Thanks for the responses to my IPv8 note.
In case people missed the point, IPv8 addresses
are smaller than IPv6. Here are the sizes.
IPv4 - 32 bits
IPv6 - 128 bits
IPv8 - 43 bits (3+8+32)
There is a natural routing hierarchy with IPv8
addressing....8 regions, 256 distribution centers
in each region and full 32 bit Internets from there.
IPv8 addresses can fit inside the IPv6 address fields.
The problem here, as I see it, is that this _imposes_ a hierarchical
structure onto the physical design on the net, which has not been the
observed pattern of growth.
Cheers,
-- jra
Yes, and when my mother didn't make me go to bed at night, I was
rather cranky, and tardy in my multiplication tables and spelling
exercises the next day.
That we can impose strict hierarchy on address allocations (like
our friends at RIPE, APNIC, and InterNIC have done) is part of the
reason our networking system has assumed a somewhat manageable
growth wrt addressnig.
Big Brother impositions are fine, if the benevolent dictatorship
really is altruistic. (in community space allocation)
-alan
DISCLAIMER -- THIS IS NOT TO IMPLY THAT I SUPPORT, CONDONE, OR
AGREE WITH JIM FLEMING. RATHER I HARBOR FEARS THAT
HE FLIRTS WITH DANGEROUS CHEMICAL SUBSTANCES, AND
PERCEIVES THE WORLD IN A MANNER UNLIKE ANY OF SANE
MIND AND BODY.
Quoting Jay R. Ashworth (jra@scfn.thpl.lib.fl.us):
Yes, and when my mother didn't make me go to bed at night, I was
rather cranky, and tardy in my multiplication tables and spelling
exercises the next day.
You know, Alan... I could'a sworn I wrote something that equates to
"Yeah, Jim, but that's not what peopler _are doing_."
<looks>
Yep, that's what I wrote.
That we can impose strict hierarchy on address allocations (like
our friends at RIPE, APNIC, and InterNIC have done) is part of the
reason our networking system has assumed a somewhat manageable
growth wrt addressnig.Big Brother impositions are fine, if the benevolent dictatorship
really is altruistic. (in community space allocation)
Any particular reason you're dragging in Big Bother, when what I cited
was _the market_?
DISCLAIMER -- THIS IS NOT TO IMPLY THAT I SUPPORT, CONDONE, OR
AGREE WITH JIM FLEMING. RATHER I HARBOR FEARS THAT
HE FLIRTS WITH DANGEROUS CHEMICAL SUBSTANCES, AND
PERCEIVES THE WORLD IN A MANNER UNLIKE ANY OF SANE
MIND AND BODY.
And what are _you_ smoking?
Quoting Jay R. Ashworth (jra@scfn.thpl.lib.fl.us):
> > There is a natural routing hierarchy with IPv8
> > addressing....8 regions, 256 distribution centers
> > in each region and full 32 bit Internets from there.
> > IPv8 addresses can fit inside the IPv6 address fields.
>
> The problem here, as I see it, is that this _imposes_ a hierarchical
> structure onto the physical design on the net, which has not been the
> observed pattern of growth.
^^^^^^^^^^^^^^^^^^^^^^^^^^
Sho nuff.
Cheers,
-- jra
And what I meant was that they should be. And if they don't want
to, they should be forced.
-a
Quoting Jay R. Ashworth (jra@scfn.thpl.lib.fl.us):
[ hierarchical addressing ]
> You know, Alan... I could'a sworn I wrote something that equates to
> "Yeah, Jim, but that's not what peopler _are doing_."
And what I meant was that they should be. And if they don't want
to, they should be forced.
And, as I noted in a private reply: justify that statement.
Cheers,
-- jra
Jay R. Ashworth demanded:
And, as I noted in a private reply: justify that statement.
Physical topology is likely to map to geographic topology.
Circuits certainly do take odd L1 paths to connect L1 endpoints,
but these are exceptions, not he rule.
Accordingly, not allocating in a geographic fashion lends to
deaggregation, which is bad.
Even me, as a proponent of fully meshed architectures, recognizes
that hierarchy is demanded, and will give rise to efficient
network announcments if properly utilized.
-alan
Correction: Not allocating in a topological fashion lends to
deaggregation. Geography often has nothing to do with it.
- paul
Paul Ferguson wrote:
>
> Accordingly, not allocating in a geographic fashion lends to
> deaggregation, which is bad.
>Correction: Not allocating in a topological fashion lends to
deaggregation. Geography often has nothing to do with it.- paul
Paul,
We are talking the *world* here, geography is actually important,
as well as topology.
It is a combination of geography *in conjunction* with topology.
I have a neighbor that is only about 2 miles away, however, a trace
runs to chicago
and back. On the other hand, I have a neighbor, where the next hop is
phoenix.....
It takes a synergistic build , using *both* these metrics.
