Farklı MTU Boyutları Altında 802.3ad ve Round Robin Modları Üzerindeki Performans Etkisinin İncelenmesi

Öz

MTU boyutu ağ performansını doğrudan etkiler; büyük MTU’lar protokol yükünü azaltıp işlemci üzerindeki yükü hafifletebilirken, uyumsuz ağlarda parçalara bölünmeye ve gecikmeye yol açabilir. Bu çalışmada, 802.3ad ve Round-Robin bağlantı birleştirme modlarının farklı MTU değerleri (450–9200 bayt) altında, tekli ve paralel veri akışı testleriyle performansı değerlendirilmiştir. Çok çekirdekli sistemlerde işlemci çekirdek kullanımı ve veri aktarım hızı analiz edilmiştir. Sonuçlar, Round-Robin modunun paralel akışlarda daha yüksek tepe verim sağlasa da işlemci kullanımında dengesizlik yarattığını ve uzun vadeli kararlılığı olumsuz etkilediğini göstermektedir. Öte yandan, 802.3ad modu daha dengeli işlemci yükü dağılımı ve tutarlı veri aktarım performansı ile yüksek verimli ve sürdürülebilir ağ ortamları için daha uygun bir çözüm sunmaktadır. Bu çalışma, modern ve gecikmeye duyarlı ağ tasarımlarında MTU farkındalığı olan, işlemci dostu bağlantı birleştirme stratejilerinin önemini vurgulamaktadır.

Anahtar Kelimeler

• Link Aggregation
• 802.3ad
• Round-Robin
• MTU Optimization
• Network Performance
• Throughput
• CPU Utilization
• LACP
• Bonding Modes
• SDN
• Jumbo Frames
• High-Performance
• Networking

Investigation of Performance Impact of 802.3ad and Round-Robin Bonding Modes Under Different MTU Configurations

Öz

MTU size plays a crucial role in performance, where larger MTUs reduce protocol overhead and CPU load but may cause fragmentation and latency in heterogeneous networks. This study evaluates the performance of 802.3ad and Round-Robin bonding under varying MTU values (450–9200 bytes) across single and parallel stream tests. CPU core utilization and throughput metrics were analyzed to assess efficiency in multi-core systems. Results show that although Round-Robin achieves higher peak throughput in parallel streams, it leads to significant CPU imbalance and potential instability. In contrast, 802.3ad ensures better CPU load distribution and consistent throughput, making it more suitable for long-term, high-performance environments. The study highlights the need for MTU-aware, CPU-efficient bonding strategies in modern, latency-sensitive network designs.

Anahtar Kelimeler

• Link Aggregation
• 802.3ad
• Round-Robin
• MTU Optimization
• Network Performance
• Throughput
• CPU Utilization
• LACP
• Bonding Modes
• SDN
• Jumbo Frames
• High-Performance
• Networking

Kaynakça

1. 1. Stallings, W. (2014). Data and computer communications. (10th ed.), Pearson Education.
2. 2. Tanenbaum, A.S. & Wetherall, D.J. (2010). Computer networks. (5th ed.), Pearson Education.
3. 3. IEEE Std 802.1AX-2020. IEEE standard for link aggregation.
4. 4. Nikolova, D. & Blondia, C. (2011). Bonded deficit round robin scheduling for multi-channel networks. Computer Networks, 55(15), 3503-3516. 5. Odom, W. (2016). CCNA routing and switching ICND2 200-105 official cert guide. Cisco Press.
5. 6. Tso, F.P., Jouet, S. & Pezaros, D.P. (2016). Network and server resource management strategies for data centre infrastructures: A survey. Computer Networks, 106, 209-225.
6. 7. Steinbacher, S. & Bredel, M. (2015). LACP meets OpenFlow: Dynamic link aggregation in SDN. IEEE INFOCOM.
7. 8. Abbas, K., et al. (2022). An efficient SDN‐based LTE‐WiFi spectrum aggregation system for heterogeneous 5G networks. Transactions on Emerging Telecommunications Technologies, 33(4), e3943.
8. 9. Forouzan, B.A. & Fegan, S.C. (2017). Data communications and networking. (5th ed.), McGraw-Hill Education.

9. 10. Comer, D. (2014). Computer networks and internets. (6th ed.), Pearson Education.
10. 11. Sharma, K. & Badarla, V. (2016). Curtailing latency in data center network by adopting jumbo frames. 2016 IEEE Int. Conf. on Advanced Networks and Telecommunication Systems (ANTS), 1-6.
11. 12. Ma, J., Lou, W. & Li, X.-Y. (2013). Contiguous link scheduling for data aggregation in wireless sensor networks. IEEE Transactions on Parallel and Distributed Systems, 25(7), 1691-1701.
12. 13. Elmadani, M. & Sati, S.O. (2024). MTU analyzing for data centers interconnected using VxLAN. 2024 ASU International Conference on Emerging Technologies for Sustainability and Intelligent Systems (ICETSIS). IEEE.
13. 14. Parpinello, D. (2024). Multibuffer implementation in XDP-testing and evaluation with jumbo frames. Master's Thesis.
14. 15. Li, T., et al. (2009). Aggregation with fragment retransmission for very high-speed WLANs. IEEE/ACM Transactions on Networking, 17(2), 591-604.
15. 16. Li, F., et al. (2024). MTU-adaptive in-band network-wide telemetry. IEEE/ACM Transactions on Networking, 32(3), 2315-2330.
16. 17. Elmadani, M. & Sati, S.O. (2024). MTU analyzing for data centers interconnected using VxLAN. 2024 ASU International Conference on Emerging Technologies for Sustainability and Intelligent Systems (ICETSIS). IEEE.
17. 18. Li, F., et al. (2024). MTU-adaptive in-band network-wide telemetry. IEEE/ACM Transactions on Networking, 32(3), 2315–2330.
18. 19. Abbas, K., et al. (2022). An efficient SDN‐based LTE‐WiFi spectrum aggregation system for heterogeneous 5G networks. Transactions on Emerging Telecommunications Technologies, 33(4), e3943.
19. 20. Parpinello, D. (2024). Multibuffer Implementation in XDP – Testing and Evaluation with Jumbo Frames. Master’s Thesis.
20. 21. Shashidhara, R., Kaufmann, A. & Peter, S. (2025). Scaling data center TCP to terabits with laminar. arXiv preprint arXiv: 2504.19058.
21. 22. Nan, H., Zhou, Z. & Yang, M. (2025). Plug & offload: transparently offloading TCP stack onto off‑path SmartNIC with PnO‑TCP. arXiv preprint arXiv: 2503.22930.
22. 23. Koşar, O. & Atak, M. (2023). A multi-criteria decision support model for cloud computing virtual server product selection. Çukurova University Journal of the Faculty of Engineering, 38(4), 939-953.
23. 24. Kırcı, P. & Zaim, H. (2015). Optical burst switching and multicasting over wavelength division multiplexing (WDM). Çukurova University Journal of the Faculty of Engineering and Architecture, 30(1), 175-182.