Hi,
On Fri, Nov 01, 2002 at 07:06:21AM +0900, Jun-ichiro itojun Hagino wrote:
> if you get a default route from three sites you are joined to, what
> are you going to do? or what if you get 2001:240::/32 from both sides,
> what are you going to do?
Same as when you have two int
I disagree with this assessment. In the v4 world sites that are,
voluntarily or forcibly, using RFC 1918 address do expect to connect
into the public Internet. They do so because these are the only IP
addresses they have, so what other choice do they have?
This is an effect of the current sh
homas Jefferson]
> -Original Message-
> From: Tony Hain [mailto:alh-ietf@;tndh.net]
> Sent: Thursday, October 31, 2002 3:26 PM
> To: 'Ralph Droms'; [EMAIL PROTECTED]
> Subject: RE: Default site-local behavior for routers
>
>
> Ralph Droms wrote:
> > ..
>
> i still think it necessary to:
> - limit nodes from joining more than (including) 2
> sites at the same
> time.
> - document site-border router's behavior in full
Does anyone on this list object to Itojuns request?
I think it's a good idea and healthy for IPv6.
/
Jefferson]
> -Original Message-
> From: Margaret Wasserman [mailto:mrw@;windriver.com]
> Sent: Thursday, October 31, 2002 9:59 AM
> To: Mark Smith
> Cc: Keith Moore; [EMAIL PROTECTED]
> Subject: Re: Default site-local behavior for routers
>
>
>
>
> >As V
>> On Thu, 31 Oct 2002 09:51:17 -0500,
>> Margaret Wasserman <[EMAIL PROTECTED]> said:
>
>> Are there any commercial routers today that include SBR support?
>
>If I remember correctly, NEC has a product that supports SBR.
>
Yes, NEC's IX1000, IX2000, and IX
> In the v4 world sites that are,
> voluntarily or forcibly, using RFC 1918 address do expect to connect into
> the public Internet. They do so because these are the only IP addresses
> they have, so what other choice do they have?
for that matter, what other choice do they have using the current
I disagree with this assessment. In the v4 world sites that are,
voluntarily or forcibly, using RFC 1918 address do expect to connect into
the public Internet. They do so because these are the only IP addresses
they have, so what other choice do they have?
The multi-address space in the v6 wo
> On Thu, 31 Oct 2002 09:51:17 -0500,
> Margaret Wasserman <[EMAIL PROTECTED]> said:
> Are there any commercial routers today that include SBR support?
If I remember correctly, NEC has a product that supports SBR.
JINMEI, Tatuya
> > Does anyone have an operational network that uses site-local
> > addresses to provide private addressing within a globally
> > connected network? Why did you choose to do this? What were
> > your experiences? Please note that I am interested in
> > deployed, operational networks, not the
>Participate in both, but not route SL prefixes between them. This is
>easy since it can track which interface is appropriate for any given
>use.
you need to have separate routing table for those, or you need to do
other tricks (like KAME's embedded link-local scope identifier). a
> are we looking for a way to
> support applications that span multiple sites that each use site-local
> addresses?
the reality is that if SLs are widely used in v6 networks, apps will
be expected to span sites using SLs, just as they are now expected
to span between the public internet and site
> > Adjacent nets that both use SLs is an interesting (potentially
> > problematic?) architecture - I would be interested in finding
> > out about
> > deployment experience with that case.
>
> This is exactly the case that Keith is concerned about. There is no
> magic here, in this situation the a
Tony,
I don't know about any magic involved; I'm just interested in hearing about
operational experience that would help us understand exactly what's
involved in all of the possible cases.
Also, perhaps it was good that I asked the question, because I'm not sure I
understand part of your respo
At 01:11 PM 10/31/02, Richard Draves wrote:
> Does anyone have an operational network that uses site-local
> addresses to provide private addressing within a globally
> connected network? Why did you choose to do this? What were
> your experiences? Please note that I am interested in
> deployed,
Not true. There is no expectation that the _same_ IP address
will point to two _different_ systems because it occurs on
different sides of an IGP/EGP transition.
Margaret
At 12:05 PM 10/31/02, Tony Hain wrote:
Margaret Wasserman wrote:
> ...
> Are there any commercial routers today that inclu
Ralph Droms wrote:
> ...
> Adjacent nets that both use SLs is an interesting (potentially
> problematic?) architecture - I would be interested in finding
> out about
> deployment experience with that case.
This is exactly the case that Keith is concerned about. There is no
magic here, in this s
itojun wrote:
> ...
> depending on your definition of site border, Microsoft
> router can
> participate both Microsoft site as well as upstream-ISP site.
>
> see Miyakawa-san's DSL service plans - CPE participates
> to both ISP
> site as well as customer site.
Participat
> My two cents about two-faced DNS: In the v4 setups I have done, a
> one-faced DNS is enough if the DNS server is inside the NAT box, because
> the router that does NAT (at least the ones I have been using, Cisco)
> will decapsulate the DNS reply and replace the IP address with the
> public one.
> > curious: do you run any router which participates to
> > multiple sites?
