid-

ers the cost threshold γ due to the task ofﬂoading from

Fog to Cloud platforms [Lines 28-33]. Newly arriv-

ing tasks are allocated directly to the Cloud platform

to address compute-sensitive requirements [Lines 35-

37]. Finally, tasks in the T

special sublist are sched-

uled sequentially, with preference given to Fog plat-

forms unless QoS constraints necessitate Cloud plat-

form allocation. Tasks unable to be scheduled re-

motely are assigned to the local machine [Lines 39-

44]. At the conclusion of the algorithm, taskList is

updated as existingTaskList for scheduling in the sub-

sequent interval [Line 46]. This structured approach

ensures efﬁcient task allocation while accommodat-

ing diverse QoS and compute requirements within the

Edge-Cloud environment.

4 ASSESSMENT

Figure 4: Internet2 Abilene Topology (Beck and Moore,

1998).

The QoS satisfaction of QoS-sensitive applica-

tions was considered to evaluate how well the sug-

gested QoS Aware Task Placement (QTP) Algorithm

performed. The RYU controller (RYU-Community,

2024) and the Mininet emulator (Mininet-Project,

2024) were used for the evaluation of the algorithm.

Enhancing Older Adults’ Well-Being Through QoS-Aware Edge-Cloud eHealth Applications

447

The RYU SDN Controller offers application program

interfaces (APIs) for creating new control for data

ﬂows, while the Mininet emulator allows to simulate

a network of virtual switches, hosts, controllers, and

links. A directed graph G = (V, E), where V is the set

of nodes or switches and E is the set of links, is typ-

ically used to depict the underlying network. Based

on the acquired link-state, each link (u, v) in E is ex-

pressed with QoS information as C(u, v) including

bandwidth, etc.

As seen in Figure 4, a network architecture based

on the Internet2 Abilene backbone network topol-

ogy (Beck and Moore, 1998) was also taken into con-

sideration for the assessment. There are eleven Inter-

net service providers (ISPs) in the customized topol-

ogy. The bidirectional link between the ISPs is con-

ﬁgured as of a 10 Mbps capacity, a random packet

loss percentage between 1 and 5, and a random delay

between 1 and 5 ms. The scenario consists of three

cloud-based infrastructure as a service (IaaS) servers

linked to respective ISPs (S3, S4, S5). In this experi-

ment, three different kinds of virtual machines (VMs)

are taken into consideration in the cloud. Table 1 dis-

plays the conﬁguration details for each type of virtual

machine. These virtual machine types vary depend-

ing on the virtual CPU, memory, and network band-

width that are available. For instance, LowConfVM

provides 1 virtual CPU, 500 MB memory, limited net-

work bandwidth, and elastic storage, while HighCon-

fVM gives 2 virtual CPUs, 3 GB of memory, and high

network bandwidth. Each type of virtual machine

(VM) has an hourly fee that is determined by the pric-

ing model provided by the Amazon EC2 cloud (Ser-

vices, 2024). Additionally, the three distinct users are

connected via S0, S1, and S2 ISPs while edge devices

are also offered by those three ISPs to assist clients, if

needed.

An experiment was carried out with a speciﬁc case

scenario and a few important QoS parameters related

to an application’s preferred data trafﬁc were identi-

ﬁed to quantify the performance. The objective is to

evaluate the created QTP algorithm’s cost and QoS

performance to that of non-QTP algorithms. The non-

QTP algorithm places compute-intensive workloads

on the cloud without taking the workload’s quality

of service (QoS) into account. To compare the per-

formance of the algorithms, crucial QoS performance

metrics — cost, throughput, latency, and packet loss

are taken into account. The total data transfer rate of

the application’s accepted ﬂows, or throughput, is the

result of adding up all of the accepted ﬂows. The per-

centage of packets lost while the packets are sent by

the accepted ﬂows is measured by packet loss. The

amount of time taken for a network communication

by a packet is characterized as latency or network de-

lay.

QoS performances are traced for both QTP and

non-QTP methods because of the network ﬂows that

are observed between a randomly selected client and

server. In this experiment, the eHealth workload

between client and server was generated using the

Iperf (Hardin et al., 2023) program, and the aver-

age performance data of many runs was measured ap-

propriately. Reducing latency and packet loss during

network communication is crucial for remote health-

care monitoring applications, since their workloads

are highly susceptible to changes in throughput, de-

lay, and packet loss (Ravi et al., 2024). The QTP algo-

rithm allows and reprovisions resources (after a ﬁxed

interval) between cloud and edge devices based on

the anticipated workload. According to the simulation

results, the QTP algorithm can successfully increase

throughput while lowering costs, packet loss, and net-

work latency for the eHealth application because it

can provide the proper environment to meet all neces-

sary QoS requirements. Alternatively, the Non-QTP

method does not take into account the anticipated

workload i.e. places the application workload stati-

cally without considering continuing changes of the

workload. Non-QTP cannot satisfy the QoS charac-

teristics (such as latency and loss) during application

execution and cannot offer effective resource provi-

sioning for ﬂuctuating workloads. As QTP chose a

superior and desired resource provisioning strategy, it

subsequently offered greater throughput, fewer packet

loss, and less latency as compared to Non-QTP (see

Figure 5).

