Re: [8023-10GEPON] Downstream wavelength review
Dear all,
A valid point I believe. I would hate to see the system cost increase without any point just for the sake of a nice label on the OLT saying it is all class compatible. The power budget differences are simply too high to consider using a single, all class compatible OLT.
On the other hand, allowing for "deviations" is asking for trouble. I am not sure how the 802.3 group would accept a system where the transmission wavelength is defined on the best effort basis. Honestly, I prefer clear cut distinctions - PR10 in this band, PR20 in that one etc. and I am afraid that most of the 802.3 voters we need to convince to our solution is quite similar.
My personal preference is quite aligned with Frank's - PR10/20 in a CDWM window to use existing transmitters and forget about filtering problems. PR30 in a specific window since we already know we need new transmitters and cooling so we can accept higher cost. No need to burdern PR20/10 systems with added cost in the name of nice uniform numbers in the tables. If the ONU receiver can handle it with no problems, I would say let's stick with this plan.
Unless someone can prove there is added value in specyfing PR10/20 classes to 1574 - 1600 nm window. Apart from uniform numbers, IMHO I see none.
BTW: should we use PX indicator for the new EPONs or rather PR ? I am confused already ...
Best wishes
Marek Hajduczenia (141238)
NOKIA SIEMENS Networks S.A., Portugal - R
Rua Irmãos Siemens, 1
Ed. 1, Piso 1
Alfragide
2720-093 Amadora
Portugal
* Marek.Hajduczenia@siemens.com
http://www.marekhajduczenia.info/index.php
(+351.21.416.7472 4+351.21.424.2082
"C makes it easy to shoot yourself in the foot; C++ makes it harder, but when you do, it blows away your whole leg." - Bjarne Stroustrup
-----Original Message-----
From: Frank Effenberger [mailto:feffenberger@HUAWEI.COM]
Sent: quinta-feira, 30 de Agosto de 2007 16:17
To: STDS-802-3-10GEPON@listserv.ieee.org
Subject: Re: [8023-10GEPON] Downstream wavelength review
Dear Ken-Ichi,
I have a question about your idea: Why would somebody want to make an OLT that supports PX10-20-30?
The whole point of making a PX10 or PX20 is to make a cheaper OLT. If you make an OLT that supports PX30, you will have already spent the extra cost to make the high-power OLT... You don't gain anything to down grade it.
If you are thinking about just being able to say, "My OLT complies with PX10/20/30 optical specs," well, I doubt you will ever be able to do that. We haven't finalized the entire power budget yet, but it seems that in most proposals the PX10/20 are considerably lower in power. Are you going to try to implement power control? That's more cost, to make an OLT that does less!
I don't think it is an attractive feature.
Marek correctly pointed out that the use of a band from 1580 to 1600 is very attractive for the reason that it coincides with the CWDM band plan. I worry that if we define the PX10 and PX20 band to be the 'odd' 1574 to 1600nm, then it confuses the issue.
Toward a compromise, would you accept a situation where we specify the PX10 and PX20 Tx bands to be 1580 to 1600nm, but we add a note like:
Note: Deviations of the PX10 and PX20 Tx wavelength down to 1574nm are permissible.
You may think this sounds strange, but I really think it adds value, in that the nominal wavelengths remain the 'normal sounding' 1580 to 1600 values, yet it gives you the latitude that you want.
How about that?
Sincerely,
Frank E.
-----Original Message-----
From: Ken-Ichi Suzuki [mailto:kenyichi@ansl.ntt.co.jp]
Sent: Thursday, August 30, 2007 4:37 AM
To: STDS-802-3-10GEPON@listserv.ieee.org
Subject: Re: [8023-10GEPON] Downstream wavelength review
Dear Marek and all,
Thank you for your answer.
So I believe we should not limit the wavelength range of PX10/20
transmitters considering both the use of CWDM grid and compatibility
to PX30 cooled-lasers, if there are no reasons for that limitation.
Best regards,
Ken-Ichi
At 2007/08/30 17:23 Hajduczenia, Marek wrote:
> Dear Otaka-san,
>>From what I gather from the presentations provided so far, the OLT transmitters required for PR30 systems have different requirements than PR10/PR20 ones - for once, they require (most likely) cooling which is not required (at least that is what I gather) for PR10s and PR20s. I would like to learn the opinions of components vendors - they are more likely to be familiar with market availability of 1580 - 1600 nm devices meeting PR10/20 requirements.
