CDN is the biggest pile of shit for serving videos.
 

i have done my testing... performance sucks imho....

where did you test? Hit me up on icq later Im heading to the store for a bit.

sticky check your ICQ...

http://www.ottiesdesigns.com/master-baiters1.gif

ADG

I figured as much. lol I don't mind if he does it though.

actually i am serious... i have been testing lower bandwidth models for a bit.. i have been doing this to see the quality of different platforms on low bandwidth..

why.. because as we all know wireless is going to continue to grow as well as cell/pda/etc surfing...

so far cdn fails horribly... reg flash on any site blows the performance away..

Hmm not sure what you are talking about. CDN as in Content Delivery Network. What does wireless really have to do with that?

Most cdn networks are used to push larger videos not meant for cell phones, pda etc.

http://www.cavecreek.com/HD-Flash-CDN.php such as that. It requires the newest version of the flash player, and unlikely to work with most cell phones, smart phones etc. However that is mainly due to what that video is, which is h264 .mp4. Adobe flash player just recently updated to allow h264 support.

Hit me up on icq Im interested in what ya got goin.

wireless end user.. and i disagree that videos are not meant for cell/pda/etc..

also, with free wireless bound to b offered more and more places, i think the performance of cdn leaves a lot to be desired..

i have visited several cdn served sites and its all the same, performance on under 150-120k sucks..

what really sucks is the buffering lag compared to reg flash..

Something screwy there. I dont have any cell phone sized examples for you, and thats not really the market I am pushing. With using ours mainly to push HD video not really a cell phone market Im aiming at.

i think cdn is fine for higher bandwidth users, but i am not impressed for lower bandwidth users and i think reg flash outperforms cdn...

also, i am not just thinking cell.. i am thinking wi-fi...

Quite possibly. Could be an issue with the way it hands off which server to pull off of. Not sure how geo ip works with cell phones. The Ip could be registered several states away from the phone. I can find out for you tomorrow on that though what cdn issues might be faced due to that.

CDN has nothing to do with "reg flash" or any other "type" of content.

You can just as easily put "regular flash" (I assume you mean something akin to youtube?) files on CDN. If you mean streaming, you can as well - many CDN services support streaming.

Basically, CDN on average should increase throughput substantially, vs. single-point hosting. The technology your site uses to push the video should be irrelevant, it will still be faster (on average) on CDN than it would be via a single location. Basically, it's simple physics due to the speed of light :)

-Phil

yes i know this.. i am saying its not faster and buffering issues with lower bandwidth as opposed to flash on non cdn...

i say flash so instead of people asking what format i tested streaming, anyone who knows cdn should get what i am talking about...

Well, then either the CDN provider used sucks, or you simply are one of the very small percentage that your single-location provider happened to be closer than the CDN provider's closest location.

Basically, could be you were downloading from a local provider for your single-location test, and going across country (which shouldn't happen of course) for the CDN test.

Also could be the CDN provider is overloaded in that location, seeing technical issues, etc.

Basically, what you describe doesn't make any sense from a technical angle. While it happens on the small scale, on the large scale CDN very much speeds things up for the huge majority of your traffic.

Take a look at Youtube - their "hot" content is CDN'ed, and their "cold" content is not. The cold content here (on comcast) is slow as hell, in comparison to CDN.

It would certainly be interesting to see the metrics and methodology used for the testing though, as it may shed some light on the underlying reasons.