Re: [802.3_SPMD] Comment 194

Thread Links	Date Links
Thread Prev	Thread Next	Thread Index	Date Prev	Date Next	Date Index

Val,

I see where you are going, but I think we need the second option.

In your first option, if we ended up getting to type 12 (heaven forbid), we couldn’t tell the difference between an MPD supporting type 1 and type 2 (type12) and an MPP supporting type 12 (type12).

What about this variation on your suggestion using “types” instead of “mtype”? I think it makes sense without needing an explanation.

type0 Type 0 MPD(s)

type1 Type 1 MPD(s)

types01 Able to operate as Type 0 or Type 1

type2 Type 2 MPD(s)

type3 Type 3 MPD(s)

types23 Able to operate as Type 2 or Type 3

types0123 Able to operate as Type 0, Type 1, Type 2, or Type 3

Regards,

Peter

_______________________________________________________________

Peter Jones Distinguished Engineer,

Cisco Networking Hardware

Chair, Ethernet Alliance

Mobile: +1 408 315 8024

Email: petejone@xxxxxxxxx

Web: https://about.me/petergjones

Webex: https://cisco.webex.com/meet/petejone

Book a call: Peter's booking page

_______________________________________________________________

From: George Zimmerman <george@xxxxxxxxxxxxxxxxxxxx>
Sent: Thursday, January 9, 2025 8:17 AM
To: STDS-802-3-SPMD@xxxxxxxxxxxxxxxxx
Subject: Re: [802.3_SPMD] Comment 194

Val & Jason – I think you get where I was going, and I’ll leave the nomenclature to you two. I think that perhaps having “mtype” for “multitype” will work in the management object, but might have some trouble in the actual text describing the types. We need to think of both.

George Zimmerman, Ph.D.

President & Principal

CME Consulting, Inc.

Experts in Advanced PHYsical Communications

george@xxxxxxxxxxxxxxxxxxxx

310-920-3860

From: Valerie Maguire <vmaguire@xxxxxxxx>
Sent: Thursday, January 9, 2025 7:44 AM
To: STDS-802-3-SPMD@xxxxxxxxxxxxxxxxx
Subject: Re: [802.3_SPMD] Comment 194

I like where Jason is going with this but think we might be getting too complicated for our own good! How about:

type0 Type 0 MPD(s)

type1 Type 1 MPD(s)

type2 Type 2 MPD(s)

type3 Type 3 MPD(s)

type01 Able to operate as Type 0 or Type 1

type23 Able to operate as Type 2 or Type 3

type0123 Able to operate as Type 0, Type 1, Type 2, or Type 3

We could do the following if it was felt that some differentiator indicating mixed was needed, although I don’t care for this as much…

type0 Type 0 MPD(s)

type1 Type 1 MPD(s)

mtype01 Able to operate as Type 0 or Type 1

type2 Type 2 MPD(s)

type3 Type 3 MPD(s)

mtype23 Able to operate as Type 2 or Type 3

mype0123 Able to operate as Type 0, Type 1, Type 2, or Type 3

Happy Thursday - Val

Valerie Maguire, BSEE

602-228-7943 mobile

From: Chad Jones (cmjones) <00000b60b3f54e8d-dmarc-request@xxxxxxxxxxxxxxxxx>
Sent: Thursday, January 9, 2025 9:36 AM
To: STDS-802-3-SPMD@xxxxxxxxxxxxxxxxx
Subject: Re: [802.3_SPMD] Comment 194

Agreed, nice solution Jason for terminology. I think this further complicates the solution in the text though and we will need a submission to consider as the solution. I’ll try to work on that today.

Regards,

Chad Jones

Principal Engineer, Cisco Systems

Executive Secretary, IEEE 802.3 Working Group

Chair, IEEE P802.3da Task Force

Principal, NFPA 70 CMP3

From: George Zimmerman <george@xxxxxxxxxxxxxxxxxxxx>
Date: Wednesday, January 8, 2025 at 9:04 PM
To: STDS-802-3-SPMD@xxxxxxxxxxxxxxxxx <STDS-802-3-SPMD@xxxxxxxxxxxxxxxxx>
Subject: Re: [802.3_SPMD] Comment 194

Jason, thanks. It looks like you have a decent solution, and we’ll deal with the multiplicity, if and when we extend the types. Both was just completely nonspecific. Multi-type_0_1 is useful and shows the path

George Zimmerman, Ph.D.

President & Principal

CME Consulting, Inc.

Experts in Advanced PHYsical Communications

george@xxxxxxxxxxxxxxxxxxxx

310-920-3860

On Jan 8, 2025, at 5:55 PM, Jason Potterf (jpotterf) <jpotterf@xxxxxxxxx> wrote:

I would be in favor of renaming as mixed is very close to mixing segment and could cause confusion. That said, I think it's more complicated than we'd like if we want to make this extensible.

This type can't indicate "Supports any Type " as the future type requirements are unknown to current devices implemented under 802.3da. So it MUST mean "Supports BOTH Type 0 and Type 1" now and forever. In a future with a Type 2 and Type 3, we'd have to have a scheme like this with additional enumerated types:

type0 Type 0 MPD(s)

type1 Type 1 MPD(s)

multi-type_0-1 Able to operate as Type 0, or Type 1

type2 Type 2 MPD(s)

type3 Type 3 MPD(s)

multi-type_2-3 Able to operate as Type 2, or Type 3

multi-type_0-3 Able to operate as Type 0, Type 1, Type 2, or Type 3

It's ugly, but it's accurate.

