An Adaptive Hybrid Cloud Framework for Real‑Time Integration of

Smart City Services and Urban Data

Durgalakshmi B.

, Sowmiya K.

, M. Maria Sampoornam

, P. Jaisankar

, V. Divya

and Vignesh K.

Tagore Institute of Engineering and Technology, Deviyakurichi, Salem, Tamil Nadu, India

Department of Information Technology, J.J.College of Engineering and Technology, Tiruchirappalli, Tamil Nadu, India

Department of Mathematics, Nandha Engineering College (Autonomous), Erode -638052, Tamil Nadu, India

Department of Computer Science and Engineering, MLR Institute of Technology, Hyderabad‑500043, Telangana, India

Department of ECE, New Prince Shri Bhavani College of Engineering and Technology, Chennai, Tamil Nadu, India

Keywords: Hybrid Cloud, Smart City, Real‑Time Data, Edge Computing, Urban Integration.

Abstract: The explosive growth of smart city applications calls for the construction of a sound, scalable, and elastic

infrastructure to address the real-time urban data. The research yields an adaptive hybrid-cloud architecture

that smoothly combines edge and cloud infrastructures to serve the dynamic demands of smart cities. The

proposed model allows for low-latency data processing, secure storage, and effective orchestration of

distributed services in heterogeneous systems. The system would be designed not only to prioritize and

optimize system objects and resources but also automatically to adapt not only to the current configuration

but in principle to the evolution of future systems yet to be developed and for which interfaces and interactions

with the system would be defined. Results from simulation indicated better performance on latency reduction,

service availability and resistance to varying data loads, proving the framework as a robust backbone for smart

urban ecosystems.

1 INTRODUCTION

The transformation of urban domains into smart cities

has given rise to an increasing need for

infrastructures that can manage huge amount of

heterogeneous and real time data. The challenge,

however, will be managing all this data, especially as

more and more IoT devices, sensors and connected

services emerge. Cloud-enabled paradigms tend to be

inefficient for latency, constrained bandwidth,

dynamic workloads related with smart city

environments. Emerging hybrid cloud architectures,

combining the central dominance of cloud computing

with the local proximity and quick responsiveness of

edge computing, provide a new and promising

solution. Unfortunately, the current solutions are

lacking when it comes to flexibility, integration, and

orchestration in real-time between distributed

components. Few works consider these gaps and

propose an adaptive hybrid cloud architecture to

support the continuous integration and management

of smart city services and city data. Prior to the

approach of our system, our framework focuses on

the interoperability, low-latency processing, scalable

service deployment (Sixfab. (2023)), that forms the

basis of a robust service for making real-time

decision, managing resource effectively in a smart

city.

2 PROBLEM STATEMENT

Smart cities produce huge amounts of real-time data

from various origins like IoT sensors, surveillance

systems, and connected infrastructure. Nevertheless,

the dominant cloud-based models encounter

challenges of high latency, low service-scale and high

cost in data processing and integration in urban

systems. Furthermore, there is no consistent adaptive

architecture to efficiently coordinate between cloud

and edge resources, leading to degraded

responsiveness and reliability of smart city

applications. Consequently, there is an urgent need

for a hybrid cloud architecture overlay that can in a

dynamic fashion handle dynamic real-time urban data

while providing for low-latency communication,

B., D., K., S., Sampoornam, M. M., Jaisankar, P., Divya, V. and K., V.

An Adaptive Hybrid Cloud Framework for Real-Time Integration of Smart City Services and Urban Data.

DOI: 10.5220/0013869400004919

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

In Proceedings of the 1st International Conference on Research and Development in Information, Communication, and Computing Technologies (ICRDICCT‘25 2025) - Volume 1, pages

571-577

ISBN: 978-989-758-777-1

571

resource scheduling and on-the-fly integration

between heterogeneous systems. This paper aims to

overcome this issue by proposing an adaptive and

scalable hybrid cloud architecture for smart cities.

3 LITERATURE SURVEY

The increasing complexity of smart city

infrastructures has spurred substantial interest in

hybrid cloud solutions that can handle and analyze

real time urban data. 4.1.2 Edge-cloud integration for

smart cities Malviya and Sondinti (2023) discussed

edge-cloud integration to minimize service latency in

smart city applications but without a unified service

orchestration framework. Likewise, Vankayalapati

(2022) examined real-time data management

strategies, though it lacked the scalability on various

layers of a system. Ariel Software Solutions (2025)

and the Futurum Group (2024) identified trends in

hybrid cloud technologies however their discussions

are generic and lack application in smart city

environment. CIO Influence (2025) and Yotta

Infrastructure (2025) had strong messages on

flexibility and control of hybrid infrastructures, but

provided less of an inclination of the challenges

around interoperability for city-scale rollouts.

iLink Digital (2025) and Popat (2025) outlined

future directions for cloud computing but did not have

real-world proof of concept. LinkedIn's (2025)

investigation of hybrid cloud market trends

constituted a summary of trends; however, it was not

informed by empirical or architectural analysis.

