CDN Latency Compared

Posted by Yejun on March 21, 2009 Leave a comment (4) Go to comments

I compared some puplar commercial CDN’s global latency by using just-ping.

The raw test result you can downloaded from here.

Result Summary:

.

NetworkAverageMedian

.

Akamai9.672.9

.

Aol CDN9.224

.

Panthercdn62.3612.7

.

LimeLight58.0413.2

.

Mosso Cloud Files56.8113.5

.

Amazon Cloudfront62.8218.6

.

Google Homepage53.5323.15

.

Cachefly54.5728.2

.

Google Ajax Library54.9628.5

.

Homemade CDN76.3129

.

Yahoo Homepage82.7738.4

.

Google App Engine76.0342.8

.

US East130.1196.9

.

SimpleCDN142.84100.8

.

US West156.32165.4

Chart:

CDN Latency Comparison

Notes:

  1. Akamai
  2. Aol CDN is served by Akamai
  3. Panther Express
  4. Limelight Networks
  5. Mosso is served by Limelight
  6. Amazon CloudFront Images on this page are served by CloudFront
  7. Google
  8. CacheFly
  9. Google AJAX Libraries
  10. This page is served by my own homemade CDN, you can test speed here
  11. Yahoo
  12. Google App Engine
  13. Single location ip in New Jersey USA listed here for comparison purpose
  14. SimpleCDN
  15. Single location ip in California USA listed here for comparison purpose

Disclamer: I am not affiliated to any company mentioned above. Nor do I endorse the accuracy of these results.

Web, , , , ,
  • rg

    interesting test! although not sure how you did it. Did you ping their sites and assume that their website was served on their content delivery network? or did you have accounts with all of them and served a common video and/or image file? hopefully it is the later!

  • http://digital-post.blogspot.com/2009/07/plumbing-torrent.html Lalit

    Many ISP’s or Operators do not see value in investing in CDN and the Torrents are having a free ride on the network. If we could find a way for end-user to pay for Torrents then the operators will see value in investing in CDNs. I have suggested asymmetric billing strategy to make end-user pay for Torrents. Hope this will help in widespread use of network based caching as against client-based.

  • http://code.deepinspace.net/ Hrish

    @rg

    I think this is critical to understanding the results you have posted. Could you enlighten us?

  • http://adminsgoodies.com/when-does-a-cdn-content-delivery-network-become-worth-it/ When does a CDN (content delivery network) become worth it? – Admins Goodies

    [...] do so great. It is really hard to objectively quantify the performance of the various CDNs, but here is one attempt. Maybe I’m being unfair to SimpleCDN here, but they wouldn’t be my first [...]