I’m amazed at the companies that get funding. All that limelight does is offload a lot of serving issues but at the same time it dramtically slows down your sites response time.
Here I am pinging limelights Myspace image domain. I am using Telus which is the largest telco in western canada…
Pinging myspace-414.vo.llnwd.net [69.28.154.221] with 32 bytes of data
Reply from 69.28.154.221: bytes=32 time=106ms TTL=56
Reply from 69.28.154.221: bytes=32 time=104ms TTL=56
Reply from 69.28.154.221: bytes=32 time=104ms TTL=56
ping images.plentyoffish.com
Pinging cache.peer1.net [69.90.111.120] with 32 bytes of data:
Reply from 69.90.111.120: bytes=32 time=7ms TTL=59
Reply from 69.90.111.120: bytes=32 time=8ms TTL=59
Reply from 69.90.111.120: bytes=32 time=7ms TTL=59
I am using a caching network in canada, The Response times are over 14 times faster then myspace. Is it really any wonder that American companies don’t grow as fast outside of the US? By using a company like limelight you are esentially crippling your service outside of the USA.
Just for fun i’m going to ping Userplane which is in the same datacenter as myspace.
Pinging www.userplane.com [8.3.208.201] with 32 bytes of data:
Reply from 8.3.208.201: bytes=32 time=38ms TTL=246
Reply from 8.3.208.201: bytes=32 time=38ms TTL=246
I tested limelight for a day, I had sooo many complaints across Canada and such brutal response times compared to what i previously had I had to revert back Asap.
Maybe they can use that money and build a system that actually works.
July 26, 2006 at 5:33 pm |
Markus, are there any CDNs like Limelight that you like or would recommend?
July 26, 2006 at 9:39 pm |
Crazy, you can’t judge a CND by it’s ping time
July 26, 2006 at 11:25 pm |
The fact of the matter is, you cannot point to increased latency and say the system is broken. The system is designed to spread the load across multiple pops. It’s not designed to be the least latent network, nor should it be. It specifically addresses the need for businesses to spread a network load across multiple areas of the globe for better overall service. You spouting off in an uneducated manner is pretty damn funny to me, since you obviously have the networking knowlege of a tier 1 tech support person. Please read up on how the process works, as well as how networking in general works BEFORE you spout off on your blog about it. Thanks.
July 27, 2006 at 3:09 am |
I know exactly how it works, unlike the posters commenting.
Try loading a page with 10-20 images with a latency of 100+ ms vs 9ms. The difference is speed is EXTREMELY noticable.
That is like saying google is completely clueless and wasting billions attempting to build datacenters around the world so they can shave off 20 to 30 ms off the latency.
July 28, 2006 at 12:07 am |
We had to move out of the MySpace shared datacenter – we are in a totally separate host.. Good thing considering they were back down on Sat/Sunday!
August 7, 2006 at 3:46 pm |
I pinged MySpace and PlentyOfFish from where I live which is about 30 miles outside of Downtown LA, on a 3-4mbit connection. And I am getting:
MySpace(myspace-414.vo.llnwd.net):
Reply from 69.28.154.221: bytes=32 time=92ms TTL=53
Reply from 69.28.154.221: bytes=32 time=90ms TTL=53
Reply from 69.28.154.221: bytes=32 time=90ms TTL=53
Reply from 69.28.154.221: bytes=32 time=92ms TTL=53
Average = 91ms
PlentyOfFish(images.plentyoffish.com):
Reply from 64.34.249.141: bytes=32 time=26ms TTL=44
Reply from 64.34.249.141: bytes=32 time=23ms TTL=44
Reply from 64.34.249.141: bytes=32 time=22ms TTL=44
Reply from 64.34.249.141: bytes=32 time=24ms TTL=44
Average = 23ms
February 8, 2007 at 2:56 pm |
How does this compare to other image serving network such as Akamai or does anyone has an idea of how these two compare?
Thanks
May 11, 2009 at 12:14 am |
party people in the house say hooooooooooo!
August 6, 2010 at 7:06 pm |
Maybe they can use that money and build a system that actually works.