Software-Defined Networking - Science topic

Have you been able to implement the attack? I urgently need the implementation, please.

In 2023 and 2024, several comprehensive surveys have been conducted on DDoS detection and mitigation strategies within Software-Defined Networks (SDNs). These surveys address the evolving landscape of SDN security, focusing on both high-rate and low-rate DDoS attacks.

One notable survey from **Cluster Computing** provides an extensive overview of existing DDoS detection, mitigation, and defense mechanisms in SDNs. It categorizes various approaches, highlighting their effectiveness and the ongoing challenges in developing robust DDoS protection strategies for SDN environments【8†source】.

Another significant study in **Multimedia Tools and Applications** explores the taxonomy of DDoS defense mechanisms in SDNs, identifying key research challenges and opportunities. It emphasizes the need for adaptive and intelligent systems that can respond to the dynamic nature of DDoS attacks【9†source】.

Furthermore, an article in **IEEE Access** published in 2024 offers a detailed examination of DDoS detection methods specifically tailored for SDNs. This study highlights recent advancements and future directions, providing a crucial resource for researchers looking to enhance security within SDN architectures【10†source】.

These surveys collectively underline the importance of continuous innovation in DDoS detection and mitigation techniques to address the complexities introduced by SDNs.

Question 1: Hot Research Areas in Anomaly-Based DDoS Detection in SDN

1. Machine Learning and AI Techniques : Enhancing anomaly detection using machine learning algorithms to identify unusual traffic patterns, especially in dynamic environments.

2. Real-time Traffic Analysis : Developing methods for real-time monitoring of network traffic to ensure immediate response to anomalies.

3. Adaptive Thresholding : Researching adaptive algorithms that can adjust thresholds dynamically based on network behavior over time.

4. Hybrid Detection Systems : Combining multiple detection techniques (signature-based, anomaly-based, etc.) to improve detection rates and reduce false positives.

5. Behavioral Analysis : Analyzing normal user behaviors to create baseline profiles, which can then be monitored for deviations that indicate DDoS attacks.

6. SDN-specific Approaches : Exploring how SDN's centralized control can be leveraged for more effective anomaly detection, including flow management and real-time policy updates.

7. Coordination of Security Measures : Investigating the orchestration of multiple security measures (e.g., firewall, IDS) within SDN to improve response to detected anomalies.

Question 2: New or Prominent DDoS Attack Types

1. Low-rate DDoS Attacks : These attacks involve sending traffic at a low rate to avoid detection by traditional thresholds, aimed at overwhelming specific resources like applications or databases.

2. Application Layer Attacks : Focusing on DDoS attacks targeting application layers (e.g., HTTP floods) rather than just network resources, often harder to detect.

3. Amplification Attacks : Leveraging certain protocols (e.g., DNS or NTP) to generate massive amounts of traffic against a target.

4. Distributed Reflection DoS (DRDoS) : This type sends requests to multiple devices, which then reflect the traffic to the victim, compounding the attack’s impact.

5. IoT-based Attacks : Utilizing compromised Internet of Things (IoT) devices, which are often less secure, to orchestrate large-scale attacks.

6. Flow Table Overflow Attacks : Targeting the flow tables in SDN switches, causing resource depletion and disrupting normal operations.

Question 3: Is the Flow Table Overflow Attack Proposed Much Later?

Yes, the flow table overflow attack is a more recent consideration in the context of SDN security. As SDN technologies evolved, researchers began to identify vulnerabilities specific to the architecture, including how flow tables can be overwhelmed by excessive rules or flows. This type of attack leverages the unique characteristics of SDN, where centralized control and the reliance on flow tables can be exploited to disrupt services or degrade performance.

Murtadha Shukur

Yes, SDN and NFV can potentially enhance handover performance and decrease latency in various ways. For example, in an SDN-enabled edge computing environment, effective strategies for access point (AP) handover can be implemented based on the received signal strength indicator (RSSI) and traffic load. This can not only improve the time taken to complete tasks but also reduce energy consumption. Similarly, NFV enables the optimization of resource allocation and load balancing, resulting in smooth handovers and addressing the issue of high latency in wireless LANs. Furthermore, the joint optimization of resources in Mobile Edge Computing (MEC) systems can further minimize energy consumption and delays.

