Re: [dispatch] I-D Action: draft-malas-dispatch-sip-egress-route-00

[Dean Willis <[e-mail]>]

Hadriel Kaplan wrote:
> 
>> -----Original Message----- From: [e-mail]
>> [mailto:[e-mail]] On Behalf Of Adam Roach Sent:
>> Monday, March 15, 2010 1:33 PM
>> 
>> Daryl is working within the strictures of reproducing the business 
>> relationships of the PSTN in a SIP network. And the path that has
>> been taken, for better or worse, involves the heavy use of SBCs and
>> other devices that impose a rather strict control on the way things
>> are done. These devices largely ensure that the media path follows
>> the signaling path, even when that's not the optimal thing to do.
> 
> Optimal for whom, and in what way?

Optimal in the "internet" sense: the shortest/least used available
route, as would be described by the IP routing topology from address A to B.

> 
>> The chances of such networks deploying video, hi-def voice, or any
>> of the other really neat things that Greger mentions are slim to
>> none. This, of course, is NOT Daryl's fault. But it's where the
>> egress-route work comes from.
> 
> You seem to have some very odd (and typical IETF bull) preconceptions
> about what SBC's can do or are capable of doing.  SBC's in general
> support video, HD voice, or any of the other neat things Gregor
> mentions.  Policies provisioned on the SBC's may limit what media one
> can do, or in some cases if transcoding has to happen then obviously
> you're limited to whatever codecs the transcoders can handle (as any
> media gateway would be), but that's really tangential.

I believe Adam is referring to the fact that it's the old-school telecom
people who are building this sort of topology. They don't have a great
history of successful innovation. They occasionally try to offer a new
service, but only after it's been much more successfully deployed in an
"Internet" fashion by the new-school people. Does Youtube require SBCs?
How about iTunes? AOL's Instant messenger? Facebook?


> 
> 
>> Greger is speaking more towards the network you seem to have in
>> mind -- one in which end-to-end SIP calls are possible, without
>> SBCs or similar control boxes clobbering new services and mangling
>> the ability to convey novel media from one endpoint to another.
>> Networks where we don't need to define alternate connection models
>> for protocols simply to accommodate network elements that do things
>> that are explicitly disallowed by protocol specification.
> 
> As a B2BUA, what SBC's do is in no way disallowed by the protocols in
> the IETF. (although certainly I don't really care if they are ;)

Right. What they do is controlled by the policies and mindsets of those
who purchase and install them. They live in a different world, where
innovation is slow and painful. It's also a world where long-term
reliability and stability are heavily rewarded.

> 
> 
>> Which is why I support a simple, non-normative, informational draft
>>  about the form of URLs that one might want to put in ENUM. It
>> satisfies the requirements of the stagnant world, and doesn't even
>> have any normative impacts on any protocol anywhere. It's harmless,
>> and lets the people in the vibrant world concentrate on solving the
>> problem in a more general way without having to accommodate the
>> peculiarities of network elements living in the stagnant world.
> 
> If you live in this "vibrant world" the rest of us don't, what
> problems are there that need solving??  Usually one needs users and a
> population >1 to have problems.  The future always looks rosy
> compared to the present.

Why are the kids migrating towards Facebook/Youtube etc. as their
communications channels?

--
Dean