Re: Multicast access

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



>>As has been pointed out, this is a little more complicated than just
>>the choice of client, in particular multicast is not widely available
>>to the "average" Internet user.
>>
>>But I still find it ironic that I can watch a webcast from an ICANN
>>meeting but I am unable to do the same for an IETF meeting (until after
>>the fact). That is but one example.

Pretty standard response every time this comes up:

0.  Arguing the merits of multicast is really a separate issue, but
some facts:  (1) the MBone is long dead,  (2) multicast is a highly
successive (revenue generating) service in a suprising number of 
enterprises, (3) multicast is certainly NOT ubiquitous in the 
wide-area infrastructure, but people really ought to understand its 
deployment by looking at measured statistics, and (4) before bashing 
"the MBone", make sure you understand the huge challenge that was 
undertaken (compare to moving the entire Internet to IPv6) and 
understand that there are a lot of non-technical challenges that 
were not properly envisioned.

1.  As Joel pointed out, the single reason for using multicast is
scalability.  We simply don't have enough bandwidth to support X
(where X > 5-10) simultaneous streams of the same content from the hotel.
A very fine idea is to have an exploder or some sort of server available
off-site.  We send one stream to them and it replicates.  Volunteers?

2.  The whole multicast effort is run on a shoe-string budget.  Until
now, and maybe even still now, there seems very little willingness by
remote users to pay for even a hypothetically perfpect service.  What 
everyone needs to realize is that of what is currently done, almost 
zero $$$ of IETF registration money goes to pay for it.  As Harald 
mentioned, it is time donated by UofO (and others), it is a grant from 
Cisco, and it is money from ISOC.

3.  Just some back of the envelope numbers:  you want every session 
encoded (even single camera) and available by unicast, I would estimate 
this to cost about $15K per meeting plus equipment (assuming someone
is willing to do replicated service for free).  Given a replacement
time for the equipment of three years (reasonable, especially since
a lot of the equipment doesn't travel well) and an esimated cost
of about $50K, that means, per meeting we are talking about $20K.

$100 per remote attendee = 200 attendees
$500 per remote attendee = 40 attendees

A bit tough to support but possibly doable.

---------------------

And if you are STILL reading...  as Harald sent in an email, we are
approaching the end of the grant period, so lots of opportunity for
recommendations.

-Kevin





[Index of Archives]     [IETF Annoucements]     [IETF]     [IP Storage]     [Yosemite News]     [Linux SCTP]     [Linux Newbies]     [Fedora Users]