Dear Murtadha Shukur,

how Software-Defined Networking and Network Function Virtualization can be used for handover is shown in my Fig. 9 (Examples of supra-regional SD-IoT system solutions with controllers on the fog and cloud sublayers) at the address:

Best regards and much success

Anatol Badach

Dynamic Resource Allocation:

I can provide some guidance on how you can find potential collaborators for your research in the areas of Networking, Cloud, SDN, NFV, and Advanced Computing:

Academic Conferences and Workshops: Attend conferences, workshops, and seminars related to Networking, Cloud, SDN, NFV, and Advanced Computing. These events are great opportunities to network with researchers and academics who share similar research interests.

Research Groups and Labs: Explore research groups, laboratories, and centers at universities or research institutions that focus on Networking, Cloud Computing, SDN, NFV, and Advanced Computing. Reach out to researchers leading these groups to discuss potential collaboration opportunities.

Online Research Platforms: Join online research platforms and academic social networks like ResearchGate, Academia.edu, or LinkedIn. These platforms allow researchers to connect, collaborate, and share research ideas.

Collaboration Tools: Utilize collaboration tools like Google Scholar, Mendeley, or Zotero to discover research papers and identify potential collaborators based on their publications and research interests.

Professional Networking: Attend professional networking events, webinars, and virtual meetups related to your research areas. Engage with other researchers, share your research interests, and explore potential collaboration opportunities.

Collaboration Requests: Consider posting collaboration requests on academic forums, mailing lists, or research group websites to reach out to researchers who may be interested in collaborating on research projects.

By actively engaging with the academic community, networking with researchers in your field, and exploring collaboration opportunities through various channels, you can increase your chances of finding potential collaborators for your research in Networking, Cloud, SDN, NFV, and Advanced Computing. Good luck with your research endeavors!

¡Claro! El Software-Defined Networking (SDN) distribuido es un tema avanzado en el campo de las redes de computadoras que implica la distribución de la lógica de control en múltiples controladores SDN en lugar de tener un controlador centralizado. Aquí hay algunas sugerencias y orientaciones para ayudarte con SDN distribuido:

Recuerda que el SDN distribuido es un campo complejo y en constante evolución, por lo que es importante mantenerse actualizado con las últimas tendencias y avances en la tecnología. ¡Espero que estas sugerencias te sean útiles para avanzar en tu comprensión y práctica de SDN distribuido!

In OMNeT++, you can create a subgraph of routers that are connected to an SDN (Software-Defined Networking) controller in a network topology by implementing the necessary logic in your simulation model. Here are the general steps you can follow to achieve this:

By following these steps and implementing the necessary logic in your OMNeT++ simulation model, you can create a subgraph of routers that are connected to an SDN controller in a network topology and analyze their interactions within the network.

To build an environment for computation inside Mobile Edge Computing (MEC) based on the benefits of Software-Defined Networking (SDN) to manage the network, you can follow these steps:

By combining MEC, SDN, and custom controller development, you can create a powerful and efficient environment for computation at the edge of the network, enabling seamless interaction between IoT devices and MEC resources. Using appropriate tools for task generation from IoT devices will ensure smooth communication and task offloading to the MEC servers for processing.

Dear Fatima Aziz,

the most common use of Software-Defined Networks based on Mobile Edge Computing relates to the design of applications to support moving cars:

See examples of this at the address:

There, Figure 9769 illustrates the implementation of SD-IoV and Figure 9771 shows the idea for a Smart City Parking System.

The idea of Digital Twins is of great importance for autonomous driving. For the great future significance of the “Digital Twin of Drives” idea, see my literature sources.

For general information about Digital Twins see:

Best Regards

Anatol Badach

Ghanishtha Bhatti, Harshit Mohan, R. Raja Singh:„Towards the future of smart electric vehicles: Digital twin technology“; Renewable and Sustainable Energy Reviews, Vol. 141, May 2021, DOI: 10.1016/j.rser.2021.110801

Sadeq Almeaibed, Saba Al-Rubaye, Antonios Tsourdos, Nicolas P. Avdelidis: Digital Twin Analysis to Promote Safety and Security in Autonomous Vehicles; IEEE Communications

Dear Thally Tha

Certainly! Here's an example MATLAB code that demonstrates the implementation of the controller placement algorithm using Antlion optimization and PSO for an SDN network:

###matlab code starts

% Parameters

numControllers = 5; % Number of controllers to be placed

numAntlions = 50; % Number of antlions in the population

maxIterations = 100; % Maximum number of iterations

c1 = 2; % Cognitive coefficient for PSO

c2 = 2; % Social coefficient for PSO

w = 0.7; % Inertia weight for PSO

% Network Topology

% Define your SDN network topology (nodes, links, etc.)

