Re: Review of draft-bberry-pppoe-credit (Informational)

in-line

Vernon Schryver wrote:
>>Cc: [email protected], [email protected], [email protected],
>>        [email protected], [email protected]
>>From: Fred Baker <[email protected]>
>>To: Vernon Schryver <[email protected]>
> 
> 
> 
>>AFAIK, it isn't. It is using the Ethernet as a local interface  
>>between a router and an interface device that has a number of special  
>>characteristics and is implemented in a separate chassis. Call it  
>>"system engineering" if you like. The actual protocol over the  
>>airwaves is PPP as known and loved, far more compact than PPPOE.  
>>PPPOE is simply used to manage the local interface between the router  
>>and the radio interface.
> 
> 
> Ok, so why not assign IP addresses (perhaps by RFC 1918) to those 
> those boxes with radios and PPPoE links to access controllers?
> 
> If this particular use of PPPoE as an inter-network protocol makes
> sense, where does does using PPPoE as a replacement for IP not make
> sense?
> 
> If PPPoE is not going over the radio waves, why are the access controllers
> distinct boxes from the boxes ("radio interfaces"?) that connect PPPoE
> to the real ether?

We needed a means to control the eb-n-flow of packets between the
router and the radio independent of PPP.  So each router-radio,
radio-radio (RLP) and radio-router link provides flow
control relative to its segment.  PPPoE fit the overall model.


> 
> I suspect they are distinct boxes from the mention of the use of PPPoE
> as a (de)multiplexing tool, which gets back to the evil of using PPPoE
> as an inter-network protocol instead of IP.
> 
> 
> Vernon Schryver    [email protected]
> 

_______________________________________________
Pppext mailing list
[email protected]
https://www1.ietf.org/mailman/listinfo/pppext