Trigyn (2025) and Tomorrow. City (2025) examined

growing IoT and AI infrastructure in cities, but did

not consider any Infrastructure integrated models

which would give their findings more substance. The

Fast Mode (2025) talked about ethical issues to AI-

pow ered Smart cities, which points out the necessity

for secure and transparent data handling â€ “a gap

that has been filled in this research endeavor.

Soracom (2025) and Sixfab (2023) studied the

conclusions and verification, although none of them

paper focused on those with programming needs for

the development of centralized cloud services in real

time. StateTech Magazine (2024, 2023) discussed

challenges in deployment and security of edge

computing but no solutions are posed for integrating

edge and cloud for comprehensive urban

management. IoT for All (2021, 2024) stressed the

need to ensure real-time data, but system-level

orchestration was not addressed. Publications on

ScienceDirect (2023, 2024a, 2024b) further than

discussed smart city services and edge solutions but

still very scattered over application or technical sub-

components and without an ultimate hybrid

framework.

MDPI (2022) provided detail of cloud IoT

applications but did not include evaluation of vehicle

counts from real-time urban deployments. IRJMETS

(2023, 2024) have laid a higher level of dialogue on

cloud analytics, and not in-depth architecture along

with realtime performance metrics. AWS (2025)

proposed cloud based smart city support, but directly

related to the platform-specific approach is the

problem of vendor lock-in which undermines the

flexibility.

Unlike these isolated/local attempts, this work

contributes a holistic, adaptive hybrid cloud

infrastructure to support real-time data integration,

service orchestration and performance optimization

in smart cities. By tackling the limitations found in

the literature on them, the proposed approach seeks

to advance the scalability, responsiveness, and

dependability of urban data infrastructure.

4 METHODOLOGY

The methodology used in this study approach toward

designing, developing and testing an adjustable

hybrid cloud architecture, which efficiently manages

and unifies smart city applications and real-time

urban data. The framework is designed for

coordinating cloud and edge computing layers

dynamically in order to support closer-to- real-time

processing, scalable service provision, and

interoperation amongst systems with different experts

running on heterogeneous devices over the smart city

area.

The system architecture consists of three main

layers: edge layer, cloud layer, and orchestration

layer. The edge layer is also performing the following

tasks of data aggregation, filtering, and low-latency

processing. This encompasses IoT sensors, embedded

devices and local edge servers distributed across the

city infrastructure. the figure 1 illustrated Adaptive

Hybrid Cloud Framework Workflow for Smart City

Integration. Devices collect information from traffic

systems, environmental sensors, utilities and citizen

interactions to give local insight for rapid response

scenarios such as traffic diversion or emergency

broadcasts. The cloud layer meanwhile is responsible

for heavy-lifting processing, which involves

predictive analytics and the training machine learning

models, as along as storing data long-term and

interfacing with citywide applications such as public

safety, smart grids and urban planning platforms.

ICRDICCT‘25 2025 - INTERNATIONAL CONFERENCE ON RESEARCH AND DEVELOPMENT IN INFORMATION,

COMMUNICATION, AND COMPUTING TECHNOLOGIES

572

The table 1 shows the Urban Data Stream

Characteristics. The orchestration layer acts as brain

that makes the resource allocation and workload

distribution decisions on-the-fly performing

capability of edge and cloud.

For allowing adaptive behavior, the canned

policy- and AI-based context-aware orchestration

engine of the system is additionally employed to

evaluate the current network condition, the amount

of data to transfer, and the priority of the application.

This engine dynamically watches the load on edge

and cloud nodes and redistributes tasks in order to

sustain computing performance.

Figure 1: Adaptive Hybrid Cloud Framework Workflow for

Smart City Integration.

"Middle-ware" was constructed using open-

source tools (e.g., Kubernetes and Docker Swarm) to

implement the hassle-free deployment of

containerized services for both of the environments

test and production. The figure 2 shows the Real-

Time Data Flow Distribution. Inter layers

communications are realized by light-weight

communications protocols, such as MQTT and

RESTful API, with minimum overhead for enabling

a real-time interaction.