% Fitness Function

fitnessFunction = @(controllers) calculateFitness(controllers, networkTopology);

% Initialization

antlions = initializeAntlions(numAntlions, networkTopology);

controllers = antlions(:, 1:numControllers);

% Antlion Optimization with PSO

bestFitness = Inf;

bestSolution = [];

iteration = 1;

while iteration <= maxIterations

% Evaluate Fitness

antlionsFitness = fitnessFunction(controllers);

% Update Best Solution

[minFitness, minIndex] = min(antlionsFitness);

if minFitness < bestFitness

bestFitness = minFitness;

bestSolution = controllers(minIndex, :);

end

% PSO Update

for i = 1:numAntlions

for j = 1:numControllers

r1 = rand();

r2 = rand();

antlions(i, j) = w * antlions(i, j) ...

+ c1 * r1 * (bestSolution(j) - antlions(i, j)) ...

+ c2 * r2 * (controllers(minIndex, j) - antlions(i, j));

end

end

% Update Controllers

controllers = antlions(:, 1:numControllers);

iteration = iteration + 1;

end

% Display Results

disp('Best Solution (Controller Placement):');

disp(bestSolution);

disp('Best Fitness:');

disp(bestFitness);

% Helper Functions

function antlions = initializeAntlions(numAntlions, networkTopology)

% Initialize antlions randomly in the network topology

numNodes = numel(networkTopology.nodes);

antlions = rand(numAntlions, numNodes);

end

function fitness = calculateFitness(controllers, networkTopology)

% Calculate fitness based on the defined fitness function and network topology

% Add your fitness calculation code here

% Return a fitness value for each set of controllers

fitness = zeros(size(controllers, 1), 1);

end

####matlab code ends

Please note that this is a basic implementation to give you an idea of how the algorithm could be structured. You'll need to customize the fitness function and add code specific to your network topology and fitness calculations.

Make sure to adapt the code to your specific requirements, including your network topology representation, fitness function calculations, and any additional constraints or objectives you may have for controller placement in your SDN network.

Remember to thoroughly test and validate the results obtained from the code against your expected outcomes and requirements.

Dynamic Entropy approaches and Machine Learning techniques are both used in the detection of DDoS attacks in SDN. Dynamic Entropy approaches utilize Network traffic statistics to calculate entropy of the network packets. And when the entropy value surpasses the threshold, the system flags the packet as a potential DDoS attack. On the other hand, Machine Learning algorithms classify incoming traffic as either normal or an attack, based on previously learned patterns and features.

Both methods have their advantages and disadvantages. Yet, in some cases, dynamic entropy approaches can provide better results compared to Machine Learning. One of the downsides of Machine Learning is its susceptibility to adversarial attacks, where attackers may exploit the model's weaknesses to evade detection. Besides, Machine Learning requires large datasets to achieve high accuracy. On the other hand, the entropy-based approach provides a quick and light-weight solution that can detect zero-day attacks and can be adapted quickly to new and evolving threats.

In conclusion, both Dynamic Entropy and Machine Learning approaches have their use-cases and can complement each other in detecting DDoS attacks in SDN.

There are a few disadvantages of using dynamic entropy in detecting DDoS attacks in SDN:

1. Increased overhead: Dynamic entropy requires additional overhead for each packet, which can introduce delays and strain network resources.

2. Limited applicability: Dynamic entropy may not be effective in detecting certain types of DDoS attacks, such as slow attacks or those that do not vary in frequency or duration.

3. False positives: Dynamic entropy may also generate false positives, where benign traffic is misclassified as suspicious, which can result in unnecessary intervention by the network.

4. Limited scalability: Dynamic entropy may become less effective as the scale of the network increases, as it becomes more difficult to accurately measure the entropy of traffic on a larger network.

Overall, while dynamic entropy can be an effective tool in detecting DDoS attacks in SDN, it is important to balance its advantages against its potential drawbacks.

Bineet Kumar Joshi

Dear Nimota Jabaar Biobaku,

attached is a short bibliography where you can find some information about the relationship between Data Mining and SDN.