Jason Potterf

jpotterf@xxxxxxxxx

From: George Zimmerman <george@xxxxxxxxxxxxxxxxxxxx>
Sent: Wednesday, January 8, 2025 6:08 PM
To: STDS-802-3-SPMD@xxxxxxxxxxxxxxxxx <STDS-802-3-SPMD@xxxxxxxxxxxxxxxxx>
Subject: Re: [802.3_SPMD] Comment 194

Peter – For extensibility I wouldn’t do that. Both isn’t extensible if anyone ends up adding new types. If we want to have a polymorph type, then it needs to generalize. However, maybe ‘Mixed’ isn’t the right name.

If we do rename the type, then we need to do it throughout clause 189.

George Zimmerman, Ph.D.

President & Principal

CME Consulting, Inc.

Experts in Advanced PHYsical Communications

george@xxxxxxxxxxxxxxxxxxxx

310-920-3860

From: Peter Jones (petejone) <00000b5d1d72f221-dmarc-request@xxxxxxxxxxxxxxxxx>
Sent: Wednesday, January 8, 2025 3:29 PM
To: STDS-802-3-SPMD@xxxxxxxxxxxxxxxxx
Subject: Re: [802.3_SPMD] Comment 194

Folks,

I think we should change

‘type Mixed’ (able to be a 0 or 1)

‘type Both (able to be a 0 or 1)

Regards,

Peter

_______________________________________________________________

Peter Jones Distinguished Engineer,

Cisco Networking Hardware

Chair, Ethernet Alliance

Mobile: +1 408 315 8024

Email: petejone@xxxxxxxxx

Web: https://about.me/petergjones

Webex: https://cisco.webex.com/meet/petejone

Book a call: Peter's booking page

_______________________________________________________________

From: George Zimmerman <george@xxxxxxxxxxxxxxxxxxxx>
Sent: Tuesday, January 7, 2025 2:04 PM
To: STDS-802-3-SPMD@xxxxxxxxxxxxxxxxx
Subject: Re: [802.3_SPMD] Comment 194

Thank you, Chad. For others, this became a big ticket item because when Chad and I were discussing the proposed responses it became clear that we were talking about different meanings of “mixed”.

It sounds now like we are talking only about the meaning of the output of the management object. An MPD can be a ‘type 0’ ‘type 1’ or a ‘type Mixed’ (able to be a 0 or 1), and an MPSE in the state diagram detects ‘type Mixed’ as a separate thing… It sounds like for the management object we want:

Type 0 – only type 0 or a mixture of type 0 and type Mixed MPDs are detected.

Type 1 – only type 1 or a mixture of type 1 and type Mixed MPDs are detected.

Mixture – a mixture of type 0 and type 1 MPDs are detected. This includes a mixture which also has Type Mixed MPDs.

We still need to cover the case where ONLY Type Mixed are discovered, which was the case considered in the comment, and we need to decide if we want to a value of

Mixed – only type Mixed MPDs are detected.

Regardless, it seems the required solution goes beyond what the commenter’s resolution (my resolution) envisioned, needing the extra words in type 0, type 1 and the addition of mixture…

George Zimmerman, Ph.D.

President & Principal

CME Consulting, Inc.

Experts in Advanced PHYsical Communications

george@xxxxxxxxxxxxxxxxxxxx

310-920-3860

From: Chad Jones (cmjones) <00000b60b3f54e8d-dmarc-request@xxxxxxxxxxxxxxxxx>
Sent: Tuesday, January 7, 2025 1:11 PM
To: STDS-802-3-SPMD@xxxxxxxxxxxxxxxxx
Subject: [802.3_SPMD] Comment 194

As I’m making my way through the comments, I will use the reflector to discuss things that I don’t think need a whole presentation. Here’s my first one:

Comment 194

30.17.1.1.3

the aMPSETypeDiscovery enumerated values miss the case where type "mixed" MPDs are discovered… they just have the case where Both type 0 and type 1 MPDs have been discovered. This case should also include type "mixed" discovered, as listed in 30.17.2.1.1 (as well as in clause 189)

Change description of "mixed" to read "Type Mixed, or a mixture of MPD Types"

PROPOSED ACCEPT IN PRINCIPLE.

TFTD.

(Editor's note: We need to determine with "Mixed" is a Type, or if it is just that the MPD change types just like the MPSE can change types.)

Commentor has a point that mixed is missing from this text, but the remedy doesn’t fix it. The answer is here is that we are misusing mixed in this context. Type0 should mean that the PSE has only discovered type 0 and type mixed MPDs. Similarly type1 means only type 1 and type mixed MPDs are present. Mixed REALLY means I have a mixture of type 1 and type 0, and in this case we don't care it there are any type mixed present. The main point is I have two PD types that are incompatible with interoperation. We should find a new name for mixed here, "blended" is a better description. There could be a fourth case called mixed where it only discovered type mixed MPDs and the PSE can then power as whatever type it is.

Regards,

Chad Jones

Principal Engineer, Cisco Systems

Executive Secretary, IEEE 802.3 Working Group

Chair, IEEE P802.3da Task Force

Principal, NFPA 70 CMP3

To unsubscribe from the STDS-802-3-SPMD list, click the following link: https://listserv.ieee.org/cgi-bin/wa?SUBED1=STDS-802-3-SPMD&A=1