BESIEC: An Adaptive Optimized Model for Task Scheduling & Offloading

Abstract

The rapid change in computational strategy and delay-sensitive applications require intense power sources of computational resources. This creates a challenge of precise latency requirements in 5G network services. Task scheduling and offloading can be promising solutions to achieve high-performance optimized output with heterogeneity, handling of tasks, conservation of energy, and reliable latency factor. Blockchain (BC), Software defined networks (SDN) and the Internet of Things(IoT) are the most promising significant technologies researched in this article, and the fusion of the three has the potential to reinvent the relationship of trust in the networks and promote the integration of confidentiality and reliability in the respective use cases. Cloud infrastructure is used to provide clients with powerful computing and storage environments. A kind of expansion of cloud computing architecture, edge computing, has been trending. Now, it is used to build distributed secure architecture to promote the safety and integrity of data throughout its lifetime and bring much-needed efficiency to IoT data processing. This paper considers a Blockchain-Enabled Software-defined network-based IoT Edge Cloud(BESIEC) scenario for data integrity during task scheduling and offloading processes while achieving optimal computational resources and minimizing end-to-end delays. The strategy shows that it is better to implement the BESIEC model than the Traditional Floodlight implementation. The BESIEC model shows better time consumption performance regarding the number of tasks completed compared to local processing, cloud offloading, and edge offloading.

Introduction

IoT nodes often need more computational resource management, including wireless communication processing units and storage and memory management. These are also highly limited and consume a lot of energy, which is inefficient in meeting the tremendous network growth with heterogeneous structural objects. To overcome heterogeneity situations, cloud infrastructure can be taken as an effective tool to use.

1. Remote cloud server for offloading complex tasks

2. Releasing computing burden

3. Effective task handling

4. Reduce task response time

5. Reduced energy consumption of IoT devices.

However, applications with delayed timelines have better choices than the cloud platform [1]. These applications require high-end computational services and storage to increase the latency bandwidth usage [2]. Various issues must be dealt with to satisfy the efficiency and effectiveness in real-time. It is better to rely on a hybrid computing environment (Edge-cloud) rather than a centralized one [3]. Edge computing allows migration of computing tasks, including edge server management, to enhance the adequate capability of resource shortage while reducing communication cost and latency during the task offloading process.

The solutions to the computational complex tasks can't solve the real-time challenges of optimal offloading decision parameters. Due to the heterogeneous nature of the network, it is difficult for the edge cloud platforms to offload the required data. So, a fusion of technologies or a hybrid computing environment is needed to effectively and efficiently offload tasks. Intelligent network and resource management are vital in providing appropriate solutions in an IoT-enabled environment [4]. Cloud, network edge, and SDN are popular approaches to allocating resources dynamically and effectively [5].

It can deal with traditional network challenges and distribute traffic geographically over IoT networks. Application task offloading, in general, is an unsafe approach because of the nontransparent environment, which results in data loss and privacy leakage. A potential solution needs to be introduced for offloading. Block-chain-enabled secure features can ensure data.

To this end, the current study proposed BESIEC, an SDN-based IoT edge and cloud model that supports blockchain, where SDN involves flexible resource management and intelligent network control. Edge-Cloud servers ensure low-latency computing and powerful computational latency by maintaining data integrity and preventing illegal offloading behaviours. A smart contract is enabled with the SDN controller to collect the information for resource management. Smart contracts enforce predefined rules and conditions for task scheduling, ensuring fairness and efficiency in resource allocation. It also allows the SDN controller to allocate available resources based on predefined scheduling algorithms, optimizing resource utilization and task performance. A signature-based approach is also used to handle IoT device abnormality.

The significant contributions of this paper are:

• Contribution 1: A Novel model based on SDN and Edge-cloud n/w that supports BC.

• Contribution 2: Adaptive task scheduling and offloading algorithm.

• Contribution 3: Signature-based detection approach.

• Contribution 4: A Smart contract for better resource management in IoT edge n/w.

• Contribution 5: An effective algorithmic implementation to show the optimized output.

Related Works

Heterogeneity leads to many security threats. Efficient task scheduling and offloading mechanisms can be a promising solution to these challenges. Ensuring data integrity during the task scheduling and offloading process while achieving optimal computational resources and minimizing end-to-end delays are also stringent requirements.

