Re: ppp - I'll stop discussing (was: [16NG] Re: multicast and IPv6 over ETHCS)

[Alexandru Petrescu <[e-mail]>]

Raj, thanks for the email.  I will not post further on this topic, not
because I don't want to reply to you, but because I've already got
private indication (from Chairs and AD) that I should stop ppp
discussion, there's nothing new or relevant to the IP-over-IP/ETHCS.

Alex

Basavaraj Patil wrote:
> The PPP that you are talking about is most likely not between your PC
>  card (that is implementing GPRS/HSDPA) and the GGSN.

I'm sorry but I don't care.  It's about my IP laptop and a certain
network.  IETF doesn't define what's GGSN nor SGSN.  I believe (not
sure) no deployment has strict GGSN/SGSN separation (separation
suggested by SDO and graduate textbooks).

> You could be doing PPP over a PDP context that is established between
>  the PC card and the GGSN. Question is, does the PPP terminate at the
>  GGSN? Of course not knowing what kind of service you are using or 
> details, I can only speculate.

YEs, I agree.  I no longer bother to even speculate, I'm torn down by
these speculations, too many traceroutes, too many pings, on every
operator's different topology.

I prefer to see it as a network into which I can ppp, with a certain
qos.  This is the typical point of view adopted by a end host.

>> Do you mean a SDO spec says that PDP context is set up?  And that 
>> that SDO document doesn't say 'ppp is used'?  I agree with this 
>> part.
> 
> PPP is an overhead and not required to transport IP packets in GPRS.

Well that's what many people do (see 'globally' interpreted at the end
of this message).  Not even Compression modules are used.  It works.

> PDP contexts are the mechanism used for carrying IP packets.

There's no IP-over-PDPcontext StdsTrack RFC, sorry.

I too run IPv6-over-RS232 w/o an RFC telling me how, it works, doesn't
upset anybody, and solves today's context sometimes (think really small
custom stuff).

SDO can do that too with its PDP contexts, unless it upsets somebody.

>>> PPPoE is NOT used in GPRS/UMTS or HSDPA.
>> What do you call a PPP ConfigReq within a MAC header with 48bits 
>> addresses, as displayed by Ethereal?  And with a pcmcia card saying
>>  hsdpa.
> 
> Your Ethereal may not really be have the drivers to read what is 
> being sent by the HSDPA card... I don't know.

WEll it does have necessary drivers telling me that.  It's enough for
making run the IPv6 stack, no more needed.

> As I said above, tell me where is PPP being terminated in the 
> network, i.e the GPRS core?

Sorry, I don't know precisely.  It terminates somewhere in the network.
  I could tell you all the experiments I've done to find that precise
question you ask, and how they were voided by simply connecting to
another operator.

>>> PDP context != PPP or PPPoE
>> I'm sorry.  PDP context is not something defined at IETF, a 
>> different beast, can't compare.
> 
> And I am not asking you to do so. I was just correcting a statement 
> about PPP usage that you made. And it appeared to be a globally 
> applicable statement to all GPRS/UMTS/HSDPA deployments.

Ok, what's globally applicable.  To me globally applicable is the
following: google APN name and find the name of operators that will
allow one to ppp into their cell networks.  One will find cell operators
all over the world.  I've personally tried a bunch of them in Europe and
the 2-3 available in US.  GPRS and UMTS and HSDPA.  Is this globally 
applicable to you?

Alex

_______________________________________________
16NG mailing list
[e-mail]
https://www1.ietf.org/mailman/listinfo/16ng