Advertisement

The Journal of Supercomputing

, Volume 65, Issue 3, pp 1020–1036 | Cite as

Leveraging bandwidth improvements to web servers through enhanced network interfaces

  • Andrés Ortiz
  • Julio Ortega
  • Antonio F. DíazEmail author
  • Mancia Anguita
Article

Abstract

Markets nowadays demand applications that require high communication throughputs to reach their adequate levels of performance. Although the bandwidth of the network links has increased allowing multiple gigabits per second, taking advantage of these links accounts for a high communication overhead, and thus a lot of processor cycles are used for communication tasks, diminishing the processor cycles that remain available for the application.

In this paper, we evaluate the performance in web applications of a network interface that as it is distributed among the processors currently available in the node takes advantage of both the hardware (multiprocessor nodes and multicore architectures, as well as programmable network interface cards) and software elements present in the system, thus improving not only the effective communication throughputs and latencies, but also the capacity of the nodes to satisfy the requirements of the applications. Here, the usefulness of this distributed network interface to improve the performance of either static or dynamic web servers is shown. The ubiquity and the different computation/communication rates that can be found in web applications make the analysis of web servers interesting, as it could provide relevant conclusions about the efficiency of the different approaches to the design of high-performance network interfaces.

Keywords

Dynamic and static web servers Full-system simulation Multicore and multiprocessor nodes Network interface Offloading Onloading Web server evaluation 

Notes

Acknowledgements

This work has been funded by projects TIN2007-60587 and TIN2012-32039. The authors would like to thank the reviewers for their useful comments and suggestions.