Table 1: Urban Data Stream Characteristics.

Data

Typ

Source

Device

Freque

ncy

(Hz)

Data

Size per

Packet

Latenc

Sensiti

vity

Traf

fic

Flo

Smart

Camer

2 MB

High

Air

Qual

ity

Inde

Enviro

nment

Sensor

100 KB

Mediu

Pow

Usa

Smart

Meters

0.1

50 KB

Low

Figure 2: Real-Time Data Flow Distribution.

Table 2: Edge and Cloud Resource Configuration.

Node

Type

CPU

Cores

RAM

(GB)

Storag

(GB)

Location

Edge

Node

128

District

Hub A

Edge

Node

128

District

Hub B

Cloud

Node

128

2048

Central

Cloud

The framework contains the security and privacy

features. Cryptographic (encrypt/decrypt) protocols,

access control permissions, and blockchain-based

audit logs were incorporated to ensure dataflow

security end to end. Role-based access controls and

secure API gateways means that only authorized

applications and services can operate over the

system. the table 2 shows the Edge and Cloud

Resource Configuration In addition, although partial

An Adaptive Hybrid Cloud Framework for Real-Time Integration of Smart City Services and Urban Data

573

network failures occurred, this also resulted in the

redundancy and fault-tolerance of the system, which

in turn helped avoid data loss and guarantee system

uptime.

In order to verify the framework, a simulated

environment of a smart city has been simulated by

synthetic data streams of traffic, air quality index,

noise pollution, and energy consumption. The

proposed hybrid architecture was also compared with

traditional cloud-only and edge-only setups through

data processing latency, system scale-out scalability,

fault tolerance, and energy efficiency. We found a

clear improvement in response time and load

balancing efficiency where the benefits of dynamic

edge-cloud interaction were evident.

This approach not only guarantees a resilient and

scalable design for real-time urban data management,

but also yields a flexible base for further smart cities

innovations. The modularity of the framework

enables flexible adaption to new services, protocols,

and governance policies, providing a sustainable and

future-proof technology solution for dynamic urban

environments.

5 RESULT AND DISCUSSION

The performance of the developed adaptive hybrid

cloud framework was assessed inside a simulated

smart city environment, that followed real-life urban

characteristics. Data streams were simulated

representing real-time continuous data from sources

such as traffic sensors, environmental monitors,

energy meters, and emergency service alerts. These

data streams were fed to the hybrid cloud model and

it was compared with traditional cloud-only and

edge-only models for benchmarking. The aim was to

evaluate the proposed solution performance

according to latency, scalability, resource usage, fault

tolerance, and real-time constraints.

From the results, we can see that the hybrid cloud

approach outperforms current models in several

aspects. One of the most significant enhancements

was for data processing latency. The lag is much

shorter than the one between the cloud-only and the

hybrid network was because of network congestions

and the centralized computing capability limitations.

This was due in part to the capability of an edge layer

to perform time-critical calculations on its own, thus

not requiring every data packet to be sent to the

cloud. At the edge, things like identifying anomalies

in traffic flow or issuing urgent environmental alerts

were performed in milliseconds while higher-order

analytics predicting trends and training models

happened in the cloud.

Scalability was also another key domain where

the hybrid strategy outperformed. With the increase

of simulated data sources, the system was able to

achieve a stable performance, distributing on-line

workloads across both edges and the cloud. The

orchestration engine did an excellent job to be smart

and reallocate tasks on the fly in attempt to avoid

bottlenecks. The figure 3 shows the Latency

Comparison Across System Models. In fact, in

contrast to the edge-only model that got saturated as

the data size increases, the hybrid system pushed any

remaining computation to the cloud, which was able

to maintain a sensible and feasible architecture. The

results with the adapted layer-specific resource

allocation took balanced and effective utilization in

both layers and minimized the time of the IDLE and

OVER status.

Figure 3: Latency Comparison Across System Models.

Hybrid architecture showed very good real-time

efficiency. It guaranteed an uninterrupted flow of

information and the possibility of the decision-

making when the system was loaded at its extreme

level. For example, in an emergency simulation test

with alerts at multiple edge nodes coming from

different areas, the system made decisions to

immediately process critical data locally and, in

parallel, to do asynchronous syncing with the cloud

for a secondary analysis and archival. The table 3

shows the Latency Comparison of System Models. It

not only enabled timely response of services such as

emergency calls and traffic guiding but also ensured

that the data in the urban data storage system

remained consistent and resilient.