4.1 Results

The tasks of remote healthcare monitoring application

exhibit ﬂuctuating workloads over time, predicted us-

ing the developed time-series prediction model. Ini-

tially, utilizing the developed provisioning algorithm,

all tasks were allocated across Cloud-Edge devices

based on their QoS characteristics and Compute Sen-

sitivity. With a deﬁned time interval, the QTP algo-

rithm analyzed the future workload of each task, re-

sulting in the identiﬁcation of the tasks to be relocated

between cloud platform and edge devices.

In the experiment, three different predicted time

periods are considered i.e. 3 hours, 6 hours, and 12

hours. An increase in throughput was observed when

the QoS of running tasks on the Cloud-Edge was as-

sessed for each interval (see Figure 5). Throughput

increased considerably as a result of the QTP algo-

rithm’s ability to make efﬁcient reallocation strate-

gies. This was made possible by the increased options

IS4WB_SC 2025 - Special Session on Innovative Strategies to Enhance Older Adults’ Well-being and Social Connections

448

Table 1: VM Conﬁguration.

Different Types of VM in Cloud and Related Costs

VM Type VCPU Memory Network Storage Hourly Cost (in Dollar)

HighConfVM 2 2 GB High Elastic 0.0188

MedConfVM 1 1 GB Moderate Elastic 0.0116

LowConfVM 1 0.5 GB Low Elastic 0.0058

Figure 5: Performance Comparison of Various Algorithms.

for applications to assign tasks to the optimal loca-

tions between the cloud and the edge. This advantage

is more pronounced when prediction and related rear-

rangement occur more often, for example, a 3-hours

prediction performs better than one made over longer

time intervals, i.e. six hours or twelve hours. Again,

for each type of prediction interval, latency or net-

work delay, was measured and the results indicate that

the more frequently the QTP algorithm validates the

relocation plan, the lower the latency (see Figure 5).

QTP algorithm also led to a signiﬁcant decrease in

data loss in a similar fashion (see Figure 5). Finally,

there are monetary costs associated with executing

tasks in cloud environment. With the efﬁcient resouce

provisioning strategy of QTP algorithm, the duration

of tasks execution in cloud is optimized so the cost is

also reduced signiﬁcantly while compared with Non-

QTP algorithm (see Figure 5). The experiment results

show that the QoS Aware Resource Provisioning in

Edge-Cloud can effectively handle the QoS require-

ments of eHealth application and the beneﬁt is more

signiﬁcant when the tasks are rearranged more fre-

quently based on forecasting model.

5 CONCLUSION

This research addresses the complex challenge of

managing Quality of Service (QoS) requirements in

Edge-Cloud computing environments, where both

Cloud and Edge/Fog computing platforms play cru-

cial roles in data processing tasks. While Edge/-

Fog computing excels in latency-sensitive applica-

tions, scalability remains a concern. The proposed

scalable model offers a strategic approach to resource

allocation, considering the speciﬁc needs of data pro-

cessing tasks and balancing QoS requirements effec-

tively. By developing an efﬁcient heuristic algorithm

and integrating a predictive model for eHealth work-

load behavior, this research signiﬁcantly enhances the

efﬁciency and effectiveness of resource management

in Edge-Cloud environments. Through simulations,

the proposed approach demonstrates superior perfor-

mance in terms of cost-effectiveness, response time,

and resource utilization compared to existing meth-

ods. Overall, this research contributes valuable in-

sights into optimizing service delivery and enhancing

Enhancing Older Adults’ Well-Being Through QoS-Aware Edge-Cloud eHealth Applications

449

user satisfaction in cloud and edge computing ecosys-

tems, paving the way for more robust and scalable ap-

plications in the future.

ACKNOWLEDGMENTS

The related work section was paraphrased using Chat-

GPT (OpenAI, 2024).

FUNDING STATEMENT

This work was funded by the King Fahd University

of Petroleum and Minerals, Dhahran, Saudi Arabia,

under the Deanship of Research (Grant-EC213004).

REFERENCES

Ahmed, S., Irfan, S., Kiran, N., Masood, N., Anjum, N.,

and Ramzan, N. (2023). Remote health monitor-

ing systems for elderly people: a survey. Sensors,

23(16):7095.