References

  1. 1.
    Apache web server. http://httpd.apache.org
  2. 2.
    Benvenuti C (2005) Understanding Linux kernel internals. O’Reilly, Sebastopol Google Scholar
  3. 3.
    Bovet DP, Cesati M (2005) Understanding the Linux kernel. O’Reilly, Sebastopol Google Scholar
  4. 4.
    Brogioli M, Willman P, Rixner S (2006) Parallelization strategies for network interface firmware. In: Proc. of 4th workshop on optimization for DSP and embedded systems, ODES-4 Google Scholar
  5. 5.
    Cascón P, Ortiz A, Ortega J, Díaz AF, Roja I (2011) Accelerating network applications by distributed interfaces on heterogeneous multiprocessor architectures. J Supercomput 58(3):302–313 CrossRefGoogle Scholar
  6. 6.
    Chakraborty S et al (2003) Performance evaluation of network processor architectures: combining simulation with analytical estimation. Comput Netw 41:641–645 zbMATHCrossRefGoogle Scholar
  7. 7.
    Competitive Comparison. Intel I/O acceleration technology vs. TCP offload engine. http://www.intel.com/technology/ioacceleration/316126.pdf
  8. 8.
    de Bruijn W, Bos H (2008) Model-T: Rethinking the OS for terabits speeds. In: Workshop on high-speed networks (HSN2008), INFOCOM’2008 Google Scholar
  9. 9.
    Foong A, Fung J, Newell D (2004) An in-depth analysis of the impact of processor affinity on network performance. In: Proceedings of the ICON, November 2004 Google Scholar
  10. 10.
    GadelRab S (2007) 10-gigabit Ethernet connectivity for computer servers. IEEE MICRO 27(3):94–105 CrossRefGoogle Scholar
  11. 11.
    Hansen T, Mainkar V, Reeser P (2002) Performance comparison of dynamic web platforms. Comput Commun 26(8):888–898 Google Scholar
  12. 12.
  13. 13.
  14. 14.
    Jan H (2009) MiAMI: multi-core aware processor affinity for TCP/IP over multiple network interfaces”. In: 17th IEEE symposium on high performance interconnects, HOTI Google Scholar
  15. 15.
    Kim H-Y, Rixner S (2006) TCP offload through connection handoff. In: ACM Eurosys’06, pp 279–290 Google Scholar
  16. 16.
    Kim H, Pai VS, Rixner S (2003) Exploiting task-level concurrency in a programmable network interface. In: Proc of the ACM PPoPP’03 Google Scholar
  17. 17.
    Magnusson PS et al (2002) Simics: a full system simulation platform. Computer 35(2):50–58 CrossRefGoogle Scholar
  18. 18.
    Mogul JC (2003) TCP offload is a dumb idea whose time has come. In: 9th workshop on hot topics in operating systems (HotOS IX) Google Scholar
  19. 19.
    Nahum EM, Yates DJ, Kurose JF, Towsley D (1994) Performance issues in parallelized network protocols. In: Proc of the operating systems design and implementation, pp 125–137 Google Scholar
  20. 20.
    Narayanaswamy G, Balaji P, Feng W (2007) An analysis of 10-gigabit Ethernet protocol stacks in multicore environments. In: 15th IEEE symp on high-performance interconnects (HOTI’07), pp 109–116 CrossRefGoogle Scholar
  21. 21.
    Ortiz A, Ortega J, Díaz AF, Prieto A (2006) Protocol offload evaluation using Simics. In: IEEE cluster computing, Barcelona, September, 2006 Google Scholar
  22. 22.
    Ortiz A, Ortega J, Díaz AF, Prieto A (2008) Comparison of offloading and onloading strategies to improve network interfaces. In: 16th Euromicro international conference on parallel, distributed and network-based processing, PDP 2008, February 2008, Toulouse Google Scholar
  23. 23.
    Ortiz A, Ortega J, Díaz AF, Cascón P, Prieto A (2009) Protocol offload analysis by simulation. J Syst Archit 55:25–42 CrossRefGoogle Scholar
  24. 24.
    Ortiz A, Ortega J, Díaz AF, Prieto A (2009) A new offloaded/onloaded network interface for high performance communication. In: 17th Euromicro international conference on parallel, distributed and network-based processing, PDP 2009, February 2008, Weimar Google Scholar
  25. 25.
    Ortiz A, Ortega J, Díaz AF, Prieto A (2009) Network interfaces for programmable NICs and multicore platforms. Comput Netw. doi: 10.1016/j.comnet.2009.09.11 Google Scholar
  26. 26.
    Pacifici G, Segmuller W, Spreitzer M, Tantawi A (2008) CPU demand for web serving: measurement analysis and dynamic estimation. Perform Eval 65(6–7):531–553 CrossRefGoogle Scholar
  27. 27.
    Papaefstathiou I. et al (2004) Network processors for future high-end systems and applications. IEEE MICRO 24(5):7–9 CrossRefGoogle Scholar
  28. 28.
    Ravi J, Yu Z, Shi W (2009) A survey on dynamic web content generation and delivery techniques. J Netw Comput Appl 32:943–960 CrossRefGoogle Scholar
  29. 29.
    Reeser P, Hariharan R (2002) An analytic model of web servers in distributed computing environments. Telecommun Syst 21(2–4):283–299 CrossRefGoogle Scholar
  30. 30.
    Regnier G et al (2004) TCP onloading for data center servers. Computer 37(11):48–58 CrossRefGoogle Scholar
  31. 31.
    Salim J (2001) Beyond SoftNet. In: Proc 5th ann Linux showcase and conf. www.linuxshowcase.org/2001/full_papers/jamal/jamal.pdf
  32. 32.
    Shalev L, Marhervaks V, Machulsky Z, Biran G, Satran J, Ben-Yehuda M, Shimony I (2006) Loosely coupled TCP acceleration architecture. In: Proceedings of the 14th IEEE symposium on high-performance interconnects (HOTI) Google Scholar
  33. 33.
    Shivam P, Chase JS (2003) On the elusive benefits of protocol offload. In: SIGCOMM’03 workshop on network-I/O convergence: experience, lessons, implications (NICELI), August, 2003 Google Scholar
  34. 34.
    Thiele L et al (2002) Design space exploration of network processor architectures. In: Proc 1st workshop on network processors (8th int. symp. on high performance computer architecture), February, 2002 Google Scholar
  35. 35.
    Titchkosky L, Arlitt M, Williamson C (2003) Performance benchmarking of dynamic web technologies. In: 11th IEEE international symposium on modeling, analysis, and simulation of computer and telecommunications systems (MASCOTS’03), vol 20, p 250 Google Scholar
  36. 36.
    Vaidyanathan K, Panda DK Benefits of I/O acceleration technology (I/OAT) in clusters. Technical report, Ohio State Univ (OSU_CISRC-2/07-TR13) Google Scholar
  37. 37.
    Williamson C, Simmonds R, Arlitt M (2002) A case study of web server benchmarking using parallel WAN emulation. Perform Eval 49(1–4):111–127 zbMATHCrossRefGoogle Scholar
  38. 38.
    Wu W, Crawford M, Bowden M (2007) The performance analysis of Linux networking—packet receiving. Comput Commun 30(5):1044–1057 CrossRefGoogle Scholar
  39. 39.
    Yeager N, McGrath R (1996) Web server technology: the advanced guide for world wide web information providers. Morgan Kaufmann, San Francisco zbMATHGoogle Scholar

Copyright information

© Springer Science+Business Media New York 2012

Authors and Affiliations

  • Andrés Ortiz
    • 1
  • Julio Ortega
    • 2
  • Antonio F. Díaz
    • 2
    Email author
  • Mancia Anguita
    • 2
  1. 1.Departamento de Ingeniería de ComunicacionesUniversidad de MálagaMálagaSpain
  2. 2.Departamento de Arquitectura y Tecnología de ComputadoresUniversidad de GranadaGranadaSpain

Personalised recommendations