At 03:11 PM 5/29/2007, Robert G. Brown wrote:
On Tue, 29 May 2007, Jim Lux wrote:
So how's Verizon going to amortize the $900 installation cost of my FiOS?
As an investment. So that they "win". Because if they don't, others
will and they'll lose. And yeah, ultimately YOU will pay for it, but
maybe not all at once and up front.
Speaking of your other issues -- one possible solution is media
replication and media servers. OK, so to be able to deliver (say) the
2000 top DVD titles on demand requires 10-20 TB of storage. That's a
trivial investment, really. Disk (even RAID disk) is perhaps $0.25/GB.
Storing a movie costs at MOST $5-10 -- small compared to the capital
investment required to sell them or rent them on physical media. And
disk is ever cheaper, servers ever faster.
I suspect the problems are not technological, but business. Say you
put a movie cache in a vault at the end of every block (or with
whatever density is needed). Who owns the vault? Who owns the
content stored in the vault? Does Verizon charge Sony Entertainment
to store Sony's movies for future distribution? Who gets to bill the
customer? Is Sony reimbursed when the customer actually watches it,
when Verizon issues the invoice, when Verizon gets paid, or
speculatively, up front. (The cable TV model has cable companies
paying, essentially in advance, for the channels they carry, with a
lot of bundling and package deals from vendor to cable co, who then
repackages to the consumer.. e.g. We'll sell you ESPN, but you also
have to carry Midnight Preschool channel and the Paint Drying
channel, so you burn 3 channels worth of bandwidth to carry the most
customers want to pay for.) Who's responsible for authenticating the
users of the content in that vault?
So if one locates "stores" of basically all the titles one might wish to
deliver that auto-replicate on demand while solving an "interesting"
problem in provisioning and optimization throughout communities (with a
suitable tree structure or network) one can avoid a lot of resource
contention on the aggregate backbone. There are already companies
trying to move into this space using the limited bw already available --
however, centralized distribution models probably will not scale.
Distributed stores do work (e.g. Google) but have challenging
non-technical aspects as described above.
James Lux, P.E.
Spacecraft Radio Frequency Subsystems Group
Flight Communications Systems Section
Jet Propulsion Laboratory, Mail Stop 161-213
4800 Oak Grove Drive
Pasadena CA 91109
tel: (818)354-2075
fax: (818)393-6875
_______________________________________________
Beowulf mailing list, [email protected]
To change your subscription (digest mode or unsubscribe) visit
http://www.beowulf.org/mailman/listinfo/beowulf