Beck, M. and Moore, T. (1998). The internet2 distributed

storage infrastructure project: An architecture for in-

ternet content channels. Computer Networks and

ISDN systems, 30(22-23):2141–2148.

Chi, H. R., Domingues, M. F., and Radwan, A. (2020). Qos-

aware small-cell-overlaid heterogeneous sensor net-

work deployment for ehealth. In 2020 IEEE SEN-

SORS, pages 1–4. IEEE.

da Rosa Righi, R., Correa, E., Gomes, M. M., and da Costa,

C. A. (2020). Enhancing performance of iot applica-

tions with load prediction and cloud elasticity. Future

Generation Computer Systems, 109:689–701.

Etemadi, M., Ghobaei-Arani, M., and Shahidinejad, A.

(2020). Resource provisioning for iot services in the

fog computing environment: An autonomic approach.

Computer Communications, 161:109–131.

Hardin, B., Comer, D., and Rastegarnia, A. (2023). On

the unreliability of network simulation results from

mininet and iperf. International Journal of Future

Computer and Communication, 12(1).

Herrera, J. L., Bellavista, P., Foschini, L., Gal

an-Jim

enez,

J., Murillo, J. M., and Berrocal, J. (2020). Meeting

stringent qos requirements in iiot-based scenarios. In

GLOBECOM 2020-2020 IEEE Global Communica-

tions Conference, pages 1–6. IEEE.

Hoseiny, F., Azizi, S., Shojafar, M., and Tafazolli, R.

(2021). Joint qos-aware and cost-efﬁcient task

scheduling for fog-cloud resources in a volunteer

computing system. ACM Transactions on Internet

Technology (TOIT), 21(4):1–21.

Hoseinyfarahabady, M. R., Tari, Z., and Zomaya, A. Y.

(2019). Disk throughput controller for cloud data-

centers. In 2019 20th International Conference on

Parallel and Distributed Computing, Applications and

Technologies (PDCAT), pages 404–409. IEEE.

Islam, M. Z., Sagar, A. S., and Kim, H. S. (2024). Enabling

pandemic-resilient healthcare: Edge-computing-

assisted real-time elderly caring monitoring system.

Applied Sciences, 14(18):8486.

ulzer, D. F., Kasparick, M., Palaios, A., Sattiraju, R.,

Ramos-Cantor, O. D., Wieruch, D., Tchouankem,

H., G

ottsch, F., Geuer, P., Schwardmann, J., et al.

(2021). AI4Mobile: Use cases and challenges of

AI-based QoS prediction for high-mobility scenarios.

In 2021 IEEE 93rd Vehicular Technology Conference

(VTC2021-Spring), pages 1–7. IEEE.

Louvros, S., Paraskevas, M., and Chrysikos, T. (2023).

QoS-aware resource management in 5G and 6G

cloud-based architectures with priorities. Information,

14(3):175.

Mininet-Project (2024). Mininet. https://mininet.org/.

Mukhopadhyay, A., Remanidevi Devidas, A., Rangan, V. P.,

and Ramesh, M. V. (2024). A QoS-aware IoT edge

network for mobile telemedicine enabling in-transit

monitoring of emergency patients. Future Internet,

16(2):52.

OpenAI (2024). ChatGPT: Conversational AI Model. https:

//openai.com/chatgpt. Accessed: [DATE].

Peng, D., Sun, L., Zhou, R., and Wang, Y. (2023).

Study QoS-aware fog computing for disease diagno-

sis and prognosis. Mobile Networks and Applications,

28(2):452–459.

Ravi, K. C., Kavitha, G., Prasad, L. H., Srinivasa Rao, N. V.,

Deivasigamani, S., Ramesh, J. V. N., and Siddiqui,

S. T. (2024). Beyond 5g-based smart hospitals: In-

tegrating connectivity and intelligence. Smart Hospi-

tals: 5G, 6G and Moving Beyond Connectivity, pages

169–193.

Rema, V. and Sikdar, K. (2021). Time series modelling

and forecasting of patient arrivals at an emergency de-

partment of a select hospital. In Recent trends in sig-

nal and image processing: ISSIP 2020, pages 53–65.

Springer.

RYU-Community (2024). Ryu sdn framework. https:

//ryu-sdn.org/.

Services, A. W. (2024). Aws pricing calculator for EC2

enhancements. https://calculator.aws/.

Swedish Society for Industrial Organization (2025).

Swedish society for industrial organization. https:

//ssio.se/.

Xu, Y., Li, J., Lu, Z., Wu, J., Hung, P. C., and Alelaiwi, A.

(2020). Arvmec: adaptive recommendation of virtual

machines for IoT in edge–cloud environment. Journal

of Parallel and Distributed Computing, 141:23–34.

IS4WB_SC 2025 - Special Session on Innovative Strategies to Enhance Older Adults’ Well-being and Social Connections

450