IoT, blockchain edge computing, and different AIML technologies are the most significant integrated systems used in various industry applications. These integrated technologies are also used in smart cities for traffic management [6, 7]. Some researchers have also implemented deep reinforcement learning for traffic management in the IoT scenario [8].In the cloud environment, the main concern is to deploy security management. This security can be achieved by implementing a blockchain in an edge cloud network [9]. The QoS services are somewhat embedded with the genetic algorithm [10]. The most significant part of this security level checking is to measure the appropriate parameters and guidance to enhance security in a 5G network environment. This can be implemented by integrating blockchain, SDN, and ML algorithms [11].The handling of the tasks is the measure concern in FOG or cloud-based scenarios [12, 13]. In these cases, the main parts are task assignment, scheduling, and offloading wisely. The difficulties are to be handled in a mobile edge computing environment [14, 15].

Researchers also used reputation-based proof-of-cooperation in supply chain management over the cloud [16] and Fair proof of reputation for blockchain technology [17].In security mechanisms, attack detection in the task scenario can be managed by SDN Flow or a collaborative deep-learning approach with SDN [18, 19].

The following parameter to manage is energy. Managing energy transactions to balance the system is an integral part of the architecture in a cloud or FOG computing environment. A decentralization blockchain framework or SDN-based architecture can be implemented for the same [20, 21].

Q-Learning algorithms are also used to locate the end devices to calculate the end-to-end delay transmission and connectivity in SDN-based IoT environments [22, 23]. In the above, different applications are discussed, and other informative sections are related to cloud, edge, and IoT. However, in all the cases, task offloading is the primary concern in this article for the establishment of a significant connection with the different tasks and scheduling them for proper execution. Tasks can be offloaded with the help of learning algorithms or smart contracts [24, 25]. Numerous studies have been conducted on task scheduling and offloading strategies for verity of problems but the current studies based on in order to make decision as quickly and cheaply as possible by maintaining data integrity. The Traditional Floodlight system is also considered for the estimation [26].

The edge offloading process is implemented in certain scenarios like the smart grid approach [27] and the Internet of Remote Things [28]. Researchers [29] also indicate the cross-edge offloading process in green mobile edge computing environments. The cloud offloading approach is also used in a fog-based or traditional approach [30], a cloud-based IoT system [31].

The current research is based upon the task scheduling and offloading in Blockchain-Enabled Software-defined network based IoT Edge Cloud(BESIEC) environment. The study comprises the implementation of novel methods with algorithms, as well as implementation and comparison with the other available processes to showcase better execution in a subsequent section.

Task Scheduling and Offloading within the BESIEC Scenario

Task scheduling within SDN-enabled edge computing environments involves a multi-step process that leverages the capabilities of SDN controllers to allocate computing resources and manage network traffic efficiently. SDN enhances agility, scalability, and efficiency of diverse cloud and edge networks by centralizing management, optimizing resource utilization, improving traffic engineering, facilitating network virtualization, enforcing policies, and promoting interoperability. Modern applications and services in organizations are in growing demand due to the merging of cloud to edge computing technologies.

Integration of valuable applications, user devices, and IoT devices in the presence of a centralized server in a centralized environment is required to begin a task-scheduling process. The SDN controller is used for intelligent implementation throughout the process to measure performance constraints over the available network during the scheduling process. Then, the SDN controller decides to allocate resources to execute the task. Then, the edge node is assigned by the controller per the task's requirement. After the execution of the task, Quality of Service (QoS) management has to be done to calculate and optimize the performance and reliability.

After the successful task completion, the allotted resources have to be cleaned up along with the temporary configuration systems and change of states. These enhance the resources' performance and make them available for upcoming tasks. SDN's centralized control and environment of edge computing ensure optimized performance, reliability enhancements, application requirements, and traffic management.

Blockchain can be used with the SDN controller to record scheduling decisions, resource allocations, and scheduling processes.

Smart contracts which are self-executing codes with the terms of the agreement directly written into code, can enforce predefined rules and conditions for task scheduling, ensuring fairness and efficiency in resource allocation. They also allow an SDN controller to automatically allocate available resources based on predefined scheduling algorithms, optimizing resource utilization and task performance.

Model Formulation, Algorithm Analysis

This section is divided into.

