EXPRO2024ewfrdusayhfusyhedfyhsduyffdsyhjsdyfhu

trewtfgwe

1. DYNAMIC TRAFFIC OPTIMIZATION: A NETWORK LOAD BALANCING Team number:41 Team Members 1.Sharon K 2.Shravan 3.Sripad J 4.Tharun R Guide Name: Dr. Bola Sunil Kamath , Assistant Professor Gd-III DEPARTMENT OF ISE, NMAMIT, NITTE(DU), Nitte 1

2. AGENDA DEPARTMENT OF ISE, NMAMIT, NITTE(DU), Nitte 2 INTRODUCTION LITERATURE REVIEW PROBLEM STATEMENT AND OBJECTIVES METHODOLOGY AND DESIGN RESULTS AND DISCUSSION CONCLUSION FUTURE WORK REFERENCE

3. INTRODUCTION In this project, we design and implement a load balancing system running on techniques of reverse proxy. Our backend will make use of HTTPS servers running on Python as well as being dynamically scaled through Kubernetes and Docker. It automatically adjusts to varying traffic loads by increasing the number of server instances, or pods. It works at the application layer (HTTP/HTTPS) and decide based on the application data levels examples include URLs, headers, and cookies. We incorporated a reverse proxy in our project with a Layer 7 Load Balancer. Since it is a reverse proxy, we can intelligently route the requests based on some applicative conditions in the client's HTTP request properties. DEPARTMENT OF ISE, NMAMIT, NITTE(DU), Nitte 3

4. LITERATURE REVIEW Literature survey Description Server Load Balancing Techniques Software-Defined Networking (SDN) can enhance load balancing strategies in response to increasing network traffic. They compare two load balancing techniques—bandwidth-based and round-robin—using Mininet for emulation. Results show that bandwidth-based load balancing outperforms round-robin in throughput and response time Load Balancing for Structured P2P Network The study highlights the limitations of current dynamic load balancing in structured peer-to-peer networks and proposes an efficient scheme using an aggregation mechanism for comprehensive load information. This approach improves decision-making, scalability, and uniformity, addressing inefficiencies and emphasizing the need for effective load management for performance and reliability. DEPARTMENT OF ISE, NMAMIT, NITTE(DU), Nitte 4

5. LITERATURE REVIEW Literature survey Description Load-Balancing Technique in Ad-Hoc Networks The Load-Balancing Technique to improve load distribution in Ad Hoc Networks by dynamically adjusting the transmission ranges of cluster-heads, effectively redistributing load and enhancing network efficiency while reducing the risk of cluster-head failure due to energy depletion. Round-Robin load balancing scheme Round-Robin load balancing scheme for software-defined networks (SDNs) in institutional settings, addressing uneven traffic distribution and improving network performance Conclusion The exploration of load balancing techniques within Software-Defined Networking (SDN) underscores the critical need for efficient traffic management in response to escalating network demands. DEPARTMENT OF ISE, NMAMIT, NITTE(DU), Nitte 5

6. LITERATURE GAP IDENTIFIED Integration of Techniques : There is limited exploration of hybrid load balancing approaches that combine the strengths of both bandwidth-based and round-robin methods, potentially leading to enhanced performance in various scenarios. Scalability in Dynamic Environments: While some strategies address scalability, there is a lack of comprehensive models that can dynamically adapt to changing network conditions in real time, especially in heterogeneous networks. Impact of SDN on Emerging Technologies: There is a gap in understanding how SDN can be leveraged to optimize load balancing in emerging technologies, such as IoT and edge computing, where traffic patterns can be highly unpredictable and resource-constrained. DEPARTMENT OF ISE, NMAMIT, NITTE(DU), Nitte 6

7. PROBLEM STATEMENT AND OBJECTIVES PROBLEM STATEMENT Increasingly complex network architectures face challenges in effectively distributing traffic loads across servers, resulting in uneven resource utilization, performance degradation, and potential downtime. Addressing these issues requires the implementation of a dynamic traffic optimization system via robust network load balancing mechanisms to ensure efficient resource allocation, high availability, and optimal user experience. OBJECTIVES Enhance network performance by dynamically distributing traffic across servers. Improve scalability and resource utilization through efficient load balancing. Ensure high availability and fault tolerance to minimize downtime. Provide comprehensive monitoring and reporting capabilities for proactive management. Reduce operational costs by optimizing resource allocation and minimizing wastage. DEPARTMENT OF ISE, NMAMIT, NITTE(DU), Nitte 7

