Re: [Tsv-art] [Last-Call] [Idr] Tsvart last call review of draft-ietf-idr-tunnel-encaps-19

Hi, Sue,

FWIW, my view of the text you have is as follows:

- it should be more clear whether you care about path MTU or receiver MTU (EMTU_R); the former ensures “fate sharing” of bytes within a packet along a path, but the latter is actually the limit of what a link can support. E.g., for IPv6, the former is 1280 and the latter 1500; for IPv4, the former is technically 64 and the latter 576.

- tunnels and layers of tunnels do not necessarily reduce the path MTU at the topmost level, e.g., if they correctly support (as should be required) encapsulating header source fragmentation at the ingress and reassembly at the egress
	a tunnel that does not so support source frag and reassembly MUST already support the IPv4 or IPv6 minimums
	it’s easy to see that X over X requires either source frag/reasy at some layer OR no MTU (e.g., Spacewire)
		IPv6 over IPv6, e.g., requires frag/reassy, but sometimes this can be supported at an intermediate layer, e.g., GRE (as noted in 3.6.3)

The doc already discusses tunnel signaling and why that’s not a solution as well, in sec 4.3.1

Joe

> On Nov 10, 2020, at 9:56 AM, Susan Hares <shares@ndzh.com> wrote:
> 
> Joe:
>  
> First of all, I’m sorry to indicate “TSV” instead of “INTAREA”.   Second, I will re-read the draft and determine why my questions were wrong. 
>  
> Thank you for pointing out the errors in my questions. 
>  
> Cheers,  Sue 
>  
> From: Joseph Touch [mailto:touch@strayalpha.com <mailto:touch@strayalpha.com>] 
> Sent: Tuesday, November 10, 2020 12:53 PM
> To: Susan Hares
> Cc: Brian Trammell; idr@ietf.org <mailto:idr@ietf.org>; draft-ietf-idr-tunnel-encaps.all@ietf.org <mailto:draft-ietf-idr-tunnel-encaps.all@ietf.org>; Last Call; tsv-art
> Subject: Re: [Idr] [Tsv-art] [Last-Call] Tsvart last call review of draft-ietf-idr-tunnel-encaps-19
>  
> Hi, Sue,
>  
> It is not and cannot be a requirement. I’ve published numerous RFCs (as have others) that cite expired drafts - even those expired many years before for many reasons - sometimes to leverage their terminology or perspective, others to include their suggestions.
>  
> E.g., RFC 8899, for which 6 draft versions and the RFC were published after the latest version of the tunnels draft expired. 
>  
> Further, a quick look at the RFC Editor queue confirms that this is a ridiculous requirement. Over half of all RFCs published cite versions of drafts that had expired, given the number of weeks delay.
>  
> This is in addition to the fact that the submission queue is laughably locked (as if that prevented any WG from posting informal versions during cutoff),.
>  
> I encourage you to consult the expired version for information on why “MTU” itself is not defined (there are many variants of MTU) and for information about how to deal with layering, just as at least three other published RFCs 
>  
> I cannot speak for the IESG, but I know I have better things to do than support their idiotic invention of requirements.
>  
> Note: it’s an INTAREA draft, not transport. 
>  
> Joe
> 
> 
> On Nov 10, 2020, at 9:14 AM, Susan Hares <shares@ndzh.com <mailto:shares@ndzh.com>> wrote:
>  
> Joe: 
>  
> Could you just resubmit a version of the draft?   
>  
> We cannot reference in the draft-ietf-idr-tunnel-encaps-20.txt draft unless you have a non-expired draft.  John Scudder and the author team added it as a recommendation, but I had them take it out since the draft was expired.  IESG members do not like “expired” drafts as references. 
>  
> Here’s the current -20.txt without your draft reference. 
>  
>                  12.  Operational Considerations               
>                                
>                    A potential operational difficulty arises when tunnels are used, if          