1. System architecture

2. Enhanced SDN Based Model for task offloading

3. Signature-based detection

4. Effective problem statement

5. Algorithm

System Architecture

Figure 1 shows the architectural design of BESIEC, in addition to the task scheduler and BC-enabled flow rule at the SDN controller level, to manage edge resources, track and maintain consistency, protect data, and minimize power consumption. Blockchain can be used to record scheduling decisions, task assignments, and resource allocations, which ensures trust and accountability among different entities involved in the scheduling process.

Fig. 1 [Images not available. See PDF.]

BESIEC Framework

The SDN controller collaborates with a smart contract and task scheduling and offloading algorithm to secure schedule checking and intelligent network resource scheduling and management in a secure manner while ensuring fairness and efficiency in resource allocation.

Enhanced SDN Based Model for Task Offloading

In a BC-enabled SDN-based framework, resource management starts with network infrastructure monitoring to gather data on metrics like traffic flow, bandwidth usage, link capacities, and device statuses. This information is acquired by SDN controllers from switches, routers, and other network devices. SDN frameworks come equipped with monitoring and analytic tools that constantly observe network performance, analyze resource usage, and identify anomalies or bottlenecks. These insights inform adjustments to resource management strategies, enhancing network efficiency and performance. Blockchain can be used to record scheduling decisions, task assignments, and resource allocations. Blockchain can facilitate decentralized task allocation by assigning available resources, such as computing nodes, storage, and network bandwidth, on the blockchain. Smart contracts can then automatically allocate these resources based on predefined scheduling algorithms, optimizing resource utilization and task performance. Blockchain can facilitate dynamic adaptation in task scheduling by enabling real-time updates to resource availability and scheduling decisions, ensuring agility and responsiveness to evolving requirements. Integrating smart contracts into SDN controllers can help increase efficiency by automating resource allocation and task scheduling, improving transparency and accountability, accelerating decision-making, enhancing security, and enabling dynamic adaptation to changing network conditions.

The SDN controller uses a scheduling algorithm to manage task-processing requests and provide information that can be used for efficient resource utilization.

Figure 2 shows internal controller modules:

1. Task scheduler module.

2. Handler module.

3. N/W monitoring and computing module.

Fig. 2 [Images not available. See PDF.]

5G Challenges and emerging technology

The task scheduler module is responsible for improved model performance and efficiency.

The task scheduler module is responsible for improved model performance and efficiency as well as enhanced ability to find optimal or near-optimal solutions.

The smart contract handler module processes the task request forwarded by the smart contract and also updates the status of task requirements. It efficiently manages and allocates resources in a decentralized manner and monitors the resource utilization of network devices such as bandwidth, CPU and memory. It ensures that tasks are scheduled in a way that optimizes resource utilization and prevents network congestion. This module helps SDN controllers increase efficiency by automating resource allocation and task scheduling, improving transparency and accountability, accelerating decision-making, enhancing security and enabling dynamic adaption to change network conditions. Call() and Feedback() are used in the mode of communication between the SDN-Controller and smart contract.

The network monitoring and computing module detects network faults in real time and initiates appropriate recovery mechanisms. It reroutes traffic, reallocates resources or applies load balancing techniques to mitigate the impact of failures and ensures interrupted task execution.

These components work together within the SDN controller to efficiently schedule and manage network policies in accordance with the requirements of the SDN deployment.

IoT systems can effectively assign tasks to edge nodes, optimize resource utilization, and meet application-specific goals while operating efficiently in dynamic and heterogeneous environments. In simplifying the Complexity of the problem into a single objective problem and reducing the difficulty of solving, we consider the following hypotheses:

By considering the following hypotheses should be considered while edge nodes are assigned tasks by the IoT system to meet application-specific goals.

1. Dynamic environment requirements should be taken into account before node assignment.

2. To ensure correct execution task, dependencies should be ignored

3. Chose optimal path nodes for assignment of tasks

4. Task assignment must consider available resources to balance overloading situations.

5. Each task should get a chance to offload at least once

✓Task offloading to Edge/Cloud: A task will be offloaded to Edge or Cloud depending on the data transmission time and execution time. The SDN controller tries to optimize the transmission time, whereas all transactions are kept and maintained by smart contract. The edge and cloud server list, as well as available resources vs requested resources, are updated in the record. The cost of these servers is calculated and compared with the available predefined cost and the minimum cost server is selected for offloading, i.e. if the available is less than the requested resources or the cost of all servers is greater than the internal tokens of the device, then the task is offloaded to the cloud otherwise to edge server.

