Re: [EFM] Banana networks

[Alan Levy <refconstandard@yahoo.com>]

Thanks, Geoff.  Quite correct... nature abhors a
vacuum.  I wish someeone would explain the physics
part to the FCC.  Back to lurker mode.

Best, 

Alan Levy

--- Geoff Thompson <gthompso@nortelnetworks.com>
wrote:
> Roy-
> 
> At 08:54 PM 12/10/2002 -0600, Roy Bynum wrote:
> >Geoff,
> >
> >A service subscription network does not work like a
> >privately owned LAN facility.
> 
> Currently true. Therein lies a great part of the
> problem that some of us are trying to solve.
> Carriers have formulated the tariff/revenue model
> forever and it is based on the "old" model that 
> bandwidth is scarce and precious. Customers and us 
> cranks from the LAN industry both want the WAN to 
> look like a LAN
> 
> >In a subscription network, there is no such thing
> >as "excess bandwidth".  A copper facility will be 
> >provisioned and operate at the maximum that the 
> >distance attenuation will allow. Fiber facilities 
> >will be provisioned to provide the maximum service
> >bandwidth that the customer is willing to pay for. 

> >When more than one customer can be put on the 
> >fiber facility, the service provider will provision
> >the maximum that the fiber will support, often for 
> >packet/frame facilities, the bandwidth will 
> >be over provisioned in the aggregate for all of the
> >customers.  This will hold true for either P2P or 
> >P2MP.  This is one of the economic realities 
> >of subscription networks.  The limitation of how
> >much bandwidth and how many customers is put on 
> >that bandwidth is strictly do to the physical 
> >limitations of the facilities and the willingness
> >of the sales and marketing people to keep putting 
> >people on the same facility.
> 
> You are assuming that the revenue/tariff model will
> not change. I believe (1) that it will, if not in 
> the US then elsewhere in the world in countries who 
> are willing to bet that a new model will provide 
> a "great lap forward" for them. and (2) if the 
> revenue model does not change then EFM not be a 
> success to any significant degree.
> 
> >The use of "lettuce", "bananas", and "peanuts" was
> >an effort to be able to indirectly discuss issues 
> >of service functionality requirements, which up 
> >until now, have not been fully explored.  Since
> >"services" delivery is the specific purpose of a 
> >subscription network, without such a discussion, 
> >how will the group know if they have achieved the 
> >basic objective of "Support Subscriber Access 
> >Network ..."
> >
> >Thank you,
> >Roy Bynum
> 
> The world's WAN space is awash in lit and unlit
> excess bandwidth. (Transatlantic lit capacity is 
> currently said to be 7X traffic) The telcom 
> industry has no way to sell a significant portion of
> it in the current revenue model. Ultimately they (or

> their scrap dealers, who won't care about the 
> current revenue. That's what makes scrap 
> dealers/scavengers different) will have to make a 
> decision about how to get some revenue out of it any

> way they can. High jitter, high bandwidth packet 
> delivery systems that are terrible for real time 
> voice and video provide the potential for this. If 
> the scrap dealers can sell it this way in the 
> regulated space I would expect them to do so. If the

> scrap dealers can sell in the unregulated space then

> they won't have to go to the trouble of making it 
> jitter.
> 
> This vast vacuum of available bandwidth in the core
> will suck players into the access space to provide 
> paths to the core.. Nature abhors a vacuum. Physics 
> always wins in the long haul (pun intended).
> 
> Geoff