The week's pick
Random Articles
Reseach Article
A Simulation Study of AI Traffic Patterns and their Impact on Data Center Network Performance
| International Journal of Computer Applications |
| Foundation of Computer Science (FCS), NY, USA |
| Volume 187 - Number 115 |
| Year of Publication: 2026 |
| Authors: Mike Adedoyin |
10.5120/ijcaaa704a80639c
|
Mike Adedoyin . A Simulation Study of AI Traffic Patterns and their Impact on Data Center Network Performance. International Journal of Computer Applications. 187, 115 ( Jun 2026), 23-30. DOI=10.5120/ijcaaa704a80639c
Abstract
AI training workloads are changing the way traffic behaves inside modern data centers. Unlike many traditional enterprise applications, distributed training jobs often create synchronized bursts during collective operations such as AllReduce and gradient synchronization. These bursts can temporarily overload shared links and expose limitations in switch queueing, transport behavior, and capacity planning methods that rely primarily on average utilization. This paper uses NS-3 3.43 to simulate AI-like burst traffic in a small leaf-spine bottleneck topology. Four scenarios are evaluated: normal TCP enterprise traffic, a moderate synchronized burst load of 120 Mbps, a heavy burst load of 808 Mbps, and the same heavy load with FQ-CoDel replacing FIFO queueing. Results show that even moderate average overload can produce severe performance degradation when traffic is synchronized. At 120 Mbps, packet loss reaches 52.8% and average delay nearly doubles. At 808 Mbps, loss increases to 92.9%. FQ-CoDel reduces average delay by approximately 27% and improves inter-flow delay fairness, although throughput remains constrained under heavy load. The study provides a compact and reproducible simulation baseline for analyzing AI-driven incast behavior under controlled overload conditions.
References
- M. Alizadeh, A. Greenberg, D. A. Maltz, J. Padhye, P. Patel, B. Prabhakar, S. Sengupta, and M. Sridharan, “Data Center TCP (DCTCP),” in Proc. ACM SIGCOMM, New Delhi, India, 2010, pp. 63–74.
- Y. Zhu, H. Eran, D. Firestone, C. Guo, M. Lipshteyn, Y. Liron, J. Padhye, S. Raindel, M. H. Yahia, and M. Zhang, “Congestion Control for Large-Scale RDMA Deployments,” in Proc. ACM SIGCOMM, London, UK, 2015, pp. 523–536.
- R. Mittal, V. T. Lam, N. Dukkipati, E. Blem, H. Wassel, M. Ghobadi, A. Vahdat, Y. Wang, D. Wetherall, and D. Zats, “TIMELY: RTT-based Congestion Control for the Datacenter,” in Proc. ACM SIGCOMM, London, UK, 2015, pp. 537–550.
- C. Guo, H. Wu, Z. Deng, G. Soni, J. Ye, J. Padhye, and M. Lipshteyn, “RDMA over Commodity Ethernet at Scale,” in Proc. ACM SIGCOMM, Florianopolis, Brazil, 2016, pp. 202–215.
- Y. Chen, R. Griffith, J. Liu, R. H. Katz, and A. D. Joseph, “Understanding TCP Incast Throughput Collapse in Datacenter Networks,” in Proc. 1st ACM Workshop on Research on Enterprise Networking (WREN), Barcelona, Spain, 2009, pp. 73–82.
- B. Vamanan, J. Hasan, and T. N. Vijaykumar, “Deadline-Aware Datacenter TCP (D2TCP),” in Proc. ACM SIGCOMM, Helsinki, Finland, 2012, pp. 115–126.
- M. Al-Fares, A. Loukissas, and A. Vahdat, “A Scalable, Commodity Data Center Network Architecture,” in Proc. ACM SIGCOMM, Seattle, WA, USA, 2008, pp. 63–74.
- A. Greenberg, J. R. Hamilton, N. Jain, S. Kandula, C. Kim, P. Lahiri, D. A. Maltz, P. Patel, and S. Sengupta, “VL2: A Scalable and Flexible Data Center Network,” in Proc. ACM SIGCOMM, Barcelona, Spain, 2009, pp. 51–62.
- K. Nichols and V. Jacobson, “Controlling Queue Delay,” ACM Queue, vol. 10, no. 5, pp. 20–34, May 2012.
- M. Alizadeh, S. Yang, M. Sharif, S. Katti, N. McKeown, B. Prabhakar, and S. Shenker, “pFabric: Minimal Near-Optimal Datacenter Transport,” in Proc. ACM SIGCOMM, Hong Kong, China, 2013, pp. 435–446.
- J. Dean, G. Corrado, R. Monga, K. Chen, M. Devin, M. Mao, A. Senior, P. Tucker, K. Yang, Q. V. Le, and A. Y. Ng, “Large Scale Distributed Deep Networks,” in Proc. Advances in Neural Information Processing Systems (NIPS), Lake Tahoe, NV, USA, 2012, pp. 1223–1231.
- L. Poutievski, O. Mashayekhi, J. Ong, A. Singh, M. Tariq, R. Wang, J. Zhang, V. Beauregard, P. Conner, S. Gribble, R. Kapoor, S. Kratzer, N. Lewis, H. Lim, K. Namyar, A. Nikore, J. S. Osborne, P. Osterhout, E. S. Ryan, and A. Vahdat, “Jupiter Evolving: Transforming Google’s Datacenter Network via Optical Circuit Switches and Software-Defined Networking,” in Proc. ACM SIGCOMM, Amsterdam, Netherlands, 2022, pp. 359–377.
- R. Jain, D. Chiu, and W. Hawe, “A Quantitative Measure of Fairness and Discrimination for Resource Allocation in Shared Computer Systems,” DEC Research Report TR-301, Sept. 1984.
- D. Abts and B. Felderman, “A Guided Tour Through Data-Center Networking,” ACM Queue, vol. 10, no. 5, pp. 10–23, May 2012.
- M. Allalouf, Y. Shavitt, and E. Steiner, “Notes on the Simulation of TCP Congestion Control Algorithms,” School of Electrical Engineering, Tel Aviv University, Tech. Rep.
- G. F. Riley and T. R. Henderson, “The ns-3 Network Simulator,” in Modeling and Tools for Network Simulation, K. Wehrle, M. Güneş, and J. Gross, Eds. Berlin: Springer, 2010, pp. 15–34.
- The ns-3 Network Simulator, “Traffic Control Layer,” [Online]. Available: https://www.nsnam.org/docs/models/html/traffic-control.html. Accessed: Jun. 2026.
- Cisco Systems, “AI/ML Networking Design Guide for Data Centers,” Cisco White Paper, 2024.
- NVIDIA Corporation, “NVIDIA Spectrum-X: Ethernet Platform for AI Networking,” NVIDIA Technical Brief, 2024.
Index Terms
Keywords