Signature-Based Detection

This approach is used to develop and maintain a lightweight signature database that targets various possible threats encountered by the environment. Analyzing the behaviour and integrating techniques for identifying attacks detects deviations from the original behaviour and targets emerging threats (Figs. 3 and 4).

Fig. 3 [Images not available. See PDF.]

Working Model

Fig. 4 [Images not available. See PDF.]

Signature-Based Detection Approach

Problem Statement

The objective is to classify tasks in a heterogeneous environment by eliminating heterogeneity using an intelligent assignment system and processing units. The classification component includes three option values: average, accumulated average, and average number of tasks in the queue. The total number of tasks is divided into short and long, as per the threshold. Then, as per the requirement, the tasks are assigned to the available low or fast processor to complete the process in minimum execution time.

Let , T = T₁,T₂………Tn are assigned over edge nodes E=E₁,E₂………E_M

The minimization of energy consumption and latency can be achieved by assigning N different tasks over M different edge devices represented by

1

$$T_1,T_2\dots \dots \dots \mathit{Tn}=E_1,E_2\dots \dots \dots E_M$$

Assignment of task T_i to E_j also optimizes the use of transformation channels. The assignment of Task Ti on edge Ej with request of time with respect of is denoted as.

Y_ij[t] and provides two values

2

$$Y_{\mathit{ij}}=\left\{\begin{array}{cc}1\hfill & \text{if the task runs on edge}\\ 0\hfill & \text{otherwise}\end{array}\right)$$

Latency time and execution time are represented when the assigned task runs on edge device

3

$${\mathit{TEC}}_{\mathit{ij}}\left(t\right)=\mathit{Ii}/\mathit{Cj}$$where TEC_ij is the task execution time with processing rate at edge node Ej and Ii is the number of instructions of Ti over time.

Or

4

$$\frac{Noofinstructionsoftask\left(\mathit{Ti}\right)}{CPUprocessingrate}$$

5

$$\mathit{Ti}\to \mathit{Ej}$$

To run the experimental setup, which consists of 15 access points, 10 edge servers, tasks from 50 to 400, task size 500 KB, edge server frequency 2 to 4 GHz, and IoT devices frequency 20 to 80 GHz, Mininet-Wifi is used as a simulator, and Solidity using Remix IDE is used for the smart contract. The Web3 protocol has been used to complete the task and interface.

It is assumed that all the parameters for the system architecture are SDN-enabled. The computational value for each server($c_i^v$) integrated to final value($\mathit{CV}$)

6

$$CV=\left\{c_1^v,c_2^v\dots \dots \dots \dots \dots c_{E_s}^v\right\}$$

Let the IoT device X have I/o operations to compute the task capability with the data input and output values(${{\mathit{data}}_i^{\mathit{in}},data}_i^{\mathit{out}}$) with a max value of time execution.

7

$$Ti=\left\{\left\{dat{a}_i^{in},\right)dat{a}_i^{out},t_{max}\right\}$$

The processing time to execute($P_i^{E_s}$) the whole task can be calculated as the sum of the data transmission($d_i^{\mathit{tr}}$) and execution time ($t_i^{\mathit{ex}}$).

8

$$P_i^{\mathit{Es}}=d_i^{\mathit{tr}}+t_i^{\mathit{ex}}$$

The data transmission time($d_i^{\mathit{tr}}$) in this case is where ${\mathit{data}}_{i\to Es}^{\mathit{in}}$ data transmission from input to each server and ${\mathit{data}}_{Es\to i}^{\mathit{out}}$ data transmission from each server as output.

9

$$d_i^{\mathit{tr}}=dat{a}_{i\to Es}^{\mathit{in}}+bdat{a}_{Es\to i}^{\mathit{out}}$$

The total time required to complete the whole task (${\mathit{CT}}_i^{\mathit{com}}$) is the sum of data transmission time($d_i^{\mathit{tr}}$),data transmission time due to gateway GW($d_{\mathit{GW}}^{\mathit{tr}}$) and execution time ($t_i^{\mathit{ex}}$).

