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

Re: [HSSG] HSSG Reach Ad Hoc "Rides Again"



At the risk of sounding obvious:

- The cost of creating a PMD which can handle higher insertion loss (due to more connectors/patch panels etc.) is not equal to the cost of increasing the distance
- From a simplistic point of view, you can understand that the transmission over the SMF media will introduce more penalties the further it is transmitted down the fiber

- Let's take an example (see Note below):
PMD-HighConnectorLoss
 - 3km SMF transmission = 1.5dB loss
 - 5.5dB connector loss (~11 connectors)
PMD-10km
 - 10km SMF transmission = 5dB loss
 - 2dB connector loss (~4 connectors)

- The PMD-HCL would require high output power light sources and high sensitivity diodes
- The PMD-10km would require similarly high output power light sources and high sensitivity diodes, but the added penalty of 7km of transmission will increase the overall penalty number
- The final penalty value will be a matter which all of the optical folks will study in detail, but we can certainly agree that the delta in penalty between 3 & 10km transmission is greater than 0.5dB and less than 3dB. (My personal opinion is between 1~2dB).

In conclusion, the reach question and the number of connectors are both very key data points which we need to answer.  However, please note that the "cost" of increasing these points is different.

Note: I used 0.5dB/km to simplify calculations for email/reflector purposes -- for accuracy strongly recommend that folks refer to anslow_03_1107.xls.  Also I used 0.5dB per connector -- I would like to understand if this is a useful "modern" value given the splice issues that Mike Dudek has pointed out.

thanks
--matt traverso

PS - Please reply to my Opnext e-mail address (mtraverso AT opnext.com) -- this gmail account is my personal account.  I only use it to bypass default notices which Opnext inserts onto every e-mail I send.  These default notices are not typically considered friendly to the IEEE reflector.

On Nov 19, 2007 9:29 AM, Chris Cole <chris.cole@finisar.com> wrote:
>  
>  
>  
>
> During discussion of the 100GE SMF reach objective at the Atlanta meeting,
> an interesting new perspective was given on how optics are used in the data
> center. Specifically, it was suggested (for example by Ted Seely of Sprint)
> that the number of connectors used in typical connections is much higher
> then the 4 or 5 connectors assumed in SMF link budgets. For these
> applications, the availability of a 10km link budget (in the 10GBASE-LR
> standard) has allowed supporting this higher number of connectors over data
> center reaches shorter then 10km.
>
>  
>
> As part of the HSSG Reach Ad Hoc, it may be beneficial to quantify the
> distribution of number of connectors in links as part of the effort to
> quantify the distribution of link reaches. We may still choose to use 2dB as
> an allocation for connector loss (corresponding to 4 or 5 connectors) in SMF
> link budgets, but would use this additional information to determine the
> best reach objectives. Further, we may include informative information as
> part of the standard to give guidance on how long reach optics can be used
> for shorter reach applications requiring a large number of connectors.
>
>  
>
> Chris
>
>  
>  
>  ________________________________
>  
>
> From: Andy Moorwood [mailto:amoorwood@EXTREMENETWORKS.COM ]
>  Sent: Monday, November 19, 2007 8:27 AM
>  To: STDS-802-3-HSSG@LISTSERV.IEEE.ORG
>  Subject: [HSSG] HSSG Reach Ad Hoc "Rides Again"
>
>  
>
>  
>
> Colleagues,
>
> At the Atlanta meeting several presentations were made in the area of
> requirements for, and relative cost of, reaches lower than 10km over single
> mode fiber.  In the spirit of promoting consensus building in the group John
> has asked me to organize a conference call so we can have further discussion
> prior to the interim meeting in January.
>
>  
>
> I will forward contact information to the reflector shortly but would ask
> those members interested in these issues to mark in their calendars  for a
> conference call at 9am to 11 am ( U.S. Pacific time zone) on Thursday
> December 6th.
>
>  
>
> Presentations from Atlanta may be found on the study group web site, links
> to some of the more relevant presentations are included here for
> convenience.
>
>  
>
> http://grouper.ieee.org/groups/802/3/hssg/public/nov07/cole_01_1107.pdf
>
> http://grouper.ieee.org/groups/802/3/hssg/public/nov07/goergen_01_1107.pdf
>
> http://grouper.ieee.org/groups/802/3/hssg/public/nov07/cole_03_1107.pdf
>
>  
>
>  
>
> For the December conf call it would really help the group if we could
> receive further contributions on :
>
> a)       the need for a sub 10km SMF reach
>
> b)       relative costing information  for the different reaches
>
>  
>
>  
>
> As of now the agenda for the call is as follows
>  
> discussion of the reach and relative costing presentations made at Atlanta
> discussion of any new material posted to the reflector
> discussion of how to resolve
>
>  
>
> Many Thanks
>
> Andy Moorwood
>
>