Aprendizados de uma avaliação experimental do eXpress Data Path (XDP) e Data Plane Development Kit (DPDK) em um ambiente de Computação em Nuvem

Autores

DOI:

https://doi.org/10.33448/rsd-v11i12.34200

Palavras-chave:

XDP; DPDK; Kernel Linux; Frameworks de Processamento de Pacotes; Computação em nuvem.

Resumo

A popularização do “kernel” Linux nos servidores de data centers tornou-se um fator significativo a ser considerado ao desenvolver ou implantar aplicações de rede. Novos “frameworks de processamento de pacotes” surgiram para garantir o rápido processamento de pacotes dentro desses servidores que dependem do “kernel” Linux, sendo o XDP e o DPDK os dois principais. Paralelamente, os ambientes de computação em nuvem dominaram o cenário de data centers, trazendo a virtualização para o centro destes servidores. Esta pesquisa compara os processadores de pacotes DPDK e XDP quando submetidos a um ambiente de computação em nuvem com E/S, CPU e carga de rede. Primeiro descrevemos uma arquitetura que implementa os frameworks e emula um ambiente de nuvem. Medimos o rendimento e a perda de pacotes enquanto variamos o número de VMs concorrentes, o uso de CPU pelos frameworks e o tamanho dos pacotes e apresentamos nossos resultados sob a forma de aprendizados. Concluímos que o ambiente virtual pode impactar o desempenho de ambos os frameworks, dependendo do tamanho do pacote recebido e da carga de uso de recursos que o ambiente em nuvem realiza. Um caso significativo a ser destacado é quando o ambiente em nuvem realiza carga de rede, e os servidores recebem pacotes pequenos, criando tanto a degradação do rendimento quanto o aumento da perda de pacotes. Concluímos nossos experimentos conduzindo testes estatísticos de hipóteses não paramétricas para aprimorar nossas conclusões e analisar mais de perto os resultados.

Referências

Cilium. (2022). Bpf and xdp reference guide. https://docs.cilium.io/en/latest/bpf/

Incubator, F. (2022). Katran - a high performance layer 4 load balancer. https://github.com/facebookincubator/katran

Cilium. (2022). What is cilium? https://cilium.io/get-started/

Belay, A., Prekas, G., Klimovic, A., Grossman, S., Kozyrakis, C., & Bugnion, E. (2014). Ix: A protected dataplane operating system for high throughput and low latency. In Proceedings of the 11th USENIX Conference on Operating Systems Design and Imple-mentation, OSDI’14, page 49–65, USA. USENIX Association.

Bruijn, W. D. & Dumazet, E. (2017). sendmsg copy avoidance with msg zerocopy. In NetDev, The Technical Conference on Linux Networking, Netdev 2.1, Montreal, Canada. Netdev.

Corder, G. W. & Foreman, D. I. (2011). Nonparametric statistics for non-statisticians.

Dantas, R., Sadok, D., Flinta, C., & Johnsson, A. (2015). Kvm virtualization impact on active round-trip time measurements. In 2015 IFIP/IEEE International Symposium on Integrated Network Management (IM), pages 810–813.

Oliveira Filho, A. T., Freitas, E., do Carmo, P. R., Sadok, D. H., & Kelner, J. (2022). An experimental investigation of round-trip time and virtualization. Computer Communications, 184:73–85.

Emmerich, P., Gallenmüller, S., Raumer, D., Wohlfart, F., & Carle, G. (2015). Moongen: A scriptable high-speed packet generator. In Proceedings of the 2015 Internet Measurement Conference, IMC ’15, page 275–287, New York, NY, USA. Association for Computing Machinery.

Fabre, A. (2018). L4drop: Xdp ddos mitigations. https://blog.cloudflare.com/l4drop-xdp-ebpf-based-ddos-mitigations/

Foundation, L. (2018). Data plane development kit. https://www.dpdk.org/

Freitas, E. (2021). Experimental evaluation on packet processing frameworks under virtual environments. Master’s thesis, Universidade Federal de Pernambuco, Recife.

Gallenmüller, S., Emmerich, P., Wohlfart, F., Raumer, D., and Carle, G. (2015). Comparison of frameworks for high-performance packet io. In Proceedings of the Eleventh ACM/IEEE Symposium on Architectures for Networking and Communications Systems, ANCS ’15, page 29–38, USA. IEEE Computer Society.

Group, D. P. D. K. (2017a). Dpdk docs: Poll mode driver. http://doc.dpdk.org/guides/prog_guide/poll_mode_drv.html

Group, D. P. D. K. (2017b). Dpdk docs: Programmer’s guide, mempool library. https://doc.dpdk.org/guides/prog_guide/mempool_lib.html

Group, D. P. D. K. (2017c). Dpdk docs: Programmer’s guide, overview. http://doc.dpdk.org/guides/prog_guide/overview.html

Group, D. P. D. K. (2017d). Dpdk docs: Programmer’s guide, ring library. https://doc.dpdk.org/guides/prog_guide/ring_lib.html

Han, S., Jang, K., Park, K., & Moon, S. (2010). Packetshader: A gpu-accelerated software router. In Proceedings of the ACM SIGCOMM 2010 Conference, SIGCOMM ’10, page 195–206, New York, NY, USA. Association for Computing Machinery.