10

$$C{T}_i^{\mathit{com}}=d_i^{\mathit{tr}}+d_{\mathit{GW}}^{\mathit{tr}}+t_i^{\mathit{ex}}$$

11

$$t_i^{\mathit{tran}}=d_{i\to GW}^{dat{a}_{\mathit{in}}}+d_{GW\to i}^{dat{a}_{\mathit{out}}}$$

12

$$d_{\mathit{GW}}^{\mathit{tr}}=t_{GW\to cloud}^{dat{a}_{\mathit{in}}}+t_{cloud\to GW}^{dat{a}_{\mathit{out}}}$$

Algorithm Implementation

The SDN controller collects state information from the fog node to make task-scheduling decisions. The controller communicates with BC n/w for flow verification and validation. The smart contract and SDN controller collaborate in task scheduling, and the BC node assigns tasks to the fog node and sends flow rules to the SDN data plane. The smart contract ensures data integrity during the task scheduling and offloading process.

Steps of block verification and validation:

• Step 1: The SDN controller sends requests to all nodes of the BC system. Only one node will be chosen and act as an ordered node.

• Step 2: All peer nodes of the network get a message from the blockchain node.

• Step 3: Subsequently, each blockchain node sends a message to all other pre-qualified nodes from which it will receive messages.

• Step 4: Finally, the SDN controller receives an acknowledgement from all nodes, including the Ordered node.After getting confirmation that all transactions are validated, the distributed ledger updates records by adding the newly validated block.

Performance Evaluation

The performance evaluation is carried out by the estimation of the output value and compared with the required resources. The output graphs in Figs. 5, 6 and 7 shows the best performance of BESIEC model in all the categories.

Fig. 5 [Images not available. See PDF.]

Tasks vs Flow completion time

Fig. 6 [Images not available. See PDF.]

Number of task vs Time consumption

Fig. 7 [Images not available. See PDF.]

Number of requests vs time consumption with and without BESIEC Model

The performance evaluation metrics are categorized into three parts.

1. Flow completion time with traditional Floodlight and BESIEC model

2. Time required to complete the task—Comparison approach

3. Request management with and without BESIEC

I Flow Completion Time with Traditional Floodlight and BESIEC Model

Figure 5 shows the performance outcome of the flow completion time with the traditional Floodlight and BESIEC model. The number of tasks and flow completion time are given in X and Y axis respectively. The output shows that the flow completion time is much less with the BESIEC model than the traditional Floodlight model.

II Time is Required to Complete the Task Using the Comparison Approach

Figure 6 shows the number of tasks to the time consumption output, where the number of tasks is shown in the X axis while time consumption is shown in the Y axis. The proposed model is compared by local processing, cloud offloading and edge offloading. The output indicates that the time completion in BESIEC is much less as compared to others.

III Request Management with and Without BESIEC

The Fig. 7 shows the number of requests vs Time consumption graph. The number of requests shows in X axis and time consumption in Y axis. The output shows the the number of handling requests with BESIEC is more efficient than handling number of request without BESIEC.

Conclusion

This article comprises the development of task assignment and offloading mechanisms for the SEN-enabled network with the help of IoT networks and Smart contracts. The formulated task assignment and offloading problem ensures a data integrity during the task scheduling and offloading. It is achieved by the optimal computational resources and minimization of end-to-end delays. The proposed algorithm enhances the optimal allocation of task assignments in real-time scenarios. The proposed approach can be performed in both local and global optimization by ensuring lower latency communication with optimized increasing energy efficiency. The proposed algorithm can be implemented in anonymous edge networks with efficient resource utilization, real-time measurement of performance metrics and optimal process of task offloading. This integration strategy with technology and edge cloud architecture can be used to enhance task offloading in the 5G network by enabling the use of modern applications and services with seamless user experiences.

The future scope of this approach to establish an and deploy Edge-Fog AI models in the real time environment for the data processing by creating a collaborative environment where task can be offloaded to both the edge nodes and intermediate fog nodes. This process may reduce the need of large amount of data transmission to the central cloud.

Author Contribution

Ms Jayashree Mohanty has developed different architecture, done experimental tests, and resultant output, and Dr. Srichandan Sobhanayak has guided in all respects.

Funding

We have no research funding.

Data Availability

As per requirement, we will provide our data.

Declarations

Conflict of Interest

We have no conflict of interest.

Informed Consent

Since we have developed our model, there is no need to obtain consent from others.

Research Involving Humans and /or Animals

We are human only for research.

