Re: [RRG] What does incremental deployment mean

[Dino Farinacci <dino@cisco.com>]

> Hi Dino,
>
> > > According to this criteria, many proposals, including v6 EID over v4
> > > RLOC of
> > > LISP have been sentenced to death.
> >
> > No, not true. We assume and it is a safe assumption that hosts  
> > already
> > have IPv6 stacks. And IPv6 EIDs with IPv4 RLOCs means your LISP  
> > router
> > has to do it.
>
> As Noel said, "there's a lot of old stuff out there, not just  
> Windows."
> Does that statement mean there are still a small part of hosts  
> without IPv6
> capability?
>
> If your assumption is correct, then what's the main obstacle for the
> transition from v4 to v6 since IPv6 is already enabled on most hosts?

What I meant was that you wouldn't any more changes to IPv6 stacks to  
support LISP at the site.

Noel's point stands strong and it would be nice for those systems that  
will never get upgraded that they can still use the Internet.  ;-)   
Err, nice.... I mean necessary.

> > That is life if I have ever seen it.
> >
> > > > > End of story.
> > > > >
> > > > > (For one, earler versions of Windows don't use/support Windows
> > > > > Update, and a
> > > > > lot of people have it turned off anyway, from paranoia/prudence/
> > > > > whatever. But
> > > > > just in general, there's a lot of old stuff out there, not just
> > > > > Windows.)
> > > >
> > > > Even if everyone had Windows Update enabled, it would take too long
> > > > to
> > > > get every system upgraded.
> > >
> > > Most of today's core routers have already been able to support more
> > > than 1
> > > million routing entries, is there enough incentive for the  
> > > carriers to
> > > deploy map&encap scheme within a short period.
> >
> > Carriers don't deploy it. Sites do because they want easier ways to  
> > do
> > multihoming.
>
> Provided the above was the main drive for the map&encap scheme,  
> though the
> site with multiple ETRs can enjoy the multihoming benefits, what's the
> incentive for the other site to deploy the ITR?

For the same incentive. Any site who wants to receive packets on  
multiple ISP links would have incentive.

I think we can all agree and a safe assumption that every site in the  
world would want bidirectional connectivity?  ;-)

> > > The approach that map implemented by hosts and encapsulation
> > > implemented by
> > > ITRs is incremental deployable. If hosts have not been changed to
> > > support
> > > this capability, the ITR can implement the map and encapsulation
> > > together.
> > > The upgraded hosts will not suffer the initial packet loss/latency
> > > pain. And
> > > the ITR with upgraded site network doesn't need a cache.
> >
> > What is the point of having them both do it?
>
> This is a transition strategy just like the dual stacks. Of course,  
> if there
> was a solution which can overcome the side-effects of the cache  
> mechanism
> and the long stretch forwarding path, it's not necessary to upgrade  
> the
> hosts to support this mapping querying function.

You think the Internet can take more than one transition? I haven't  
seen any yet in 25 years. And NCP to TCP happened because a few  
researchers just pushed it through.

I bet the number of people using NCP were less than the number of  
people on this mailing list.  ;-)

Dino






--
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