8. METHODOLOGY AND DESIGN Methodology The methodology for the Dynamic Traffic Optimization: A Network Load Balancing System project follows an Agile approach, emphasizing adaptability and iterative improvements at each development phase. This allows for responsive adjustments based on ongoing feedback and performance metrics. Major phases of the project includes planning, system design, implementation, testing, and deployment. Design With reverse proxy algorithm, Docker, Kubernetes, and AWS, the network load balancer efficiently manages and distributes client requests for high availability and scalability. Clients' requests reach the Network Load Balancer at the transport layer in AWS, where loads are spread across multiple backend servers running on the Kubernetes cluster. DEPARTMENT OF ISE, NMAMIT, NITTE(DU), Nitte 8

9. RESULTS AND DISCUSSION Network load balancer helps manage the traffic in a different way as incoming requests are to be divided on multiple servers, thereby ensuring that only one server gets overwhelmed hence keeping it fast and online for its users.. Servers make much better use of their resources because network load balancer ensures proper balance and maintains things quite efficient. A network load balancer merely ensures that a user experience will be smoother and more reliable as traffic is distributed evenly throughout the network. DEPARTMENT OF ISE, NMAMIT, NITTE(DU), Nitte 9

10. RESULTS AND DISCUSSION Server Status Server status is the status given in server management systems, which indicates whether a server is active or inactive. When a server is deactivated, its status is set to ‘false' meaning it is not currently in use or not responding to requests. On activation of servers, it changes to ‘true' status and is considered available for traffic. Dashboard Once a socket is run, the dashboard displays, most probably signifying that the socket has managed to establish a connection between the server and client; data then keeps flowing real-time and then shows on the dashboard. DEPARTMENT OF ISE, NMAMIT, NITTE(DU), Nitte 10

11. CONCLUSION This renders the network load balancer a very important part of application delivery, particularly in the cloud, since the efficient distribution of incoming traffic across multiple server instances helps prevent server overload-a real plus for faster response times and generally improved user experience. Since network load balancer provide significant value in handling traffic when the workload experienced by applications fluctuates, they allow organizations to scale resources horizontally for better efficiency and performance. DEPARTMENT OF ISE, NMAMIT, NITTE(DU), Nitte 11

12. FUTURE WORK Network Load Balancer are going to experience huge growth in the future, as cloud computing advances, along with application architecture and emerging technologies. Automation and AI will be critical in smart management of traffic. They will allow network load balancers to optimize resource allocation according to real-time traffic but also enhance their reliability in real time through self healing. Advanced cyber-attacks will find future network load balancers infused with innovative security techniques like automated threat detection as well as a zero-trust architecture that ensures applications operate securely. DEPARTMENT OF ISE, NMAMIT, NITTE(DU), Nitte 12

13. REFERENCE Labdah ALGhafran and Zulkefli Bin Muhammed Yusof, "Load-Balancing Technique in Clustered Mobile Ad-Hoc Networks," Proceedings of the 2013 International Conference on Advanced Computer Science Applications and Technologies, 2013. Farruh Ishmanov, "Distributed Clustering Algorithm with Load Balancing in Wireless Sensor Network," Proceedings of the 2009 World Congress on Computer Science and Information Engineering, 2009. Edison F. Aza, "Implementation of Round-Robin Load Balancing Scheme in a Wireless Software Defined Network," Proceedings of the 2015 18th International Conference on Network-Based Information Systems, 2015. Atsushi Takeda, Akiko Takahashi "Efficient Dynamic Load Balancing for Structured P2P Network," Proceedings of the 2015 18th International Conference on Network-Based Information Systems, 2015. DEPARTMENT OF ISE, NMAMIT, NITTE(DU), Nitte 13

14. QUESTIONS? Thank you DEPARTMENT OF ISE, NMAMIT, NITTE(DU), Nitte 14