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

References:

1. Mirzapour-Moshizi, M; Sattari-Naeini, V; Molahosseini, AS. Application placement in fog-cum-cloud environment based on a low latency policy-making framework. Clust Comput; 2024; 27, 1 pp. 199-217. [DOI: https://dx.doi.org/10.1007/s10586-022-03954-z]

2. Magotra, B; Malhotra, D; Dogra, AK. Adaptive computational solutions to energy efficiency in cloud computing environment using VM consolidation. Arch Comput Methods Eng; 2023; 30, 3 pp. 1789-1818. [DOI: https://dx.doi.org/10.1007/s11831-022-09852-2]

3. Yuan, H; Bi, J; Wang, Z; Yang, J; Zhang, J. Partial and cost-minimized computation offloading in hybrid edge and cloud systems. Expert Syst Appl; 2024; 250, [DOI: https://dx.doi.org/10.1016/j.eswa.2024.123896]

4. Jeyaraj, R; Balasubramaniam, A. Resource management in cloud and cloud-influenced technologies for internet of things applications. ACM Comput Surv; 2023; 55, 12 pp. 1-37. [DOI: https://dx.doi.org/10.1145/3571729]

5. Kanellopoulos, D; Sharma, VK. Dynamic load balancing techniques in the IoT: a review. Symmetry; 2022; 14, 12 2554. [DOI: https://dx.doi.org/10.3390/sym14122554]

6. Lilhore, UK; Imoize, AL; Li, CT; Simaiya, S; Pani, SK; Goyal, N; Kumar, A; Lee, CC. Design and implementation of an ML and IoT based adaptive traffic-management system for smart cities. Sensors; 2022; 22, 8 2908. [DOI: https://dx.doi.org/10.3390/s22082908]

7. Hazarika, A; Choudhury, N; Nasralla, MM; Khattak, SBA; Rehman, IU. Edge ML technique for smart traffic management in intelligent transportation system. Access; 2024; [DOI: https://dx.doi.org/10.1109/ACCESS.2024.3365930]

8. Jarwan, A; Ibnkahla, M. Edge-based federated deep reinforcement learning for IoT traffic management. IEEE Internet Things J; 2022; 10, 5 pp. 3799-3813. [DOI: https://dx.doi.org/10.1109/JIOT.2022.3174469]

9. Ahmed, A; Abdullah, S; Iftikhar, S; Ahmad, I; Ajmal, S; Hussain, Q. A novel blockchain based secured and QoS aware IoT vehicular network in edge cloud computing. IEEE Access; 2022; 10, pp. 77707-77722. [DOI: https://dx.doi.org/10.1109/ACCESS.2022.3192111]

10. Karakus, M. GATE-BC: genetic algorithm-powered QoS-aware cross-network traffic engineering in blockchain-enabled SDN. Access; 2024; [DOI: https://dx.doi.org/10.1109/ACCESS.2024.3374213]

11. Rahman, A; Khan, MSI; Montieri, A; Islam, MJ; Karim, MR; Hasan, M; Kundu, D; Nasir, MK; Pescapè, A. BlockSD-5GNet: Enhancing security of 5G network through blockchain-SDN with ML-based bandwidth prediction. Trans Emerg Telecom Tech; 2024; 35, 4 [DOI: https://dx.doi.org/10.1002/ett.4965]

12. Sen, P; Pandit, R; Sarddar, D. A survey on methods and apparatus of offloading in mobile cloud computing. Intern J Sys Eng; 2024; 14, 2 pp. 212-225.

13. Sen, P; Islam, T; Pandit, R; Sarddar, D. A comparative review on different techniques of computation offloading in mobile cloud computing. Fog Comp Intel Cloud IoT Sys; 2024; [DOI: https://dx.doi.org/10.1002/9781394175345.ch2]

14. Liao, Z; Hu, W; Huang, J; Wang, J. Joint multi-user DNN partitioning and task offloading in mobile edge computing. Ad Hoc Netw; 2023; 144, [DOI: https://dx.doi.org/10.1016/j.adhoc.2023.103156]

15. Chen, J; Deng, Q; Yang, X. Non-cooperative game algorithms for computation offloading in mobile edge computing environments. J Par Distr Comput; 2023; 172, pp. 18-31. [DOI: https://dx.doi.org/10.1016/j.jpdc.2022.10.004]

16. Sarfaraz, A; Chakrabortty, RK; Essam, DL. Reputation based proof of cooperation: an efficient and scalable consensus algorithm for supply chain applications. J Ambient Intell Humaniz Comput; 2023; 14, 6 pp. 7795-7811. [DOI: https://dx.doi.org/10.1007/s12652-023-04592-y]