>
> Why would they? It is an address space that they can use for internal
> purposes. If they wanted to communicate to an external entity, they
> would either have to coordinate use of the SL space, or simply use
>
> > > Yes it uses two-faced DNS. It works fine.
> >
> > not everyone uses Microsoft software, you know.
>
> That was not a sales pitch from Rich, so don't turn it into one.
sorry, not how I read it at all - I read it as saying "it works for us".
but since Microsoft mostly runs Microsoft softwar
>> curious: do you run any router which participates to
>> multiple sites?
>Why would they? It is an address space that they can use for internal
>purposes. If they wanted to communicate to an external entity, they
>would either have to coordinate use of the SL space, or simply use
>globals.
Perhaps Microsoft has a requirement for multiple, independent address
spaces (there's nothing that requires Microsoft == one-site)? Or the
Microsoft net is in some way adjacent to another network using SLs?
Adjacent nets that both use SLs is an interesting (potentially
problematic?) architectu
Keith Moore wrote:
> > Yes it uses two-faced DNS. It works fine.
>
> not everyone uses Microsoft software, you know.
That was not a sales pitch from Rich, so don't turn it into one. The
question Margaret asked was if anyone had an example of running code. A
yes answer tends to deflate the argumen
itojun wrote:
> curious: do you run any router which participates to
> multiple sites?
Why would they? It is an address space that they can use for internal
purposes. If they wanted to communicate to an external entity, they
would either have to coordinate use of the SL space, or simply use
My two cents about two-faced DNS: In the v4 setups I have done, a
one-faced DNS is enough if the DNS server is inside the NAT box, because
the router that does NAT (at least the ones I have been using, Cisco)
will decapsulate the DNS reply and replace the IP address with the
public one. In a rather
> Yes that appears to create a problem for multi-party apps, but the
> problem of disconnectedness exists without a defined SL. Since SL makes
> it clear that there are places where the network will be disconnected,
> there should be a note to application developers stating what the
> pitfalls are,
> Yes it uses two-faced DNS. It works fine.
not everyone uses Microsoft software, you know.
IETF IPng Working Group Mailing List
IPng Home Page: http://playground.sun.com/ipng
FTP archive:
>> Does anyone have an operational network that uses site-local
>> addresses to provide private addressing within a globally
>> connected network? Why did you choose to do this? What were
>> your experiences? Please note that I am interested in
>> deployed, operational networks, not theoretic
> curious: do you run any router which participates to
> multiple sites?
Not to my knowledge, but I don't run the routers.
Rich
IETF IPng Working Group Mailing List
IPng Home Page: http://playground.
> Does anyone have an operational network that uses site-local
> addresses to provide private addressing within a globally
> connected network? Why did you choose to do this? What were
> your experiences? Please note that I am interested in
> deployed, operational networks, not theoretical de
2 7:41 AM
> To: [EMAIL PROTECTED]
> Subject: Re: Default site-local behavior for routers
>
>
> Margaret Wasserman wrote:
>
> >> What doesn't really exist is the filtering of prefixes
> being put into
> >> route exchange messages based on an arbitrary
Margaret Wasserman wrote:
> ...
> Are there any commercial routers today that include SBR support?
By definition, every IGP/EGP transition is at least one example of site
border, so the answer to your question is yes.
IETF IP
Margaret Wasserman wrote:
You're probably right.
On the other hand, as per Ole Troan's earlier email (which I agree
with), I don't think all router implementations should be required to
support multi-sites.
I think Ole's comments apply to specialized routers. If you are
marketing a general
Margaret Wasserman wrote:
What doesn't really exist is the filtering of prefixes being put
into route exchange messages based on an arbitrary index (zone
id).
The other big issue is how the routing table(s) are built and
managed. That can be a big hit on memory/storage space.
Brian, could yo
You're probably right.
On the other hand, as per Ole Troan's earlier email (which I agree
with), I don't think all router implementations should be required to
support multi-sites.
I think Ole's comments apply to specialized routers. If you are
marketing a general purpose router, you almost h
As Vint Cerf wrote in a RFC recently, The Internet is for Everyone. Once
everyone has it (I'd say one of the fundamental inherent goals of IPv6),
hopefully the world can become a more tolerant place through
communication, allowing better understanding of different peoples view
points and beliefs
> Does this make me a terrorist network administrator, for trying to help
> by showing how I might try to use one of the features of IPv6 in the
> real world ?
No, of course not. It's just that recent events have provided such glaring
examples of the utter stupidity of arguments of the form
"we
Does this make me a terrorist network administrator, for trying to help
by showing how I might try to use one of the features of IPv6 in the
real world ?
Please do not bring up terrorism on this mailing list, not only is it
in-appropriate, it is in particularly bad taste after the recent
bombings
Hi Brian,
> I'm not sure that it does much, though, to address the issues
> that site-locals raise for transport protocols, applications,
> DNS and management protocols. Am I missing something?