My two cents...
PS. Keep the change. 
Duh.
And geography is *not* important. I can show you several
organizational networks located in one country, with their
principle Internet connectivity located in another country
entirely. Geographic location means absolutely nothing.
It is where they connect in the global hierarchy.
In the example above, if addresses are allocated solely based
on geographic criteria, I can assure you that you will have
a much, much larger number of prefixes in the global routing
table. Why do you think that the allocating authorities
automagically ask requesters to first ask their upstreams
for address space?
- paul
Paul,
We agree on all but this small point. Certainly one should
allocate ip address space in a topological manner. However, I
know of several large NSP/ISPs that don't because there's no
[technological or economic] punitive incentive for them to do so.
I believe that the correlation between topology and geography will
increase as a function of time.
In fact, I believe that today the correlation is quite high.
The lack of correlation is the exception, in my experience, than
the rule.
-alan
Quoting Paul Ferguson (ferguson@cisco.com):
I'm not sure I agree, but that's not really the point.
I believe the point is that we should not assume that
they are one and the same. Topology (or as they say in
IPv6-speak, aggregator, next-level-aggregator, etc.) is
quite critical in maintaining sufficient levels of
aggregation.
- paul
I have left off a clause that makes my point of view inconsistent.
My comments regarding geographic correlation were with regards to a
specific [example of a] provider's network allocation.
On a provider by provider basis, topology does, and will
increasingly, match geography.
Networks that span large geographic areas should be deaggregated
into regional ares such that aggregation can be efficiently
implemented.
Additionally, within this provider, they should allocate their
chunk of the address space in a geographic manner.
So you have correlation at the high level (providers in an area)
and at a low level (particular provider's IP addressing in a
small area). I did not mean to [though I certainly did] imply
that 2 separate providers would have a strong correlation of IP
address allocation in a locallized geographic region.
Let's take these examples:
KINDISP has a network in 4 continents. They obtain 4
separate sets of network space from the corresponding NICs.
They then develop an allocation plan [ignoring crystal
balls, but utilizing inferences from business plans and
historical trends] and allocate netblocks in a regional
manner.
DUMBISP had a network in 1 continent. They expand into 3
more and continue to use their same original netblock. They
allocate addresses chronologically, instead of
geographically.
This continental allocation is being done today, the NSP's
internal allocation is what I was attempting to focus on.
-alan
Quoting Paul Ferguson (ferguson@cisco.com):
Alan,
Thinking about your response, I am reminded of the
MUST and SHOULD semantics in the context of IETF
documents. 
- paul
Paul,
Duh.
You won't mind if I quote you, here?
And geography is *not* important. I can show you several
organizational networks located in one country, with their
principle Internet connectivity located in another country
entirely. Geographic location means absolutely nothing.It is where they connect in the global hierarchy.
This is true [at the interprovider level]. Today. However, as
deregulation continues [like a freight train with constant
acceleration] what will incentivize "organizational networks" to
maintain these geographically disparate connections?
When Internet demand continues, and more fiber is run, and
competitive LEC and IXCs develop in europe, pacrim, and africa,
the infrastructure will increase, and these distant connections
will become financially prohibitive.*
The market will make topology follow geography. But not for a
while. The next interesting model to study is that of ubiquitous
banwidth. With this model geography will have no bearing on
topology.
So I see three phases of bandwidth growth, with corresponding
effects on the correlation:
Time Bandwidth Correlation
--------- ----------------------- ------------
Today Consolidated Bandwidth Low
Soon Available Bandwidth High
Someday Oversupply of Bandwidth Low
With an oversupply of bandwidth (like we had w/ long distance
providers in the mid to late 80s) backhauling traffic is
cost-effective allowing a noncommodity based market to emerge.
When South African ISPs can interconnect to a continental operating
company/network ina cost effective manner, we can be they'll do
so. And this is the trend I envision towards correlating
topology and geography.
-alan
* the key premise I'm making here is that as the value of a local
connection increases, geography and topology will converge. As
the value of such decreases, they will diverge. The value is a
function of the benefit [gained by local content and useful
aggregation towars the content [like caches, and large networks
building pipes to far-away places, and gaining nice aggregation]]
and the cost.
The cost is going down as competition increases. The benefit is
going up as interesting technologies grow, and providers build
larger international networks.
Make sure that the alternic crowd (when they
get out of jail) controls one of those 8 regions. This
scheme imposes an administrative hierarchy to addressing/networking
which is not conducive to the kind of growth we have seen to date.
Granted, there will be an administrative hierarchy no
matter how you structure addresses, but I would rather that the
consumer decides who is going to administer the tiers
of such a hierarchy instead of leaving that decision
to the protocol fairy.
brad reynolds
ber@cwru.edu
"Faith: not wanting to know what is true."
-- Friedrich Nietzsche