17. Zhang, T; Huang, Z. FPoR: Fair proof-of-reputation consensus for blockchain. ICT Express; 2023; 9, 1 pp. 45-50. [DOI: https://dx.doi.org/10.1016/j.icte.2022.11.007]

18. Buzzio-García, J; Vergara, J; Ríos-Guiral, S; Garzón, C; Gutiérrez, S; Botero, JF; Quiroz-Arroyo, JL; Pérez-Díaz, JA. Exploring traffic patterns through network programmability: introducing sdnflow, a comprehensive openflow-based statistics dataset for attack detection. IEEE Access; 2024; 12, pp. 42163-42180. [DOI: https://dx.doi.org/10.1109/ACCESS.2024.3378271]

19. Zhou, H; Zheng, Y; Jia, X; Shu, J. Collaborative prediction and detection of DDoS attacks in edge computing: A deep learning-based approach with distributed SDN. Comput Netw; 2023; 225, [DOI: https://dx.doi.org/10.1016/j.comnet.2023.109642]

20. Gawas, M. and Govekar, S., 2023. Energy efficient Qos aware decentralized blockchain framework for fog computing.

21. Roux, R; Olwal, TO; Chowdhury, DS. Software defined networking architecture for energy transaction in smart microgrid systems. Energies; 2023; 16, 14 5275. [DOI: https://dx.doi.org/10.3390/en16145275]

22. Huang, G; Ullah, I; Huang, H; Kim, KT. Predictive mobility and cost-aware flow placement in SDN-based IoT networks: a Q-learning approach. J Cloud Comp; 2024; 13, 1 26. [DOI: https://dx.doi.org/10.1186/s13677-024-00589-w]

23. Diro, AA; Reda, HT; Chilamkurti, N. Differential flow space allocation scheme in SDN based fog computing for IoT applications. J Amb Intel Human Comp; 2018; [DOI: https://dx.doi.org/10.1007/s12652-017-0677-z]

24. YLiu, T., Wu, J., Chen, L., Wu, Y. and Li, Y., ,. Smart contract-based long-term auction for mobile blockchain computation offloading. IEEE Access; 2020; 8, pp. 36029-36042. [DOI: https://dx.doi.org/10.1109/ACCESS.2020.2974750]

25. Li, J; Zhu, M; Liu, J; Liu, W; Huang, B; Liu, R. Blockchain-based reliable task offloading framework for edge-cloud cooperative workflows in IoMT. Inf Sci; 2024; 668, [DOI: https://dx.doi.org/10.1016/j.ins.2024.120530]

26. Floodlight. Accessed: Jan. 20, 2024. [Online]. https://github.com/floodlight/floodlight

27. Arcas, GI; Cioara, T; Anghel, I; Lazea, D; Hangan, A. Edge offloading in smart grid. Smart Cities; 2024; 7, 1 pp. 680-711. [DOI: https://dx.doi.org/10.3390/smartcities7010028]

28. Zhang, S; Liu, A; Han, C; Liang, X; Xu, X; Wang, G. Multiagent reinforcement learning-based orbital edge offloading in SAGIN supporting Internet of Remote Things. IEEE Internet Things J; 2023; 10, 23 pp. 20472-20483. [DOI: https://dx.doi.org/10.1109/JIOT.2023.3287737]

29. Fang, G; Zhou, R; Li, X; Huang, H; Li, Z. A cross-edge offloading framework for green MEC systems. IEEE Trans Cons Elect; 2023; [DOI: https://dx.doi.org/10.1109/TCE.2023.3319816]

30. Kar, B; Yahya, W; Lin, YD; Ali, A. Offloading using traditional optimization and machine learning in federated cloud–edge–fog systems: A survey. IEEE Com Survey Tutor; 2023; 25, 2 pp. 1199-1226. [DOI: https://dx.doi.org/10.1109/COMST.2023.3239579]

31. Wang, T; Liang, Y; Zhang, Y; Zheng, X; Arif, M; Wang, J; Jin, Q. An intelligent dynamic offloading from cloud to edge for smart iot systems with big data. IEEE Transact Network Sci Eng; 2020; 7, 4 pp. 2598-2607. [DOI: https://dx.doi.org/10.1109/TNSE.2020.2988052]