Best regards and much success

Anatol Badach

Kyriakos Sideris, Reza Nejabati, Dimitra Simeonidou: „Seer: Empowering Software Defined Networking with Data Analytics“; 15th International Conference on Ubiquitous Computing and Communications and 2016 International Symposium on Cyberspace and Security (IUCC-CSS), Dec 2016

Albert Mestres et al.: Knowledge-Defined Networking; ACM SIGCOMM Computer Communication Review, Vol. 47 Issue 3, Jul 2017

Haojun Huang et al.: Data-Driven Information Plane in Software-Defined Networking; IEEE Communications Magazine, Vol. 55, Issue 6, Jun 2017

Tam Nguyen: “The Challenges in SDN/ML Based Network Security: A Survey”; arXiv:1804.03539v2 [cs.CR], Apr 2018

Juliana Arevalo Herrera1, Jorge E. Camargo: A Survey on Machine Learning Applications for Software Defined Network Security; International Conference on Applied Cryptography and Network Security (ACNS), Aug 2019

Yuhong Li, Xiang Su, Aaron Yi Ding et al.: „Enhancing the Internet of Things with Knowledge-Driven Software-Defined Networking Technology: Future Perspectives”; Sensors (MDPI), Vol. 20, Jun 2020

No, Open vSwitch (OVS) does not calculate the FCS (Frame Check Sequence). The FCS is calculated by the Ethernet hardware.

hello Wasswa Shafik

thanks alot for this suggestion

I see some of these simulations like EdgeCloudSim

i will read more details about what of these will be more useful to my ideas

Regards …

The response time of a floodlight controller can be calculated by measuring the time it takes for the controller to react to a specific event or request. This can be done by sending a command to the controller and measuring the time it takes for the controller to respond to the command.

Here are the steps you can follow to calculate the response time of a floodlight controller:

Determine the event or request you want to use to measure the response time of the controller. For example, you could measure the response time of the controller to a request to turn on a specific light.

Use a tool such as Wireshark to capture the network traffic between the floodlight controller and the device sending the request.

Send the request to the floodlight controller and record the time the request was sent.

Monitor the network traffic in Wireshark and record the time the response was received.

Calculate the response time by subtracting the time the request was sent from the time the response was received.

Repeat the process multiple times to get an average response time.

It's worth noting that the response time of a floodlight controller can be affected by various factors such as the network latency, processing power of the controller, and the load on the network. Therefore, it's important to perform multiple tests under different conditions to get an accurate measurement of the response time.

Dear Fatima Aziz,

the idea for "Recursive Hierarchical SDN Architecture" might help you.

See:

Figure 13 – Peers as symmetric requestors and providers

Figure 9548: Recursive hierarchical SDN architecture: a) general idea, b) implementation of handover in SD-IoV

Best regards and much success

Anatol Badach

One option is to check out the NetSim SDN library, which has a range of features to help you simulate and analyze the behavior of software-defined networks. Additionally, there are several projects on network security available on the NetSim website at the following URL: https://tetcos.com/file-exchange.html. These projects provide resources to model and simulate some recent developments and techniques in network security

Yes you can in Linux.

The Software Name is OMNET++ 4.6 or Higher Version.

Exaple Solution:

void Txc::handleMessage(cMessage *msg) { // ... simtime_t eed = simTime() - msg->getCreationTime(); }

simTime() returns current simulation time.

The same problem is happening to me. How do you solve the problem? Manisha Rajoriya

You cannot base your research on the responses of others; the responses will be useful, but not sufficient to guarantee high research quality. This is entirely your job. I believe you should read more survey articles about these topics. Following that, you'll be able to present new ideas to fill the research gaps you discovered.

Dear Noor Thamer Mahmood,

large SDNs can be set up in the form of hierarchical structures.

As the idea for this see

Fig. 9546: Two-tier hierarchical SDN architecture

and

Fig. 9548: Recursive hierarchical SDN architecture: a) general idea, b) implementation of handover in SD-IoV

under the addresses

Best regards

Anatol Badach

Open Networking Foundation: „ONF TR-521: SDN Architecture, Issue 1.1, April 2016

Dear Manash Chakraborty ,

I believe this article can help you find inspiration. This is a Q1 journal about network communication and network architecture. It proposed an internet of things platform with p2p wireless communication, high speed, and security. It integrated with ION DID (based on blockchain) too. Finally, the source code is available on the GitHub link in this article.

