Distance Objectives: 2nd motion

["Jonathan Thatcher" <jonathan@xxxxxxxxxxxxx>]

All,

Assuming general agreement on my first proposed motion (or maybe even if not): there are several, potential interesting distances that are candidates (there may be more, but this seems to be an adequate list to start with):

1. 100 meters over MMF (850 and 1300)
2. 300 meters over MMF (850? and 1300?)
3. 500 (550) meters over MMF (850? and 1300?)
4. 3 km over SMF (1300)
5. 5 km over SMF (1300)
6. 10 km over SMF (1300)
7. 20-40 km over SMF (1300)
8. 80-120 km over SMF (1500)

To sort through these, I think we need to consider some of the "traditional" assumptions/arguments/positions that were used in Gigabit Ethernet when making decisions about fiber, wavelength, and distance:

a. We should support the existing infrastructure (meaning the existing cable plants)
b. We should minimize the number of PHY choices
c. We should minimize the cost of implementation

There is a corollary to "a." (called a' = "a prime") which is: we should try to go as far as we went at slower speeds over the same media.

Let me start with 4, 5, and 6. Since we currently specify 5 km in the 1000BASE-LX standard, according to "a.", we should support at least 5 km. Using this assumption, we would eliminate 4. To the best of my knowledge, all the LX solutions in existence support 10 km (these parts typically meet the Fibre Channel 10 km solution while running at the GbE speed), effectively making "a" a 10 km requirement

Debatable issue 1 is therefore: support the LX distance as specified or as commonly implementable in the industry (5 or 6).

Questions about 7 are most interesting if the group picks 6 over 5 but could be independent, I will assume the former. We are all aware that solutions to the 10 km LX actually run at much greater distances because of the conservatism of the standard. We therefore see practical applications that achieve 20 to 40 km. These are not guaranteed to be interoperable.

Debatable issue 1a is therefore (assuming issue 1 selects 6 over 5): support the LX distance as commonly implementable or as available in the industry (6 or 7).

Questions about 8 are similar to those of 7 when taken independently. Most arguments supporting 7 would also apply to 8. The major rationale for doing either would be to take what are now proprietary solutions and make them interoperable. If members believe that the market for these solutions will be significant for 10 gig, then there will be interest in making these solutions interoperable (standards based). If not, this will continue to be a custom space.

Debatable issue 2 is therefore: How interesting is this potential market space. Is it sufficient to create a standard?

1, 2, and 3 are significantly more difficult. Especially since we do not yet know if we can support assumption "a." without significantly impacting "c." (I can barely imagine the flurry of opinions on this topic). We also don't know where the majority opinion will go with respect to specifying new fiber types (e.g. high bandwidth MMF). In short, it might not be possible to satisfy both "a" and "c" simultaneously to the convincing of 75% of the members.

Debatable issue 3 is therefore: what gives first "a" or "c"?

There is more that can be said here, but I have a more important idea I would like to convey.

I would recommend the following: first, conclude the 1st general motion on space per my previous append. The second section on that motion will help to the members to resolve what they are going to do about about the debatable issues here. The third portion of that motion does not impact this discussion, since connection to the WAN can be accomplished with any distance solution.

Next, bring forward a main motion that is easily passable and covers the least controversial components of 1 through 8. I recommend a selection of 4, 5, or 6. (assuming you think that only one of the three will be kept (see "b")). Each can be placed in the motion text with the intent to select only one. Votes can be taken from least to most likely to succeed to see where the 75% vote can be achieved.

Note: This technique is frequently used in discussions of allocation of money: "We move to collect from the membership ($0, $100, $1000, $50,000, $1,000,000) to be given to the HSSG chair to use at his discretion." You vote starting with the $1,000,000 and continue down the list till you reach 75% consensus,  then stop. This is simpler and faster than picking a single number and having it amended to death. The 0$ is, of course, silly here because if something higher than or equal to $100 is not passed, the motion failed anyway.

Then make a motion(s) to add a line(s) to the previous objective (now a passed motion) that includes something about 1 through 3. Given the diversity of the group, a more general motion has a greater chance of success than a more restrictive motion. An overly general motion ("do the best you can on MMF") has no chance. The same technique could be used as above (support a distance of at least (100, 300, 550) meters over multimode fiber). I am assuming here that the group will want to have some MMF solution (see "a") and will therefore pass at least 100 meters. Note that this motion would set the lower bound!

Make a motion to add to the previous objective a direction to investigate, or not investigate the use of higher BW MMF.

Make a motion to add to the previous objective support of a 1550 nm solution.

Make a motion to add to the previous objective support of a longer distance 1300 nm solution.

etc.

Explain the process and content of each motion to the group prior to starting. This way the membership will understand when the appropriate time is to discuss/amend the motion (i.e. per the above process, the chair would consider out of order any amendment and/or discussion during the initial motion on LW, SMF, that relates to MMF as this would be covered in a subsequent motion). The only problem is, I think, that someone might think that at this point of the process the HSSG should investigate more potential solutions than should be supported in the end. If so, then add another motion that adds a line to the objective to limit the number of solutions to N in the final standard.

I am suggesting this process because I believe that we must exit the process with an objective on distance, which is the mission of this ad-hoc. I am optimistic that this or a similar procedure would accomplish that goal. There are alternatives to this which would be simpler and quicker (for instance, a single motion that contains the correct wording to allow it to pass with 75% vote). The problem is, if that vote fails, we got nuthin' and that ain't acceptable.

jonathan