Well, this is the ipv6 working group's mailing list after all. We've
been admonished before for mes
Mark Smith wrote:
On Thu, 2002-10-31 at 16:29, Keith Moore wrote:
however I'd be really surprised if SL filtering added to the
cost of a router.
You're probably right.
On the other hand, as per Ole Troan's earlier email (which I agree
with), I don't think all router implementations should b
> Enough managers of real networks created them, and still demand them
> that despite your claim that there is no need, there is a requirement
> that we provide something.
that's like saying that we have to do _something_ about bin Laden,
so we might as well bomb a few thousand people who have not
On Thu, 2002-10-31 at 16:29, Keith Moore wrote:
>
> however I'd be really surprised if SL filtering added to the
> cost of a router.
>
You're probably right.
On the other hand, as per Ole Troan's earlier email (which I agree
with), I don't think all router implementations should be required to
Keith Moore wrote:
>
> there's no need for such addresses.
Enough managers of real networks created them, and still demand them
that despite your claim that there is no need, there is a requirement
that we provide something.
Tony
if multi-site routers really did cost more than single-site routers,
that's even more reason to not use SLs - since the same effect could
be achieved at less cost using globals and prefix-based filtering.
however I'd be really surprised if SL filtering added to the
cost of a router.
Keith
-
I think there is in Australia ...
Have a read of my previous emails.
If I was to build a very simple enterprise network between 8 capital
cities, with an single ethernet segment in each, and 7 wan links
connecting them, if I follow the current site-local definition
(geographical boundaries defin
> Oops, sorry, I think I overloaded an already defined term.
>
> Maybe "enterprise local addressing" or something similar that doesn't
> imply a geographical size or location, and indicates the addressing
> uniqueness is only local to the organisation using it.
there's no need for such addresses.
Oops, sorry, I think I overloaded an already defined term.
Maybe "enterprise local addressing" or something similar that doesn't
imply a geographical size or location, and indicates the addressing
uniqueness is only local to the organisation using it.
On Thu, 2002-10-31 at 15:04, Keith Moore wr
> Would "provider independent local addressing" be a better name for site
> local addressing if Tony's model is the most commonly followed ?
you don't want PI addresses to be constrained to be "local". you want
to be able to privately route them between sites.
Keith
-
> Margaret Wasserman writes:
>
>>>In any case, the site boundary should never be larger
>>>than the IGP scope, so if we are going to talk about
>>>defaults, rather than assuming every interface is in a
>>>different site, why not assume every EGP/IGP boundary
>>>identifies a different site? If we c
Would "provider independent local addressing" be a better name for site
local addressing if Tony's model is the most commonly followed ?
I would find that a more descriptive name, as it doesn't suggest that I
have to artificially place a boundary on the addressing due to physical
geography.
Mark.
In any case, the site boundary should never be larger than the IGP
scope, so if we are going to talk about defaults, rather than assuming
every interface is in a different site, why not assume every EGP/IGP
boundary identifies a different site? If we can get past that, maybe we
can start talking
Tony,
That is a reasonable approach and one that I could live
with. It allows SLs to exist and control is based on tools
that are in wide use today.
Brian
Tony Hain wrote:
The whole discussion about lack of definition of site boundary is bogus,
and causing a large waste of energy. We don't
For the record, my opinion follows Ole's comments.
Brian
Rob Austein wrote:
What Ole said.
IETF IPng Working Group Mailing List
IPng Home Page: http://playground.sun.com/ipng
FTP archive:
What Ole said.
IETF IPng Working Group Mailing List
IPng Home Page: http://playground.sun.com/ipng
FTP archive: ftp://playground.sun.com/pub/ipng
Direct all administrative requests to [EMA
> So, one of the items that Margaret suggested was some text in
> the node requirements doc or the scoped addr arch that states
> that nodes default to being in one site.
>
> However, there has been some mention that people would prefer
> different behavior in routers. That is, the stated desire
>
Margaret Wasserman wrote:
> ...
> I've had an action item for a while to summarize the thread
> that led to this conclusion to the IPv6 list, but I haven't
> gotten to it yet. I'll do so soon.
I appreciate that work loads make a summary effort challenging, but in
this particular case it would se
An organization should probably start with the assumption that a site
boundary is exactly congruent with an OSPF area, but they may choose to
restrict it further, or expand it when it makes sense for their network.
In any case, the site boundary should never be larger than the IGP
scope, so if w
The whole discussion about lack of definition of site boundary is bogus,
and causing a large waste of energy. We don't tell people how to bound
areas in OSPF, yet we are expected to spell out the universal definition
of a site. To a first order, the concepts are exactly the same, how much
informati
> What are others' opinions on this issue?
I'm actually thinking that the most desirable default behavior
for routers is one that discourages use of SLs unless they're
explicitly configured. So I am inclined to believe that a
router (or a multi-interface host when acting as a router) should
by
> This suggestion leads to the model where hosts with multiple
> interfaces will assume that all its interfaces are in the
> same site (e.g. have the same site-local zone id) unless
> explicitly configured to have multiple sites. While routers
> will default to having a unique site-local zone id f
61 matches
Mail list logo