BRs.

Dear Muddassar Mushtaq,

see my answer at the address:

There you have good sources of literature on the subject of "SDN Security".

The idea of “setting up Honeypots” as a defense strategy is very interesting. See this literature source:

Alejandro Molina Zarca, Jorge Bernal Bernabe, Antonio Skarmeta, Jose M. Alcaraz Calero: Virtual IoT HoneyNets to Mitigate Cyberattacks in SDN/NFV-Enabled IoT Networks, IEEE Journal on Selected Areas in Communications, Vol 38, Issue 6, Jun 2020; DOI: 10.1109/JSAC.2020.2986621

Best regards and much success

Anatol Badach

Ali El Kamel thank you so much sir, I will go through this paper and let you know

Good Morning,

I think you will have to try this code and see;

Tou may read this:

NetSim V11 now has Software Defined Network (SDN) simulations with SDN Controller and Open flow compatibility module.

, me to, I want to simulate vanet project. my project is about the secure communication protocol. and I implemented it with python. So, can I use ns2 and sumo in my project. I must complete my project with the simulation before the dateline

Thank you so much for your suggestion and sharing

Look at this link may help u

Dear Varatharasasarma Thirukkumar,

I have included some literature sources on the subject of "Anomaly Detection in SDN". I hope these literature sources can help you.

Best regards and good luck

Anatol Badach

Daojing He, Sammy Chan, Xiejun Ni, Mohsen Guizani: Software-Defined-Networking-Enabled Traffic Anomaly Detection and Mitigation; IEEE Internet of Things Journal, Vol. 4, Issue 6, Dec 2017

Junfeng Xie, F. Richard Yu, Tao Huang; Renchao Xie, Jiang Liu, Chenmeng Wang, Yunjie Liu: "A survey of machine learning techniques applied to software defined networking (SDN): Research issues and challenges", IEEE Communications Surveys & Tutorials, Vol. 21, Issue 1, Firstquarter 2019, DOI: 10.1109/COMST.2018.2866942

Li-Der Chou, Chien-Chang Liu , Meng-Sheng Lai , Kai-Cheng Chiu, et al.: „Behavior Anomaly Detection in SDN Control Plane: A Case Study of Topology Discovery Attacks“; Hindawi Wireless Communications and Mobile Computing, Vol 2020, Article ID 8898949, Nov 2020; DOI: 10.1155/2020/8898949

Trung V. Phan, Tri Gia Nguyen, Nhu-Ngoc Dao, Truong Thu Huong, Nguyen Huu Thanh, Thomas Bauschert: „DeepGuard: Efficient Anomaly Detection in SDN With Fine-Grained Traffic Flow Monitoring“; IEEE Transactions on Network and Service Management, Vol. 17, Issue 3, Sep 2020; DOI: 10.1109/TNSM.2020.3004415

Dear @ Esmaeil Amiri

There are some trusted sources for the datasets. I recommend you to search on these websites:

https://datasetsearch.research.google.com/

Certainly you can

Dr. MOHAMMAD FAISAL currently working at the department of computer science and IT, University of Malakand Pakistan, received his M.S. degree in information security management from SZABIST, Pakistan, in 2012, and the Ph.D. degree in network security from the Department of Computer Science and Information Technology, University of Malakand in 2018. His research interests include ML and security of wireless ad hoc networks MANETs, VANETs, IoT, Cloud, Fog, Edge, Blockchain and digital forensics.

Dear Deepanshu Garg,

Please look at the links below:

https://www.researchgate.net/post/Can_any_bod_help_in_ns3_and_sdn_implimentation

NS3 or omnet++ for simulating sdn cellular network ? (researchgate.net)

https://ns3-code.com/ns3-sdn-projects/

https://www.nsnam.org/wp-content/uploads/2015/04/WNS3_2015_submission_39.pdf

I recommend EXata Network Simulator/Emulator. Its new versions have models that support SDN like EXata 5G and EXata Cyber. For sure, it is not open source as you requested, because it is a powerful commercial tool. However, sometimes they provide trial licenses for students and discounts for academic licenses. You may refer to the simulator website for more details:

This article can help you know how to achieve load balancing in SDN architecture.

Good luck!

