Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss
GFP Vs GMP
1,407 views
Skip to first unread message
Ramakrishna Pratapa
Aug 14, 2015, 3:26:31 AM8/14/15
to
Hi All,
Did GFP replace GMP? Or are there cases where GMP is preferred over GFP? If yes, what are they?
Could you help?
Rgds
RK
Did GFP replace GMP? Or are there cases where GMP is preferred over GFP? If yes, what are they?
Could you help?
Rgds
RK
Huub van Helvoort
Aug 14, 2015, 4:51:10 AM8/14/15
to
Hello Ramakrishna,
You wrote:
> Did GFP replace GMP? Or are there cases where GMP is preferred over
> GFP? If yes, what are they?
>
> Could you help?
GFP, as described in ITU-T G.7041 is used to map variable bitrate
packet streams into constant bitrate SDH (VC-n) and OTN (OPUk) frames.
To each received packet GFP-F adds GFP overhead, and when insufficient
packets arrive it inserts special idle frames.
GFP-T is similar, it uses code blocks instead of packets.
GMP, as described in G.709, is used in OTN to map any client
signal into constant bitrate OPUk, difference in bitrate is
compensated by inserting stuff-bytes. The number of stuff-bytes
and they location is calculated constantly, so variable bit rate
cleint signals can be mapped too.
Best regards, Huub.
--
reply to hhelvooort with 2 'o's
================================================================
http://www.van-helvoort.eu/
================================================================
Always remember that you are unique...just like everyone else...
Ramakrishna Pratapa
Aug 14, 2015, 5:55:36 AM8/14/15
to
Hi Huub,
Thank you, for the reply. Let me restate two of your statements:
1) GFP, as described in ITU-T G.7041 is used to map variable bitrate packet streams into constant bitrate SDH (VC-n) and OTN (OPUk) frames.
2) GMP, as described in G.709, is used in OTN to map any client signal into constant bitrate OPUk,
In both the cases, the output is a constant bit rate frame. Further, when we say GMP can map any client signal, does that include 'variable bit rate packet stream' as in case of GFP? If yes, this gives a sense of GMP superseding GFP. That is where my confusion is!
Rgds
RK
Thank you, for the reply. Let me restate two of your statements:
1) GFP, as described in ITU-T G.7041 is used to map variable bitrate packet streams into constant bitrate SDH (VC-n) and OTN (OPUk) frames.
2) GMP, as described in G.709, is used in OTN to map any client signal into constant bitrate OPUk,
In both the cases, the output is a constant bit rate frame. Further, when we say GMP can map any client signal, does that include 'variable bit rate packet stream' as in case of GFP? If yes, this gives a sense of GMP superseding GFP. That is where my confusion is!
Rgds
RK
Huub van Helvoort
Aug 14, 2015, 7:30:22 AM8/14/15
to
Hello Ramakrishna,
You replied:
> Thank you, for the reply. Let me restate two of your statements:
>
> 1) GFP, as described in ITU-T G.7041 is used to map variable bitrate
> packet streams into constant bitrate SDH (VC-n) and OTN (OPUk) frames.
And GFP-F is mapping based on packetstream which is variable bitrate.
While GFP-T is based on the bitstream which is constant bitrate.
> 2) GMP, as described in G.709, is used in OTN to map any client signal
> into constant bitrate OPUk,
GMP supersedes AMP, and maps constant bitrate signals.
> In both the cases, the output is a constant bit rate frame.
Correct.
> Further, when we say GMP can map any client signal, does that include
> 'variable bit rate packet stream' as in case of GFP?
No, in this case GFP-F has to be used to make a CBR first.
See clause 17 of G.709 for a description of all mappings.
> If yes, this gives a sense of GMP superseding GFP. That is where my confusion is!
I hop I removed the confusion.
Best regards, Huub.
You replied:
> Thank you, for the reply. Let me restate two of your statements:
>
> 1) GFP, as described in ITU-T G.7041 is used to map variable bitrate
> packet streams into constant bitrate SDH (VC-n) and OTN (OPUk) frames.
While GFP-T is based on the bitstream which is constant bitrate.
> 2) GMP, as described in G.709, is used in OTN to map any client signal
> into constant bitrate OPUk,
> In both the cases, the output is a constant bit rate frame.
> Further, when we say GMP can map any client signal, does that include
> 'variable bit rate packet stream' as in case of GFP?
See clause 17 of G.709 for a description of all mappings.
> If yes, this gives a sense of GMP superseding GFP. That is where my confusion is!
Best regards, Huub.
Ramakrishna Pratapa
Aug 17, 2015, 12:38:31 AM8/17/15
to
Hi Huub,
Thank you very much for the comments and very sorry for the delay from my end.
Just to be clear..
For variable bit rate client signals, the two stage process is:
step1: GFP-F would map the variable bitrate client signals onto OPUk
step2: OPUk gets mapped into ODUk using GMP. Correct?
But On the other hand, in case of constant bit rate client signals, is GFP-T necessary followed by GMP? GMP alone seems to be sufficient?
