V K EY D ERIVATIONS AND G UIDELINES In this section we present guidelines

V k ey d erivations and g uidelines in this section

This preview shows page 12 - 14 out of 18 pages.

V. K EY D ERIVATIONS AND G UIDELINES In this section, we present guidelines concerning the dif- ferent techniques that were detailed in the previous sections. They include general guidelines and caveats that should be taken into account when designing and developing VNFs (Section V-A ). Furthermore, we provide guidelines that are specific to the two main categories of acceleration techniques that are covered in this work, i.e., software (Section V-B ) and hardware (Section V-C ) acceleration. Finally, we demonstrate how these guidelines can be applied to the exemplary NGFW function (Section V-D ).
Image of page 12

Subscribe to view the full document.

13 A. Guidelines for Design and Development of Network Func- tions Ideally, as outlined in Section II , the VNF ecosystem should be designed based on a detailed performance requirement specification for the intended deployment scenario. However, due to internal policies or previous design choices, the es- sential execution environment is often already given when the performance becomes constraining. In the following, we concentrate on this deployment optimization scenario when the task is to improve VNF performance for a given envi- ronment. We highlight best practices for the VNF adoption, derive corresponding suggested implementation patterns , and summarize caveats in antipatterns , which should be avoided. 1) Know relevant performance metrics: To bootstrap the performance optimization process, it is necessary to be aware of the network functions, the current workload, the resource utilization of the whole ecosystem, as well as the underlying network topology. 3 Pattern: Deployment of dedicated monitoring and ver- ification tools providing up-to-date reports about the actual performance and resource descriptors in the system. 7 Antipattern: Inaccurate view due to missing monitoring information and insufficient insights may result in resource over-provisioning, e.g., by vertical scaling. 2) Optimize for the most constraining resource: Apply the “Pareto-principle” of software optimization to identify the 10% of the code affecting the most critical resource (e.g., most of the running time is spent, most of the memory is used) and apply the previously outlined software and hardware acceleration techniques. Table I is a good starting point to identify the bottleneck resources for the specific VNF type. 3 Pattern: Repeat the following performance optimiza- tion loop: (1) Identify per-VNF critical resources; (2) ob- tain deployment-specific insights using monitoring tools; (3) optimize for the most constraining resource by choosing appropriate software or hardware acceleration; (4) test and evaluate the results. 7 Antipattern: Rewriting a VNF for using a specific acceler- ation technique will improve one specific bottleneck resource . For instance, [ 102 ] shows that SGW acceleration with DPDK improves raw packet-per-second performance while user space table lookup remains a bottleneck, resulting in poor perfor- mance as soon as the table is sufficiently populated.
Image of page 13
Image of page 14
  • Spring '16

What students are saying

  • Left Quote Icon

    As a current student on this bumpy collegiate pathway, I stumbled upon Course Hero, where I can find study resources for nearly all my courses, get online help from tutors 24/7, and even share my old projects, papers, and lecture notes with other students.

    Student Picture

    Kiran Temple University Fox School of Business ‘17, Course Hero Intern

  • Left Quote Icon

    I cannot even describe how much Course Hero helped me this summer. It’s truly become something I can always rely on and help me. In the end, I was not only able to survive summer classes, but I was able to thrive thanks to Course Hero.

    Student Picture

    Dana University of Pennsylvania ‘17, Course Hero Intern

  • Left Quote Icon

    The ability to access any university’s resources through Course Hero proved invaluable in my case. I was behind on Tulane coursework and actually used UCLA’s materials to help me move forward and get everything together on time.

    Student Picture

    Jill Tulane University ‘16, Course Hero Intern

Ask Expert Tutors You can ask 0 bonus questions You can ask 0 questions (0 expire soon) You can ask 0 questions (will expire )
Answers in as fast as 15 minutes