OMNeT++ is also very good, but a little complex at the start

Dear Dr. Ali Elkamel thank you very much for your help

I will try your opinion, it gave me a lot of inspiration@Ali El Kamel

Install your SDN VM on your VMware or Virtualbox as like other operating system, then will install like Linux OS because it used the Linux carnel and login the Mininet environment using special commands. Read and follow this link you will get more about the Mininet and issues related to it.

Check out this wok :

Thanks Elisa Rojas for the link provided.

You can do simulation only check this link https://www.youtube.com/watch?v=eC5QDLgCPO8

Important topic, as a backhaul driver in the 5G era, which is not just about the mobile segment! Core network aspect are even more important than the radio link in this context.

Wishing great success to the publication, Isaias Martinez-Yelmo !

Dear Nur A Mahfila,

a very current and interesting theme would be, for example, the conception of ideas for the realization of

"Citizen Services based on Software Defined IoT".

In particular, various Emergency Services (Fire Services, Medical Emergency Services, ...) can be implemented very intelligently on the basis of IoT Devices Proxies.

See the figure 9772 under the address

Best regards and much success

Anatol Badach

Hi,

The following paper finds the cost of SDN network in USD against the workload. They also present a model to calculate the cost. Hopefully you find this useful.

Link: https://www.computer.org/csdl/proceedings-article/aina/2017/07920946/12OmNxwnccx

Thanks for your answer Ahmad, Could you guide me how to implement this idea. I have not found any thing in open flow specifications or Ryu controller demonstration.

This discussion is still open?

Hossein Akbari

AI algorithm or intelligent routing are useful in these fields.

I have experience with software defined radio where one can implement the digital down and conversion using programmable devices mostly field programmable gate arrays. One can also software implement the base band processing functions using programmable devices such as FPGA and digital signal processing DSPs.

Similarly one can implement network functions using similar techniques where one can represent the functions in a software code that can be implemented on programmable devices.

In fact any digital processing functions can be implemented on programmable platforms such as DSPs and FPGAs.

For specific applications you can search the web.

As a matter of definition software defined device= digital programmable platform + software deification of the function to be implemented. The platforms are in form of programmable DSP and FPGA kits.

Best wishes

You can use a demultiplexer to connect the switch to the two controllers.

The demultiplexer is controlled by a select signal.

Best wishes

You can use either Mininet or NS3 Open Flow for IDS in SDNs...

You can develop a .py file in RYU app and edit the run configuration. When starting RYU, the developed .py file will be launched.

K-means and K-center algorithms may help you for the controller placement....

This paper may help you:

Thanks for your reply & help Aleksandar S. Sokolovski .

Actually I am using ubuntu 14 because this is the recommended software in development tennison guide & it is tested with the framework. I will try to upgrade it and keep you posted about the results.

Thanks a million

Good Answer Jeisson Sanchez

See my paper: Article Controller selection in software defined networks using best...

no dear i need a SDN controller algorithm

please send me as soon as possible

thanks@Esmaeil

Interested question

In my opinion, it depends on many factors, i.e. it depends on several factors such as: what you want to prove/implement with your research, what is your previous knowledge of SDN controllers, etc.

For example, I have worked with NOX, POX, Ryu, Floodlight, OpenDaylight and ONOS, and my experience is that Ryu and ONOS are currently the best ones. I usually leverage Ryu for fast prototyping/proof-of-concepts, while I use ONOS when I'm implementing a next-step of the previous prototype (something that could be standardized/used by industry, etc.). In particular, in you haven't even worked with an SDN controller, I'd start with Ryu.

Good luck!

New routing algorithms are implemented in specific SDN platforms. Afterwards, the SDN switches managed (deployed in Mininet, for example) will behave as designed by these platforms. The most popular platforms could be Ryu, ODL and ONOS. Maybe the easiest to start with is Ryu.

You will find it here https://docs.opendaylight.org/en/latest/downloads.html

found an answer in https://stackoverflow.com/questions/60018102/how-to-connect-virtualbox-machines-as-hosts-to-mininet-ovs-switches-for-sdn-expe/60054218#60054218

Dear Kaustubh,

I think the best way to manage the wireless communication in the ISM band is to use the techniques of the cognitive radio systems. So one has a radio which can sense communicate with control center and get a decision and a connection information from the center. There is the compressive sensing techniques to reduce the sensing signal load.

You can find many many papers when you make a literature survey.

Best wishes

Md. Billal Hossain can u find the code for measuring the Delay??? i am also working on the same area can u suggest me the pox code for measuring delay?