>-----Original Message-----
>From: [EMAIL PROTECTED]
>[mailto:[EMAIL PROTECTED]]On Behalf Of Tomohide Nagashima
>Sent: Wednesday, 28 March 2001 9:52
>To: [EMAIL PROTECTED]; [EMAIL PROTECTED]; [EMAIL PROTECTED];
>[EMAIL PROTECTED]
>Cc: [EMAIL PROTECTED]
>Subject: Re: what is a site?
>
>
>Hi,
>
>At least, it is clear that "Site" is a set of links. But is
>the subset of Site
>the Site ?
This means that a site can be a part of another site....
At least, if I get this right...
Why would we make a site a part of a site?
You get something like this:
Company-wide Network
|- Sites
|- Sites
| |- Links
| |- Sites
| |- ......
|- Links
You might benefit from this up to certain extend, but I tends to get a mess
of sites and 'sub'-sites. So what's what at the end? Or for site-local
addressing, who's local to me? You might get strange siturations:
A is in the same site as B and B is in the same site as C, but A isn't (form
his/her point of view) not in the same site as C. (However, C things A is in
the same site).
>I begin to believe this answer is YES.If else, what
>do we call that ?
>I belive this will be a key of definition of "Site".
So a site can be hierarchical and can a site also be a part of two (totally)
different sites?
Wouldn't this get quite messy...
>
>I propose definition of "Site" as follows;
>I also propose new word "Full-Site".
>
> (Definition of Site)
> A set of link "S" is a Site <==>
> "N" is a integer set N = { n | n is integer , 0=< n =< 2^16 }.
~~~~
Why an upper limit?
... N = {n | n is integer, 0=<n }
> For the "M" which is subset of N;
> There is a one-to-one projection from S to M.
I got lost here...
>
> (Definition of Full-Site)
> Especially, We call that a Site with M=N is a Full-Site.
>
>I think Full-Site has no reality , but this word is very useful.
>Then we will have these theorem,
>
> (Theorem)
> 1. |M| = |S| i.e. the number of M members is equal to the
>number of S members.
> 2. Any subset of Site is Site.
> 3. Full-Site is Site
>
> These are obvious.
>
>Here is a example.
>
> (Example) There are three links in my network like this;
> Link 1 Link 2
> -----+---- --+-------
> | |
> --+---------+---
> Link 3
>
> We select projection as follows,
> Projection = {
> return 0x000a for Link 1
> return 0x000b for Link 2
> return 0x000c for Link 3
> }
>
> This is off-cause one-to-one projection from "S" to "M".
> M = { 0x000a, 0x000b, 0x000c } is subset of N
>
> So this "my network" is Site.
>
> If we select subnet ID as follows,
> subnet ID = 0x000a for Link 1
> subnet ID = 0x000b for Link 2
> subnet ID = 0x000c for Link 3
>
> then we can select Site-Local Address and Global
>Address as follows,
> Site-Local Global
> Link 1 fec0:0:0:a::/64 BLOB:A:L:a::/64
> Link 2 fec0:0:0:b::/64 BLOB:A:L:b::/64
> Link 3 fec0:0:0:c::/64 BLOB:A:L:c::/64
> (BLOB:A:L::/48 is a prefix from upstream for
>this site.)
As long as Site-Local addresses are routable over Link 1, 2 and 3, this
would be OK for me...
>
> If we add Link 4 with subnet ID = 0x000d , this new
>"my network"
> is also Site. We can add link until Site will became
>Full-Site.
>
> (Discussion)
>
> Let's think about following topology.
> [TLA1] [TLA2]
> | |
> <Network1>---<Network2>
>
> 1) If administrator of Network1 and 2 regards that both
>Network1 and 2 should
> be in same Full-Site,
> ~~~~~~~~~~~~~~
> Projection for numbering of Site-Local Address and
>Aggregatable Global
> Address for Network1 and Network2 are same projection
>ID(link_name).
>
> Link "s1" in Network1 will be allocated
> fec0:0:0:ID(s1)::/64
> T:L:A1:ID(s1)::/64
> Link "s2" in Network2 will be allocated
> fec0:0:0:ID(s2)::/64
> T:L:A2:ID(s2)::/64
>
> Note that
> It is possible that "s2" will be allocated T:L:A1:ID(s2)::/64.
> It is possible that "s1" and "s2" connect with Site-Local address.
>
> 2) If administrator of Network1 and 2 regards that Network1
>and 2 should be
> in different Full-Sites,
> ~~~~~~~~~~~~~~~~~~~~
> Projection for numbering of Site-Local Address and
>Aggregatable Global
> Address for Network1 and Network2 are differnt. Let's call
> the projection which is used in Full-Site of Network1 is
>ID1(link_name)
> the projection which is used in Full-Site of Network2 is
>ID2(link_name)
>
> Link "s1" in Network1 will be allocated
> fec0:0:0:ID1(s1)::/64
> T:L:A1:ID1(s1)::/64
> Link "s2" in Network2 will be allocated
> fec0:0:0:ID2(s2)::/64
> T:L:A2:ID2(s2)::/64
>
> Note that
> It is impossible that "s2" will be allocated T:L:A1:ID1(s2)::/64.
> It is impossible that "s1" and "s2" connect with
>Site-Local address.
>
I would fully agree....
> Which way administrator will select is depends on policy.
>
> Another way of define "Site" is that we call that only
>"Full-Site" in
> previous definition is "Site", and "Site" in previous is like ,,,
> "Sub-Site". But "Full-Site" in prev is too ideal to use it as usual,
> I believe we would define that as "Full-Site".
>
>Give me your comment?
>
I don't see why we should use 'sub-sites'. From my point, this would be
quite a stupid move to use sub-sites, it can cause networks to get quite
messy. So actually, for me, the definition site, would be enough...
Link, in this case, would be a part that requires only a TTL value of 1.
This link-local is possible by sending a message with a TTL of 1.
>----
>Tomohide Nagashima
>[EMAIL PROTECTED]
>--------------------------------------------------------------------
>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 [EMAIL PROTECTED]
>--------------------------------------------------------------------
>
- Joris
--------------------------------------------------------------------
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 [EMAIL PROTECTED]
--------------------------------------------------------------------