>                   the size of packets entering the tunnel exceeds the maximum               
>                   transmission unit (MTU) the tunnel is capable of supporting.  This         
>                   difficulty can be exacerbated by stacking multiple tunnels, since            
>                   each stacked tunnel header further reduces the supportable MTU.  This            
>                   issue is long-standing and well-known.  The tunnel signaling provided 
>                   in this specification does nothing to address this issue, nor to  
>                   aggravate it (except insofar as it may further increase the         
>                   popularity of tunneling).            
>                
> It would be stronger if we can point to your draft or another TSV draft that explains the details. 
>  
> If you and the TSV-art directorate has changes to this section to deal with MTU, it would very helpful  to receive this information this week. 
>  
> Cheers, Sue 
>  
>  
> From: Idr [mailto:idr-bounces@ietf.org <mailto:idr-bounces@ietf.org>] On Behalf Of Joseph Touch
> Sent: Monday, November 9, 2020 4:52 PM
> To: Susan Hares
> Cc: Brian Trammell; idr@ietf.org <mailto:idr@ietf.org>; draft-ietf-idr-tunnel-encaps.all@ietf.org <mailto:draft-ietf-idr-tunnel-encaps.all@ietf.org>; Last Call; tsv-art
> Subject: Re: [Idr] [Tsv-art] [Last-Call] Tsvart last call review of draft-ietf-idr-tunnel-encaps-19
>  
> Hi, Sue,
>  
> I have a couple of other drafts currently being wrapped up (UDP options, TCP control block sharing bis). The tunnels is next on my list and I hope to finalize a version that we can consider for WGLC by the end of the year.
>  
> That doc (even the latest expired version) has the text we’ve recommended elsewhere, e.g., in the TCP core (793) and elsewhere. 
>  
> Joe
> 
> 
> 
> On Nov 9, 2020, at 1:26 PM, Susan Hares <shares@ndzh.com <mailto:shares@ndzh.com>> wrote:
>  
> Joe and Brian: 
>  
> As the replacement shepherd for draft-ietf-idr-tunnel-encaps-19.txt,  I am looking for the INT area statement on tunnels and MTU in tunnels. 
>  
> Your intarea draft seems to have expired without any replacement. 
>  
> Where is the latest set of comments on tunnels and MTU issue from INT area? 
>  
> Sue 
>  
> From: Joseph Touch [mailto:touch@strayalpha.com <mailto:touch@strayalpha.com>] 
> Sent: Thursday, October 15, 2020 9:12 PM
> To: Brian Trammell
> Cc: tsv-art; Last Call; draft-ietf-idr-tunnel-encaps.all@ietf.org <mailto:draft-ietf-idr-tunnel-encaps.all@ietf.org>; idr@ietf.org <mailto:idr@ietf.org>
> Subject: Re: [Last-Call] Tsvart last call review of draft-ietf-idr-tunnel-encaps-19
>  
>  
> 
> 
> 
> 
> On Sep 28, 2020, at 11:32 PM, Brian Trammell via Datatracker <noreply@ietf.org <mailto:noreply@ietf.org>> wrote:
>  
> First and foremost, I was surprised to find no reference to tunnel or MTU 
> anywhere in the document, especially given the guidance in section 6 to
> stack tunnels. MTU issues are operationally difficulty in single-tunnel
> environments and become more likely to cause problems in multiple-tunnel
> environments. 
>  
> +1
>  
> This is discussed in detail, with some much more specific terminology, in draft-ietf-intarea-tunnels
>  
> In particular, *path MTU* is different from the received MTU, etc. It’s important to get this correct (note the many examples of current standards that do not).
>  
> Joe
> _______________________________________________
> Tsv-art mailing list
> Tsv-art@ietf.org <mailto:Tsv-art@ietf.org>
> https://www.ietf.org/mailman/listinfo/tsv-art <https://www.ietf.org/mailman/listinfo/tsv-art>
>  
> -- 
> last-call mailing list
> last-call@ietf.org <mailto:last-call@ietf.org>
> https://www.ietf.org/mailman/listinfo/last-call <https://www.ietf.org/mailman/listinfo/last-call>