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

Re: [Ecrit] Forest guides



I'll change the subject line to reflect the more substantive issue you are raising.

I'm trying to walk a fine line in providing an architectural description that illustrates what's possible, without trying to nail down various practical "governance" approaches. In practice, I think a few options are possible:

- A real root, run by somebody. This is plausible for commercial directories for non-emergency services, possibly for a more limited geographic region. Effectively, a single FG or a single tree, depending on how you look at it.

- A "cooperative club": A group of organizations that propagate each other's announcements, but have some mechanism for admitting new members.

- A "semi-cooperative club": Everybody propagates most records, maybe with exceptions.

- Full peering: Each FG peers with all (say) countries it wants to work with. Thus, the "official" US FG would not talk to Iran and Cuba, say, except maybe through Switzerland.

In each case, records may be signed or, in a more tightly controlled environment, not.

There are probably more variations. I'm not sure I want to detail them all, since I'm reluctant to rule out "business models". Clearly, the text needs to be consistent, so I'll work on that.



Henning


On Dec 18, 2006, at 11:06 AM, Otmar Lendl wrote:


Regarding Forrest Guides:

We're running into a contradiction here, e.g.:

| For scalability and reliability, there will need to be a
| large number of forest guides, all providing the same information. A
| seeker can contact any forest guide and will then be directed to the
| right tree or, rarely, set of trees.


vs.

| Trees can also restrict
| their cooperation to parts of the information. For example, if
| country C does not recognize country T, C can propagate tree regions
| for all but T.


Either the FGs offer all the same information or they don't.

As I see the problem, we have basically two choices:

One choice is to keep the forrest guides in sync with each other, with
an automated protocol between them. This is the first statement from
above plus these three sentences:


| A tree node at the top of a tree can contact any forest
| guide and inject new coverage region information into the system.
| One would expect that each tree announces its coverage to more than
| one forest guide. Each forest guide peers with one or more other
| guides and distributes new coverage region announcements to all other
| guides.


In that case, this set of forrest guides act just as another layer in
the hierarchical resolution model. Thus we don't actually have a forrest,
but a single tree with the cluster of forrest guides as the root.


This, of course, reopens the pandora's box of who controls what to put
in there, which was the reason for the introduction of the forrest in
the first place.

Reading between the lines of the security section: Do you propose that
the forrest guides just sync a set of digitally signed tree referrals
amongst each other and then each FG decides based on local configuration
which ones he considers when answering questions? And who is going to
decide who can participate in this FG coverage exchange system?


The other option is to leave the question of forrest guide synchronization
completely out of scope and thus don't even try to solve a political problem
with technical means.


I have this nagging feeling what we want to have our cake and eat it, too.

/ol
--
< Otmar Lendl (lendl at nic.at) | nic.at Systems Engineer >

_______________________________________________
Ecrit mailing list
Ecrit at ietf.org
https://www1.ietf.org/mailman/listinfo/ecrit


_______________________________________________
Ecrit mailing list
Ecrit at ietf.org
https://www1.ietf.org/mailman/listinfo/ecrit