Hohlfeld, O., Krude, J., Reelfs, J. H., R¨uth, J., and Wehrle, K. (2019). Demystifying the performance of xdp bpf. In 2019 IEEE Conference on Network Softwarization (NetSoft), pages 208–212.

-Jørgensen, T., Brouer, J. D., Borkmann, D., Fastabend, J., Herbert, T., Ahern, D., & Miller, D. (2018). The express data path: Fast programmable packet processing in the operating system kernel. In Proceedings of the 14th International Conference on Emerging Networking EXperiments and Technologies, CoNEXT ’18, page 54–66, New York, NY, USA. Association for Com-puting Machinery.

IDG (2020). Cloud computing survey. Executive summary, IDG. https://resources.foundryco.com/download/cloud-computing-executive-summary

Jeong, E., Wood, S., Jamshed, M., Jeong, H., Ihm, S., Han, D., & Park, K. (2014). mtcp: a highly scalable user-level TCP stack for multicore systems. In 11th USENIX Symposium on Networked Systems Design and Implementation (NSDI 14), pages 489– 502, Seattle, WA. USENIX Association.

Kourtis, M.-A., Xilouris, G., Riccobene, V., Mc-Grath, M. J., Petralia, G., Koumaras, H., Gardikis, G., & Liberal, F. (2015). Enhancing vnf performance by exploiting sr-iov and dpdk packet processing acceleration. In 2015 IEEE Conference on Network Function Virtualization and Software Defined Network (NFV-SDN), pages 74–78.

Liu, J. (2010). Evaluating standard-based self-virtualizing devices: A performance study on 10 gbe nics with sr-iov support. In 2010 IEEE International Symposium on Parallel Distributed Processing (IPDPS), pages 1–12.

Liu, M., Cui, T., Schuh, H., Krishnamurthy, A., Peter, S., & Gupta, K. (2019). Offloading distributed applications onto smartnics using ipipe. In Proceedings of the ACM Special Interest Group on Data Communication, SIGCOMM ’19, page 318–333, New York, NY, USA. Association for Computing Machinery.

Miano, S., Doriguzzi-Corin, R., Risso, F., Siracusa, D., & Sommese, R. (2019). Introducing smartnics in server-based data plane processing: The ddos mitigation use case. IEEE Access, 7:107161–107170.

Monnet, Q. (2016). Dive into bpf: a list of reading material. https://qmonnet.github.io/whirl-offload/2016/09/01/dive-into-bpf/#what-is-bpf

Rizzo, L. (2012). netmap: A novel framework for fast packet i/o. In 2012 USENIX Annual Technical Conference (USENIX ATC 12), pages 101–112, Boston, MA. USENIX Association.

Rybczyńska, M. (2019). Bounded loops in bpf for the 5.3 kernel. https://lwn.net/Articles/794934/

Scholz, D., Raumer, D., Emmerich, P., Kurtz, A., Lesiak, K., & Carle, G. (2018). Performance implications of packet filtering with linux ebpf. In2018 30th International Teletraffic Congress (ITC 30), volume 01, pages 209–217.

Soares, L. & Stumm, M. (2010). Flexsc: Flexible system call scheduling with exception-less system calls. In Proceedings of the 9th USENIX Conference on Operating Systems Design and Implementation, OSDI’10, page 33–46, USA. USENIX Association.

Surveys, W. W. W. W. T. (2021). Usage of operating systems broken down by data center providers. https://w3techs.com/technologies/cross/operating_system/data_center

Tsuna (2010). How long does it take to make a context switch? https://blog.tsunanet.net/2010/11/how-long-does-it-take-to-make-context.html

Van Tu, N., Yoo, J.-H., & Hong, J. W.-K. (2019). evnf - hybrid virtual network functions with linux express data path. In 2019 20th Asia-Pacific Network Operations and Management Sym-posium (APNOMS), pages 1–6.

Vieira, M. A. M., Castanho, M. S., Pacífico, R. D. G., Santos, E. R. S., Júnior, E. P. M. C., & Vieira, L. F. M. (2020). Fast packet processing with ebpf and xdp: Concepts, code, challenges, and applications. ACM Comput. Surv., 53(1).

Wu, W., Crawford, M., & Bowden, M. (2007). The performance analysis of linux networking – packet receiving. Computer Communications, 30(5):1044 – 1057. Advances in Computer Commu-nications Networks.

Downloads

Publicado

06/09/2022

Como Citar

FREITAS, E.; OLIVEIRA FILHO, A. T. de; CARMO, P. R. X. do .; SADOK, D. F. H.; KELNER, J. Aprendizados de uma avaliação experimental do eXpress Data Path (XDP) e Data Plane Development Kit (DPDK) em um ambiente de Computação em Nuvem. Research, Society and Development, [S. l.], v. 11, n. 12, p. e26111234200, 2022. DOI: 10.33448/rsd-v11i12.34200. Disponível em: https://rsdjournal.org/index.php/rsd/article/view/34200. Acesso em: 27 set. 2024.

Edição

Seção

Ciências Exatas e da Terra