Rgds
RK
Thank you very much for the comments and very sorry for the delay from my end.
Just to be clear..
For variable bit rate client signals, the two stage process is:
step1: GFP-F would map the variable bitrate client signals onto OPUk
step2: OPUk gets mapped into ODUk using GMP. Correct?
But On the other hand, in case of constant bit rate client signals, is GFP-T necessary followed by GMP? GMP alone seems to be sufficient?
Rgds
RK
Huub van Helvoort
Aug 25, 2015, 9:57:20 AM8/25/15
to
Hello Ramakrishna,
You replied:
> Just to be clear..
>
> For variable bit rate client signals, the two stage process is:
> step1: GFP-F would map the variable bitrate client signals onto OPUk
Yes, this is correct.
Note that GMP is not needed, because GFP can fill the OPU completely.
> step2: OPUk gets mapped into ODUk using GMP. Correct?
No, the OPU is part of the ODU, the difference between OPU and ODU is
only the overhead bytes.
> But On the other hand, in case of constant bit rate client signals,
> is GFP-T necessary followed by GMP? GMP alone seems to be sufficient?
When GFP-F or GFP-T is used the GFP will completely fill the OPU
no GMP is required.
Best regards, Huub.
--
reply to hhelvooort with 2 'o's
================================================================
http://members.chello.nl/hhelvoort/
You replied:
> Just to be clear..
>
> For variable bit rate client signals, the two stage process is:
> step1: GFP-F would map the variable bitrate client signals onto OPUk
Note that GMP is not needed, because GFP can fill the OPU completely.
> step2: OPUk gets mapped into ODUk using GMP. Correct?
only the overhead bytes.
> But On the other hand, in case of constant bit rate client signals,
> is GFP-T necessary followed by GMP? GMP alone seems to be sufficient?
no GMP is required.
Best regards, Huub.
--
reply to hhelvooort with 2 'o's
================================================================
Ramakrishna Pratapa
Aug 28, 2015, 1:59:09 AM8/28/15
to
Hi Huub,
So, GFP-F/ GFP-T can perform the encapsulation of variable/ constant bit rate signals. And GMP is not a necessity in either of the cases.
Is there something that GMP can do but GFP cannot? If not, can we say GFP can replace GMP? (I am ignoring the cost, even if it is a factor driving that choice).
Rgds
RK
So, GFP-F/ GFP-T can perform the encapsulation of variable/ constant bit rate signals. And GMP is not a necessity in either of the cases.
Is there something that GMP can do but GFP cannot? If not, can we say GFP can replace GMP? (I am ignoring the cost, even if it is a factor driving that choice).
Rgds
RK
Huub van Helvoort
Sep 1, 2015, 11:09:29 AM9/1/15
to
Hello Ramakrishna,
You replied:
> So, GFP-F/ GFP-T can perform the encapsulation of variable/ constant
> bit rate signals. And GMP is not a necessity in either of the cases.
Indeed.
> Is there something that GMP can do but GFP cannot? If not, can we say
> GFP can replace GMP? (I am ignoring the cost, even if it is a factor
> driving that choice).
Yes, GMP can be used for any new client signal. From now on it is not
necessary to standardise the mapping for any new client signal.
This was mandatory in the past, before GMP was introduced.
So conclusion: always use GMP because it is simpler, cheaper.
The GMP circuit can be re-used. Only the receiver and transmitter
circuits for each new signal will be different.
Best regards, Huub.
You replied:
> So, GFP-F/ GFP-T can perform the encapsulation of variable/ constant
> bit rate signals. And GMP is not a necessity in either of the cases.
> Is there something that GMP can do but GFP cannot? If not, can we say
> GFP can replace GMP? (I am ignoring the cost, even if it is a factor
> driving that choice).
necessary to standardise the mapping for any new client signal.
This was mandatory in the past, before GMP was introduced.
So conclusion: always use GMP because it is simpler, cheaper.
The GMP circuit can be re-used. Only the receiver and transmitter
circuits for each new signal will be different.
Best regards, Huub.
Ramakrishna Pratapa
Sep 2, 2015, 3:05:03 AM9/2/15
to
Thank you Huub!!
Kwati Gurunath
Sep 20, 2021, 8:22:43 AM9/20/21
to
GMP and GFP mapping procedure have their own advantages and disadvantages.
Since GMP is completely transparent in nature, any failure like link toggling on line side for short duration like 4ms/5ms will make node to send this out transparently to far end device which is connected to Gige client, due to this small glitch/failure on far end device router Gige interface any protocols like BGP/OSPF session gets terminated or find another root till this path gets recovered completely in their route table. Advantage of this BMP is that small latency can be achived.
With GFP mapping used, above mentioned issue can not be seen as in GFP mapping mapping procedure terminates complete signal and so it can hold any small duration like 4/5ms duration link toggling issues so far end device do not experience this small duration link toggling. And disadvantage of is that as there payload termination involved, there is a processing time involved hence high latency time when to GMP mapping signal. So both GMP and GFP mapping procedures are bad and good.
0 new messages