> Please note also that the ONU receiver remains a universal device, with the sensitivity window spanning between 1574 and 1600 nm, thus covering both PR10/20 and PR30 devices on the other end of the link. The only differentation here would be the OLT transmitter, nothing else.
> Any other comments ?
> Thank You for Your feedback
>
> Marek Hajduczenia (141238)
> NOKIA SIEMENS Networks S.A., Portugal - R
> Rua Irmãos Siemens, 1
> Ed. 1, Piso 1
> Alfragide
> 2720-093 Amadora
> Portugal
> * Marek.Hajduczenia@siemens.com
> http://www.marekhajduczenia.info/index.php
> (+351.21.416.7472 4+351.21.424.2082
> "C makes it easy to shoot yourself in the foot; C++ makes it harder, but when you do, it blows away your whole leg." - Bjarne Stroustrup
>
> -----Original Message-----
> From: Akihiro Otaka [mailto:ootaka@ansl.ntt.co.jp]
> Sent: quinta-feira, 30 de Agosto de 2007 9:18
> To: Hajduczenia, Marek; STDS-802-3-10GEPON@listserv.ieee.org
> Subject: Re: [8023-10GEPON] Downstream wavelength review
>
> Dear Marek and all.
>
> Thank you for your prompt comment.
> This is Akihiro Otaka.
>
> I think if there are someone who try to realize B++ and PX20/10
> OLT with identical device (it may be a cooled device), the superset
> band idea is better for them.
>
> Are there no such requirement in practice?
>
> Best regards,
> Akihiro Otaka
>
>
> At 16:18 07/08/30, Hajduczenia, Marek wrote:
> >Dear Suzuki-san,
> >
> >I believe I may answer this question since Frank is probably still at night
> >time (Frank, please confirm if I what I am saying is OK) ...
> >The main reason why Frank proposes to have PR10/PR20 PMDs use the 1580 -
> >1600 nm window in the downstream is the compatibility with the CDWM
> >wavelength grid and the availability of uncooled transmitters centered
> >around 1590 nm with the power putput sufficient to cope with these
> >particular power budgets. You are right that it does little harm to expand
> >the band to 1574 - 1600 though the big question is whether it will be used
> >in practice. I do not see a reason to block part of the band which will not
> >be used by the PMDs anyway.
> >
> >Hope that answers Your question
> >
> >Marek Hajduczenia (141238)
> >NOKIA SIEMENS Networks S.A., Portugal - R
> >Rua Irmテ」os Siemens, 1
> >Ed. 1, Piso 1
> >Alfragide
> >2720-093 Amadora
> >Portugal
> >* Marek.Hajduczenia@siemens.com
> >http://www.marekhajduczenia.info/index.php
> >(+351.21.416.7472 4+351.21.424.2082
> >"C makes it easy to shoot yourself in the foot; C++ makes it harder, but
> >when you do, it blows away your whole leg." - Bjarne Stroustrup
> >
> >-----Original Message-----
> >From: Ken-Ichi Suzuki [mailto:kenyichi@ansl.ntt.co.jp]
> >Sent: quinta-feira, 30 de Agosto de 2007 8:00
> >To: STDS-802-3-10GEPON@listserv.ieee.org
> >Subject: Re: [8023-10GEPON] Downstream wavelength review
> >
> >Dear Frank
> >
> >Thank you for your proposal.
> >Basically, I agree to your proposal.
> >But I have a comment on Option D.
> >
> >I believe the full wavelength range of 1574 to 1600 nm can be used
> >for PX10 and PX20 in Option D.
> >If someone wants to use the range of 1574 to 1580 for PX10 and PX20
> >as well as the range of 1580 to 1600 nm, I think we should not limit
> >the wavelength range of Option D.
> >
> >So I would like to confirm whether we should limit the wavelength
> >range because I believe that the specifications should be accepted
> >by as many people as possible (although I do not have a strong
> >opinion to PX10 and PX20).
> >
> >Best regards,
> >Ken-Ichi
> >
> >At 2007/08/29 0:07 Frank Effenberger wrote:
> >> Dear All,
> >>
> >>
> >>
> >> I have put together some slides that review the downstream wavelength issue,
> >> and put forward a solution that I think may have some common support.
> >>
> >>
> >>
> >> Please give me your comments, and if you would like to support it, let me
> >> know that, also.
> >>
>
>
>
--
Ken-Ichi Suzuki
NTT Access Network Service Systems Labs.
E-mail:kenyichi@ansl.ntt.co.jp
Tel:+81-43-211-3189/Fax:+81-43-211-8250