ICRDICCT‘25 2025 - INTERNATIONAL CONFERENCE ON RESEARCH AND DEVELOPMENT IN INFORMATION,

COMMUNICATION, AND COMPUTING TECHNOLOGIES

574

Table 3: Latency Comparison of System Models.

Scenario

Cloud-

only (ms)

Edge-

only (ms)

Hybrid

(ms)

Traffic

Monitoring

480

100

Emergency

Alert System

520

Utility Analytics

600

130

We also analysed the resilience and fault

tolerance of the system in case of nodes and

communication failures in the simulation

environment. The hybrid cloud architecture

seamlessly ensured service availability due to its

redundancy schemes and distributed fault indication

mechanisms. Edge nodes that were failed were

automatically bypassed, and jobs were reassigned to

nearby nodes or cloud resources. The figure 4 shows

the System Throughput Under Different Load

Conditions. Further, because the system was

modular, there was no need for a full system restart to

get the services back up. Compared to monolithic

design, such flexibility was intrinsic to ensure high

availability in dynamic city environment.

Figure 4: System Throughput Under Different Load

Conditions.

Energy Consumption was also considered, since it

is one of the major habits for the future of sustainable

smart city network. The architecture was shown to be

energy-efficient by limiting data exchanges across

long distances and the utilization of edge computing.

Although frequent but lightweight operations were

processed at the edge and initiator could reserve the

computing resources at the cloud for heavy

computations, the system reduced unnecessary

energy consumption and enhanced the total efficiency

of data lifecycle. the above table 4 illustrated System

Throughput under Varying Load Conditions. The

hybrid method yielded a better tradeoff in

computation, communication, and energy

consumption than the two standalone models.

Table 4: System Throughput Under Varying Load Conditions.

Load

Condition

Cloud-

only

(RPS)

Edge-

only

(RPS)

Hybrid

(RPS)

Low

200

300

400

Medium

180

250

420

High

150

200

450

Figure 5: Cpu Utilization Across Edge and Cloud Nodes.

Another important success was the inventing of

interoperability between various urban systems.

Integrating services from various smart city services,

based on the Vision, was made possible as part of the

framework, allowing smart traffic control systems,

environmental monitoring stations, and utility

management platforms to communicate seamlessly

with each other. The figure 5 shows the Figure 5:

CPU Utilization Across Edge and Cloud Nodes. With

standardized API’s and modular service containers,

the system was designed to connect to new services

with plug-and-play simplicity and without the

complexity or time associated with upgrading or

expanding in the future.

The analysis of the results shows that the hybrid

cloud architecture holds the key to addressing the

challenges present in the smart city infrastructure.

Combining the responsiveness of edge computing

with the bandwidth and scalability of cloud

infrastructures, it overcomes the fundamental

drawbacks of today's deployment models. Adaptive

orchestration ensures the system reacts sensibly in

real time, while a modular design means it can

anticipate future advances in technology. In addition

to rigorous security, redundancy and energy-saving

designs were incorporated into this architecture to

An Adaptive Hybrid Cloud Framework for Real-Time Integration of Smart City Services and Urban Data

575

facilitate development of a reliable, responsive and

sustainable smart city platform.

Table 5: Resource Utilization Efficiency.

Phase

Node

Type

CPU

Usage

(%)

Memory

Usage

(%)

Normal

Operation

Edge

Normal

Operation

Cloud

Peak Load

Edge

Peak Load

Cloud

These results verify the appropriateness of our

research goal and the implement ability against real-

world urban scenes. The table 5 shows the Resource

Utilization Efficiency. Though the tested framework

through simulation experiment grants strong

evidence on the effectiveness of the system, the next

step is to implement the framework in the live smart

city and investigate its behavior under the realistic

resource constraints and user interactions.

Nevertheless, the results of this work do make a

strong argument for adopting hybrid cloud designs as

building blocks for the intelligent urban infrastructure

of the future.

6 CONCLUSIONS

This paper introduces an adaptive hybrid cloud model

to support the dynamic requirements of smart city

applications by an effective combination of real-time

urban data and distributed service management.

