Re: [mpls] [EXTERNAL] [Technical Errata Reported] RFC3031 (6450)

> On 4 Mar 2021, at 14:12, Duane Anderson <duanela@edgewater.ca> wrote:
> 
> Stewart -
> 
> I work in the military/aerospace sector. I know for a fact that MPLS likely will be implemented into new component designs by any number of manufacturers of safety-critical systems (e.g., aircraft, spacecraft, automobiles, trains, etc.). This will go on for years to come. 

That is interesting. I had feared the MPLS was getting to EoL. It is good to see it applied to new applications.

> 
> I have no doubt that there is little perceived implementation difficulty. However, given the magnitude of the errata in RFC3031, I think there's a existence proof that statistically someday, sometime in the future, people will be hurt and likely will die. 
> 

An existence proof would require that someone had already been hurt not a statistical possibility that it might happen at some time in the future.

In any case MPLS is so simple in the data plane that any competent tester should be able to force a new implantation to experience all the corner cases. 

What is more likely to happen is an MPLS application error, or a hardware error, than a misunderstanding of the MPLS data plane itself.

- Stewart

> - Duane
> 
> -----Original Message-----
> From: Stewart Bryant <stewart.bryant@gmail.com> 
> Sent: March 4, 2021 7:16 AM
> To: RFC Errata System <rfc-editor@rfc-editor.org>
> Cc: Stewart Bryant <stewart.bryant@gmail.com>; erosen@cisco.com; arun@force10networks.com; rcallon@juniper.net; DEBORAH A BRUNGARD <db3546@att.com>; Alvaro Retana <aretana.ietf@gmail.com>; Martin Vigoureux <martin.vigoureux@nokia.com>; loa Andersson <loa@pi.nu>; n.leymann@telekom.de; Tarek Saad <tsaad.net@gmail.com>; Duane Anderson <duanela@edgewater.ca>; mpls@ietf.org
> Subject: [EXTERNAL] Re: [mpls] [Technical Errata Reported] RFC3031 (6450)
> 
> I think that there is an existence proof that this is not causing implementation difficulty.
> 
> - Stewart
> 
> 
>> On 4 Mar 2021, at 03:43, RFC Errata System <rfc-editor@rfc-editor.org> wrote:
>> 
>> The following errata report has been submitted for RFC3031, 
>> "Multiprotocol Label Switching Architecture".
>> 
>> --------------------------------------
>> You may review the report below and at:
>> https://www.rfc-editor.org/errata/eid6450
>> 
>> --------------------------------------
>> Type: Technical
>> Reported by: Duane L. Anderson <Duane.Anderson@Edgewater.CA>
>> 
>> Section: GLOBAL
>> 
>> Original Text
>> -------------
>> 2.2. Terminology defines the terms
>> 
>>       Layer 2         layer 2 the protocol layer under layer 3 
>>                       (which therefore offers the services used by layer 3)
>>       Layer 3         the protocol layer at which IP and its associated 
>>                       routing protocols operate
>> 
>> 2.3. Acronyms and Abbreviations defines
>> 
>>       L2              Layer 2
>>       L3              Layer 3
>> 
>> However, in 3.14. Scope and Uniqueness of Labels, 4.3. Label Stacks and Implicit Peering, 4.5. LSP Trees as Multipoint-to-Point Entities, and 4.6. LSP Tunneling between BGP Border Routers, L1, L2 and L3 are used as differentiating names for certain labels attached to packets. 
>> 
>> Of course, in 3.23. Time-to-Live (TTL), L2 is used to refer to layer 2 frame header and to a layer 2 switch, which is correct. 
>> 
>> However, in 4.3. Label Stacks and Implicit Peering, the term level 1 is used to refer to the LIFO (stack) ordinal number of a label then named L1 and given a protocol layer 2 protocol of layer 2 (L2). Furthermore, labels named L2 and then L1 are pushed onto the stack of labels prefixed to the packet. To top it all off the packet's stack attribute as protocol level 2 (L2). 
>> 
>> Of course, in 3.17. LSP Next Hop, 4.1.5. The Implicit NULL Label, 5.1.1.2. PushConditional, 5.1.1.4. PulledConditional, 5.1.2.2. RequestWhenNeeded, 5.1.3. Upstream LSR: NotAvailable Procedure, 5.1.4. Upstream LSR: Release Procedure, 5.1.4.2. NoReleaseOnChange, 5.1.5. Upstream LSR: labelUse Procedure, 5.2.2. Schemes for LSRs that do not Support Label Merging, refer to L3 meaning level 3, which is correct. 
>> 
>> Furthermore, in 3.1. Labels, 3.2. Upstream and Downstream LSRs, 3.4. 
>> Label Assignment and Distribution, 3.5. Attributes of a Label Binding, 
>> 3.14. Scope and Uniqueness of Labels, 4.1.2.2. Distributing Labels, 
>> 5.1.5. Upstream LSR: labelUse Procedure, 5.1.5.2. 
>> UseIfLoopNotDetected, 5.1.6. Downstream LSR: Withdraw Procedure
>> 
>> * L is used as a name for a certain label attached to packet, and
>> 
>> * L is used as a arbitrary value assigned to a label attached to a 
>> packet
>> 
>> 
>> Corrected Text
>> --------------
>> I have not provided any corrected text as I've literally "highlighted" 44 places in a pdf format file of RFC 3031 that are ambiguous. 
>> 
>> As there is no facility to attach a file to this Report Errata for RFC3031 form, i will send the file commented pdf file upon request. 
>> 
>> Notes
>> -----
>> My rational for highlighting (no pun intended) these problems is that the overloading of the L2, L3 abbreviations layer 2 and layer 3, with the names L1, L2, L3 and L for labels, plus the use of L1 and L2 as indexed names for the ordinal position of a label prefixed to a payload, then to use L2 and L3 as to actually mean layer 2 and layer is uh ... sloppy. 
>> 
>> Honestly, I can't understand how RFC 3031 has been posted for twenty years and that it is on the Standards Track and no one has found these problems. 
>> 
>> Its similar to when someone publishes a mathematical treatise and use the same set of variable names {x, y, z, t} over and over again in different contexts spread throughout the paper. Its intractable and practically gibberish. 
>> 
>> I apologize if my criticism is harsh regarding this problem but I spent a considerable amount of my time reading this document trying to make sense of it before I realized that the fault is not mine but it is of the document.
>> 
>> Instructions:
>> -------------
>> This erratum is currently posted as "Reported". If necessary, please 
>> use "Reply All" to discuss whether it should be verified or rejected. 
>> When a decision is reached, the verifying party can log in to change 
>> the status and edit the report, if necessary.
>> 
>> --------------------------------------
>> RFC3031 (draft-ietf-mpls-arch-06)
>> --------------------------------------
>> Title               : Multiprotocol Label Switching Architecture
>> Publication Date    : January 2001
>> Author(s)           : E. Rosen, A. Viswanathan, R. Callon
>> Category            : PROPOSED STANDARD
>> Source              : Multiprotocol Label Switching
>> Area                : Routing
>> Stream              : IETF
>> Verifying Party     : IESG
>> 
>> _______________________________________________
>> mpls mailing list
>> mpls@ietf.org
>> https://www.ietf.org/mailman/listinfo/mpls