[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

RE: [RRG] GSE?



> The ESD would be a constant when changing ISPs.  That's the whole point.
> Identifiers are decoupled from locators.

But that is not all. Forming a socket requires a port number. Basically a "complete" application to application "address" looks like ASN+IP Address+Port. How about devices with no ports, e.g. smart dust?
What I am trying to say is that the "true" ESD needs to name / identify an ultimate source / destination of bits.

Thanks,

Peter


--- On Wed, 6/11/08, Tony Li <tony.li@tony.li> wrote:

> From: Tony Li <tony.li@tony.li>
> Subject: RE: [RRG] GSE?
> To: "'Mayutan A.'" <mayutan.arumaithurai@gmail.com>, "'Robin Whittle'" <rw@firstpr.com.au>
> Cc: "'Routing Research Group'" <rrg@psg.com>
> Date: Wednesday, June 11, 2008, 1:34 PM
> Hi Mayutan, Robin,
> 
> 	Isn't the Six-One proposal by Christian Vogt an
> enhancement of the
> GSE. 
> 	http://www.watersprings.org/pub/id/draft-vogt-rrg-six-one-01.txt
> 	
> 	Correct me if I am wrong.
> 
> You are exactly correct.  I still encourage folks to read
> GSE independently,
> just so you have some perspective on Christian's
> changes.
> 
> Also, some of the work that Ran Atkinson has done has been
> in part derived
> from GSE.
> 
> 		
> 		GSE seems to have been developed briefly around 1997.  I
> understand
> 		that applying it to IPv6 as used today would involve
> major
> changes
> 		in routers, host stacks and some or all applications.
> 		
> 		There may well be some major attractions in doing this,
> if
> it could
> 		be done, but it sounds like a radical thing on which to
> bet
> the
> 		future of the Net.
> 
> 
> Welcome to the IRTF.  Our job is research.  No job too
> large, no change
> unthinkable.
> 
> 		
> 		Could you or someone else put together a proposal and
> link
> to it
> 		from the RRG wiki?  An 8 page summary and analysis
> document
> would be
> 		good too.
> 
> 
> Others should feel free to step up here.  I'm trying to
> remain neutral.
> 
> 		
> 		A crucial part of this would be the time-frame for
> transitioning the
> 		current IPv6 system to whatever it is you are planning,
> and
> then
> 		having a transition plan for most end-users from IPv4 to
> the
> new system.
> 		
> 		I think it would also be good to explain why you would
> prefer to do
> 		this in a hurry for IPv6 - due to whatever urgency you or
> other
> 		people might think about the IPv4 scaling problem -
> rather
> than
> 		fixing the IPv4 problem with a map-encap scheme and then
> being able
> 		to take more time on whatever it is you propose for IPv6.
> 
> 
> I'm not in a hurry to do anything.  There's no
> need.  I'd much rather Get It
> Right.  Whatever we do here is forever.
> 
> 		
> 		I haven't read enough to know how it provides
> multihoming
> and
> 		portability (of the ESD part of the address) when
> changing
> ISPs.
> 
> 
> The ESD would be a constant when changing ISPs.  That's
> the whole point.
> Identifiers are decoupled from locators.
> 
> 		
> 		So how does the Routing Goop and STP get set when the
> packet
> leaves
> 		the site for another?  Does a router change them or does
> the
> sending
> 		host have to get it right.  Does there need to be a
> mapping
> function
> 		and consequently a mapping database to determine what to
> set
> these
> 		to, since the ESD is what uniquely identifies the
> destination host?
> 
> 
> Presumably set by a router when you exit the subnet and/or
> the site.  Yes,
> there needs to be a mapping database to determine
> destination RG and ESD.
> One might reasonably extend DNS to do this.  No mapping
> database is needed
> in the site's local routers as they would presumably be
> configured with the
> RG or learn it via some other management mechanism such as
> SNMP, DHCP, the
> IGP, or your favorite NMS.
> 
> 		
> 		What lead to the demise of GSE ten years ago?
> 
> 
> I wasn't directly involved, but my read was that it was
> politics.  Because
> it modified v6, it was unacceptable to those that felt that
> v6 was perfect.
> We seem to be over that now...
> 
> Regards,
> Tony
> 
> 
> --
> to unsubscribe send a message to rrg-request@psg.com with
> the
> word 'unsubscribe' in a single line as the message
> text body.
> archive: <http://psg.com/lists/rrg/> &
> ftp://psg.com/pub/lists/rrg


      

--
to unsubscribe send a message to rrg-request@psg.com with the
word 'unsubscribe' in a single line as the message text body.
archive: <http://psg.com/lists/rrg/> & ftp://psg.com/pub/lists/rrg