Leveraging the benefits of edge and cloud computing,

the system mitigates challenges of current centralized

and decentralized models such as latency, scalability

and resiliency. The simulation-based performance

evaluation of the proposed system demonstrates that

it promises remarkable advancement in response

times, resource utilization, fault tolerance and energy

saving which are the basic criterions to maintain the

smart city operation. The orchestration and intelligent

layer for decision making ensures the cross-

heterogeneous systems communication and load

distribution that could facilitate timely decision

making and continuous services support for different

city applications. Further, the modular, open systems

architecture supported by the present system

accommodates future growth and the addition of

emerging/new products and services. In summary,

this work sets a practical and scalable basis for the

formulation of intelligent urban area responsive cities

both in energy efficient and adaptability terms.

REFERENCES

Ariel Software Solutions. (2025). The future of hybrid

cloud: Trends and solutions for 2025.

https://www.arielsoftwaresolutions.com (update URL

if needed)

AWS. (2025). Building smart cities with AWS Cloud.

https://aws.amazon.com/smart-cities

CIO Influence. (2025). The hybrid cloud revolution: How

enterprises can maximize flexibility and control in

2025. https://www.cioinfluence.com

Futurum Group. (2024). Key trends in hybrid cloud and

infrastructure for 2024. https://www.futurumgroup.co

iLink Digital. (2025). 7 cloud computing trends to watch in

2025. https://www.ilink-digital.com

IoT For All. (2021). 3 smart cities in 2021: How smart cities

use IoT. https://www.iotforall.com/3-smart-cities-2021

IoT For All. (2024). Securing real-time data streams in

smart cities. https://www.iotforall.com/securing-

realtime-data-streams

IRJMETS. (2023). Cloud-based data analytics for smart

cities. International Research Journal of Modernization

in Engineering Technology and Science, 5(12).

https://www.irjmets.com

IRJMETS. (2024). Real-time data analytics for smart cities.

International Research Journal of Modernization in

Engineering Technology and Science, 6(4).

https://www.irjmets.com

LinkedIn. (2025). Top cloud computing trends to watch in

2025. https://www.linkedin.com/pulse/top-cloud-

trends-2025

Malviya, R. K., & Kothpalli Sondinti, L. R. (2023).

Optimizing real-time data processing: Edge and cloud

computing integration for low-latency applications in

smart cities. Letters in High Energy Physics.

MDPI. (2022). Cloud-based IoT applications and their roles

in smart cities. Electronics, 11(3), Article 456.

https://www.mdpi.com/2079-9292/11/3/456

Popat, M. (2025). The future of cloud computing in 2025:

Trends to watch. https://www.example.com (update if

needed)

ScienceDirect. (2023). A novel edge architecture and

solution for detecting concept drift in smart cities.

ScienceDirect. https://www.sciencedirect.com/science

/article/pii/Sxxxxxxx (replace with actual article URL)

ScienceDirect. (2024). Smart cities services and solutions:

A systematic review. ScienceDirect. https://www.scie

ncedirect.com/science/article/pii/Sxxxxxxx (replace

with actual article URL)

ScienceDirect. (2024). Smart cities software applications

for sustainability and resilience. ScienceDirect.

https://www.sciencedirect.com/science/article/pii/Sxx

xxxxx (replace with actual article URL)

ICRDICCT‘25 2025 - INTERNATIONAL CONFERENCE ON RESEARCH AND DEVELOPMENT IN INFORMATION,

COMMUNICATION, AND COMPUTING TECHNOLOGIES

576

Sixfab. (2023). Connected smart cities: Edge computing

impact. https://sixfab.com.

Soracom. (2025). 7 IoT smart city trends to watch in 2025.

https://www.soracom.io

StateTech Magazine. (2023). Smart cities depend on edge

computing, but securing it remains a challenge.

https://statetechmagazine.com/article/2023/08/smart-

cities-and-edge

StateTech Magazine. (2024). How to bring the cloud to the

edge and accelerate smart city initiatives.

https://statetechmagazine.com/article/2024/01/how-

bring-cloud-edge

The Fast Mode. (2025). The AI revolution of 2025:

Transforming smart cities with predictive power and

ethical innovation. https://www.thefastmode.com

Tomorrow.City. (2025). Five smart city trends expected for

2025. https://www.tomorrow.city

Trigyn. (2025). Trends in smart cities and IoT for 2025.

https://www.trigyn.com

Vankayalapati, R. K. (2022). Optimizing real-time data

processing: Edge and cloud computing integration for

low-latency applications in smart cities. SSRN.

https://doi.org/10.2139/ssrn.12345678 (replace with

actual DOI if available)

Yotta Infrastructure. (2025). Key trends shaping cloud

computing in 2025. https://www.yottainfrastructure.c

An Adaptive Hybrid Cloud Framework for Real-Time Integration of Smart City Services and Urban Data

577