Papers
Topics
Authors
Recent
Gemini 2.5 Flash
Gemini 2.5 Flash
162 tokens/sec
GPT-4o
7 tokens/sec
Gemini 2.5 Pro Pro
45 tokens/sec
o3 Pro
4 tokens/sec
GPT-4.1 Pro
38 tokens/sec
DeepSeek R1 via Azure Pro
28 tokens/sec
2000 character limit reached

5GC$^2$ache: Improving 5G UPF Performance via Cache Optimization (2404.13991v1)

Published 22 Apr 2024 in cs.NI

Abstract: Last Level Cache (LLC) is a precious and critical resource that impacts the performance of applications running on top of CPUs. In this paper, we reveal the significant impact of LLC on the performance of the 5G user plane function (UPF) when running a cloudified 5G core on general-purposed servers. With extensive measurements showing that the throughput can degrade by over 50\% when the precious LLC resource of UPF is not properly allocated, we identify three categories of performance degradation caused by incorrect LLC usage: DMA leakage problem, hot/cold mbuf problem and cache contention. To address these problems, we introduce the design and implementation of 5GC$2$ache that monitors the LLC status as well as the throughput performance and dynamically adjusts key parameters of the LLC resource allocation. Our experiments show that 5GC$2$ache enables a commercial 5G core to increase its throughput to 76.41Gbps, 39.41\% higher than the original performance and 29.55\% higher than the state-of-the-art.

