Anyone besides jra remember the last Super Bowl?
Better this year? Worse? I'm sure whomever is listening in would like to know as well.
http://www.multichannel.com/blogs/translation-please/multicast-unicast-and-super-bowl-problem
Well, in fact, the most recent Massive Failure was the webcast of the
Concert For Boston, on 5/31. They were using a vendor called LiveAlliance.tv,
who did not appear to be farming it out to Limelight or Akamai or Youtube, as
far as I could tell, and they apparently only figured for a scale 5 audience,
and then got more than 500k attempts.
They got rescued by a vendor named Fast Hockey who are an amateur hockey
webcast aggregator, I gather, and *are* an Akamai client.
My estimation is that the reason that webcasting will never completely
replace broadcasting is that -- because it is mostly unicast -- its
inherent complexity factor is a) orders of magnitude higher than bcast, and
b) *proportional to the number of viewers*. Like Linux, that doesn't scale.
And broadcasters are not prone to think of the world in a view where you
have to provide technical support to people just to watch your show.
"He's at the 40... the 30... the 20... this is gonna be the Super Bowl,
folks... the 10... [buffering]"
Cheers,
-- jra
Jay Ashworth wrote:
"He's at the 40... the 30... the 20... this is gonna be the Super Bowl,
folks... the 10... [buffering]"Cheers,
-- jra
lol...tnx Jay!
I was at an incentive auction discussion earlier in the week where it
was suggested that the broadcasters see a rosy future with ATSC
beaming to mobile, but there is still work to be done.
From: "Michael Painter" <tvhawaii@shaka.com>
Anyone besides jra remember the last Super Bowl?
Better this year? Worse?
I'm sure whomever is listening in would like to know as well.http://www.multichannel.com/blogs/translation-please/multicast-unicast-and-super-bowl-problem
Well, in fact, the most recent Massive Failure was the webcast of the
Concert For Boston, on 5/31. They were using a vendor called LiveAlliance.tv,
who did not appear to be farming it out to Limelight or Akamai or Youtube, a..
far as I could tell, and they apparently only figured for a scale 5 audience,
and then got more than 500k attempts.
Such a common story. ..
They got rescued by a vendor named Fast Hockey who are an amateur hockey
webcast aggregator, I gather, and *are* an Akamai client.My estimation is that the reason that webcasting will never completely
replace broadcasting is that -- because it is mostly unicast -- its
inherent complexity factor is a) orders of magnitude higher than bcast, and
b) *proportional to the number of viewers*. Like Linux, that doesn't scale.
This is the primary reason companies including Internap, Peer1 and XO (The list goes on and on, and includes several company that only provide CDN services) all used to run their own CDN networks and now all three have outsourced this CDN service / sold their customers to Limelight. Edgecast even sold off all their services in Asia and just runs a US based CDN.
The general policy in data centres has been 30 - 40% utilisation to allow for bursting and unexpected temporary increases, in CDN its more like 5 - 10% especially when you are a CDN for hire you really can't make any predictions about what your customers might do. Its common for CDN's to have entire rack's sitting powered off that only need to be powered up to join the cluster, our company has multiple full racks per data centre just a alert to the NOC staff or email away from being turned on.