Dan, Margaret, et al: I have given some thought to the matters raised below. This note will address the questions raised by Dan in his first paragraph; the second paragraph will see a later response. Please be aware that the below should be considered, at its most ambitious, not more than a starting point for group discussion; it is not to be considered a formal proposal.
AEB On Fri, 14 Feb 2003, Dan Lanciani wrote: > "Alan E. Beard" <[EMAIL PROTECTED]> wrote: > > |Yes, I think you are probably right in speculating that any technically > |sound mechanism which preserves the virtues of provider independence, > |portability, and stability in configuration of end-user-network devices > |will satisfy the functional requirements of the end-user community. It > |seems to be up to us to architect the mechanism(s). > > Any thoughts on how we might jump start some activity in this area? Over > the years I've tried the bottom-up approach by suggesting some possible > mechanisms, and I've tried the top-down approach by suggesting that we try > to reach consensus on how much overhead we are willing to accept in return > for the solution. The former generally provokes protests that the solution > is too complicated to sketch and the latter usually results in silence. How > can we get some serious discussion going? > [...] > > Dan Lanciani > ddl@danlan.*com > It seems to me that we have in recent weeks seen a shouting-match form of the top-down, how-much-penalty-are-we-willing-to-pay discussion, with one camp asserting loudly that the PA-with-aggregation scheme is strictly necessary for the operation of IPv6, and another group insisting vigorously that any address allocation scheme other than unrestricted PI will be inevitably and utterly fatal to the continued deployment of IPv6. This is roughly equivalent to: do we smother the baby now, or poison it when it reaches puberty? During this debate, most people with any reasonable sense of self-preservation (which immmediately excludes me) were cowering in gator holes, preferring the possibility of a confrontation with an enraged she-gator to the near certainty of an encounter with all the lead flying around out in the sawgrass. In both cases, the magnitude of the postulated adverse effect was deemed to be well beyond the pale, and the discussion subsided to a series of exasperated splutters. There are a number of paths we might explore to encourage some productive discusions. I will suggest below the outline of a sequence which seems to me attractive. First, we may wish to start a top-down discussion with the objective of assembling a list of functional objectives for the address allocation schemes which have been under discussion here. Please note that we have a possible charter problem here - more on that below. Once we have got rough concensus on the functional objectives, we can proceed to a bottom-up approach on mechanisms, and from there to specific proposals for standards, or BCP, or guidance to the registries. Now, we have a scope problem here, which is a result of the wording of the WG charter, and is further complicated by the instructions which came of out the Atlanta meeting. We probably need to consult Margaret before she concludes that we have taken the bit in our teeth (which we have) and are about to put at risk every horse and all the coaches in the staging yard (still undecided). At IETF 55 (Atlanta), floor discussions in the WG meeting indicated that a proposal to limit use of site-local address allocations was contingent, at least in the minds of many of the attendees, on making provision for some form of PI address space which is available to most (preferably _all_) networks. As this last issue did not fall within scope of any of the WG's defined work items, a decision was taken (how's that for passive voice?) to refer the matter to another area (SubIP or OAM, if my memory serves correctly) for consideration. What started in this WG as a discussion peripheral to the issue of restricting use of SL has now devolved into a reconsideration of the fundamental principles and assumptions underlying the address allocation practices for IPv6. This latter discussion is clearly beyond the scope of the current work items for this working group. However, the fundamental issues seem unwilling to fold their tents and steal away quietly into the night. As I read the WG charter, these matters are incontestably within the bounds of the charter, although not subsumed by any extant work item. IMHO, our discussions so far have identified a matter within the scope of the charter which urgently requires a resolution and requires further work toward that end. So far, we have, by dint of quite extraordinary self-restraint (yeah, right), avoided rogue status by confining our activities to disussion of the nature and scope of the perceived problems, rather than commencing formal work upon definition and solution of the problems. However, absent explicit assent from the ADs, formal work of sort suggested above would leave Margaret fully justified in coming 'round with her elephant gun, placing the muzzle squarely between our ears, and discharging as many rounds as she sees fit. (Margaret, I know that you don't carry an elephant gun: the above description is entirely metaphorical, and no imputation of violence attaches to you.) Unlike many judicial bodies, the IETF does not have a compelling tradition of stare decisis: we are not prohibited from revising previous recommendations when changed conditions or new information indicate that the recommendations may no longer comport with reality as we know it. I would suggest that the continuing discussions and controversy over the issue of PI vs PA strongly indicate that additional consideration of IPv6 address allocation practice is urgently needed, if for no other reason than to assure the IETF and the user community that we are recommending technically and operationally feasible models for address allocations. (I will further suggest that the current practices leave that conclusion somewhat in doubt.) It seems to me not unreasonable to ask of Margaret, Rob, Thomas and Erik permission to augment our list of work items with a specific task to review current guidance to the address registries concerning address allocation philosophy and procedure with a view to encouraging user acceptance of IPv6 while maintaining technically sound policies concerning routing overhead. (Yes, that language is ambiguous and imprecise to the point of being downright sloppy - we will need better wording if this is to proceed.) The multi6 WG has a some papers extant which are related to the address allocation practices, one of which is Michel Py's gapi-00 draft. However, there does not seem to be in progress anywhere a comprehensive reconsideration of address allocation practices as related to PA, PI, and routing table burden, nor any current process which explicitly considers end-user expectations concerning IPv6 address allocations and uses permitted thereof. There are, in my view, two extant items in the IPv6 WG charter which subsume the work outlined above, to wit: - Provide technical input to the IAB, IANA and Internet Address Registries with regard to IPv6 address allocation policies and procedures. This covers the full discussion and the specific work outlined above. and - Serve as a review board and body of competence and coordination for IPv6 architectural issues that span multiple IETF working groups. The specific issues regarding PA, PI, and aggregation will likely afflict IDR and OSPF; further, Multi6 has open work on related issues. In summary, I think that we may have sufficient justification to ask the WG chairs and the ADs to add a specific task to the list of open WG work items: that being a review of current address allocation philosophy and procedures in light of current state of routing protocol specifications and in light of end-user expectations concerning allocation and permitted use of IPv6 addressing. What think ye all? AEB -- Alan E. Beard <[EMAIL PROTECTED]> AEBeard Consulting; 4109 Chelsa Ln; Lakeland FL 33809 863.815.2529 -------------------------------------------------------------------- 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] --------------------------------------------------------------------