Definition Search Book Streamline Icon: https://streamlinehq.com
References (56)
  1. 3GPP TS 23.501. System architecture for the 5g system;(rel-15), 2015.
  2. 3GPP TS 23.502. Procedures for the 5g system (5gs);(rel-15), 2017.
  3. 3GPP TS 29.244. Interface between the control plane and the user plane nodes;(rel-16), 2018.
  4. 3GPP TS 33.513. 5g security assurance specification (scas); user plane function (upf);(rel-16), 2019.
  5. Reverse engineering of cache replacement policies in intel microprocessors and their evaluation. In 2014 IEEE International Symposium on Performance Analysis of Systems and Software (ISPASS), pages 141–142, 2014.
  6. Design considerations for a 5g network architecture. IEEE Communications Magazine, pages 65–75, 2014.
  7. On the complexity of traffic traces and implications. SIGMETRICS ’20, page 47–48, 2020.
  8. A study of networking software induced latency. In 2015 International Conference and Workshops on Networked Systems (NetSys), pages 1–8, 2015.
  9. Cisco. Cisco TRex: Realistic Traffic Generator. https://trex-tgn.cisco.com. (Accessed on 2024-01).
  10. Kpart: A hybrid cache partitioning-sharing technique for commodity multicores. In 2018 IEEE International Symposium on High Performance Computer Architecture (HPCA), pages 104–117, 2018.
  11. Reexamining direct cache access to optimize I/O intensive applications for multi-hundred-gigabit networks. In 2020 USENIX Annual Technical Conference, USENIX ATC 2020, pages 673–689, 2020.
  12. Make the most out of last level cache in intel processors. In Fourteenth EuroSys Conference (EuroSys’20), 2019.
  13. Optimizing intel data direct i/o technology for multi-hundred-gigabit networks. In Fifteenth EuroSys Conference (EuroSys’20), 2020.
  14. Intel Forum. Tuning the buffers: a practical guide to reduce or avoid packet loss in dpdk applications, 2017.
  15. Linux Foundation. Perf: Linux profiling with performance counters. https://perf.wiki.kernel.org/index.php/Main_Page. (Accessed on 2024-01).
  16. free5GC Organization. free5GC: An Open-Source Project for 5G Mobile Core Networks. https://www.free5gc.org/. (Accessed on 2024-01).
  17. Ginseng: Market-driven llc allocation. In 2016 USENIX Annual Technical Conference USENIX (ATC’16), pages 295–308, 2016.
  18. Bufferbloat: dark buffers in the internet. Communications of the ACM, pages 57–65, 2012.
  19. A. Gupta and R. K. Jha. A survey of 5g network: Architecture and emerging technologies. IEEE Access, pages 1206–1232, 2015.
  20. Cache qos: From concept to reality in the intel® xeon® processor e5-2600 v3 product family. In 2016 IEEE International Symposium on High Performance Computer Architecture (HPCA), pages 657–668, 2016.
  21. Design and implementation of a generic 5g user plane function development framework. In MobiCom 2021, page 846–848, 2021.
  22. Direct cache access for high bandwidth network i/o. In 32nd International Symposium on Computer Architecture (ISCA’05), pages 50–59, 2005.
  23. Intel. Data Plane Development Kit (DPDK). https://www.dpdk.org. (Accessed on 2024-01).
  24. Intel. Intel Data Direct I/O Technology. https://www.intel.com/content/www/us/en/io/data-direct-i-o-technology.html. (Accessed on 2024-01).
  25. Intel. Intel 64 and ia-32 architectures software developer’s manual volume 4: Model-specific registers, 2019.
  26. L25gc: A low latency 5g core network based on high-performance nfv platforms. In SIGCOMM 2022, page 143–157, 2022.
  27. High performance packet processing with flexnic. In 2016 International Conference on Architectural Support for Programming Languages and Operating Systems (ASPLOS’16), pages 67–81, 2016.
  28. Characterization of direct cache access on multi-core systems and 10gbe. In 2009 IEEE 15th International Symposium on High Performance Computer Architecture (HPCA), pages 341–352, 2009.
  29. A case for stateless mobile core network functions in space. In Proceedings of the ACM SIGCOMM 2022 Conference, SIGCOMM ’22, page 298–313. Association for Computing Machinery, 2022.
  30. A new server i/o architecture for high speed networks. In 2011 IEEE 17th International Symposium on High Performance Computer Architecture (HPCA), pages 255–265, 2011.
  31. Contention-aware performance prediction for virtualized network functions. In SIGCOMM 2020, page 270–282, 2020.
  32. Network stack specialization for performance. In SIGCOMM 2014, page 175–186, 2014.
  33. Reverse engineering intel last-level cache complex addressing using performance counters. In Research in Attacks, Intrusions, and Defenses: 18th International Symposium, RAID 2015, pages 48–65, 2015.
  34. Sparsh Mittal. A survey of techniques for cache partitioning in multicore processors. ACM Computing Surveys (CSUR), 50:1–39, 2017.
  35. Probability and Computing: Randomized Algorithms and Probabilistic Analysis. 2005.
  36. OpenAirInterface Organization. OpenAirInterface 5G Core Network Project. https://openairinterface.org/oai-5g-core-network-project/. (Accessed on 2024-01).
  37. Shenango: Achieving high CPU efficiency for latency-sensitive datacenter workloads. In NSDI 2019, pages 361–378, 2019.
  38. Experimenting with localization management functions in 5g core networks. In MobiCom 2022, MobiCom ’22, page 806–807, 2022.
  39. 5G Core Networks: Powering Digitalization. Academic Press, 2019.
  40. Mobile network architecture evolution toward 5g. IEEE Communications Magazine, pages 84–91, 2016.
  41. Samsung. Samsung Achieves 305 Gbps on 5G UPF Core Utilizing Intel® Architecture. https://images.samsung.com/is/content/samsung/assets/global/business/networks/insights/white-papers/1217_samsung-achieves-305gbps-on-5g-upf-core-utilizing-intel-architecture/Wireless-Core-Intel-Samsung-Performance-Whitepaper-Design-1217.pdf. (Accessed on 2024-01).
  42. A reliable intelligent routing mechanism in 5g core networks. In MobiCom 2020, 2020.
  43. Dma cache: Using on-chip storage to architecturally separate i/o data from cpu data for improving i/o performance. In 2010 IEEE 16th International Symposium on High-Performance Computer Architecture (HPCA), pages 1–12, 2010.
  44. Packet chasing: Spying on network packets over a cache side-channel. In 47th ACM/IEEE Annual International Symposium on Computer Architecture (ISCA’20), pages 721–734, 2020.
  45. Resq: Enabling slos in network function virtualization. In NSDI 2018, page 283–297, 2018.
  46. Conor Walsh. Optimize memory usage in multithreaded data plane development kit (dpdk) applications, 2018.
  47. Conor Walsh. Optimize memory usage in multithreaded applications, 2020.
  48. Understanding i/o direct cache access performance for end host networking. Proceedings of the ACM on Measurement and Analysis of Computing Systems, pages 1–37, 2022.
  49. Swap: Effective fine-grain management of shared last-level caches with minimum hardware support. In 2017 IEEE International Symposium on High Performance Computer Architecture (HPCA), pages 121–132, 2017.
  50. Intel performance counter monitor - a better way to measure cpu utilization, 2017.
  51. Granularnf: Granular decomposition of stateful nfv at 100 gbps line speed and beyond. SIGMETRICS Perform. Eval. Rev., page 46–51, 2022.
  52. Dcaps: Dynamic cache allocation with partial sharing. In Proceedings of the Thirteenth EuroSys Conference, EuroSys ’18, 2018.
  53. Dcat: Dynamic cache management for efficient, performance-sensitive infrastructure-as-a-service. In Proceedings of the Thirteenth EuroSys Conference, EuroSys ’18, 2018.
  54. A new control plane for 5g network architecture with a case study on unified handoff, mobility, and routing management. IEEE Communications Magazine, pages 76–85, 2014.
  55. Don’t forget the I/O when allocating your LLC. In 48th ACM/IEEE Annual International Symposium on Computer Architecture, ISCA 2021, pages 112–125, 2021.
  56. ZTE. ZTE 5G Core Network. https://sdnfv.zte.com.cn/en/products/VNF/5G-core-network. (Accessed on 2024-01).

Summary

We haven't generated a summary for this paper yet.

Ai Sparkles Streamline Icon: https://streamlinehq.com Summarize Now
X Twitter Logo Streamline Icon: https://streamlinehq.com

Tweets