[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: draft-otani-ccamp-interas-gmpls-te-00.txt

Hi Adrian and Tomo,

Thank you for your interesting discussion.

Though I am not an EGP expert, I suspect that it will be problematic to advertise the TE information by an EGP-TE from the viewpoints of scalability and complexity. Particularly, the TE information will be confidential in the Inter-Provider case even if such TE information can be aggregated.
The PCE model can solve this problem because no TE information is advertised to the externals and only the result of path computation is notified to the external node requesting the path computation (Head-end LSR, ASBRs, other PCEs, etc.). Even in the PCE model, the crankback function is necessary because the TE sate may change rapidly and the time lag between the path computation and the path establishemt may be relatively large.

As you pointed out, different SPs may have TE links included in the same SRLG. Moreover, even if appropriate SRLG can be allocated to each TE-link gloabally, such SRLG information will also be confidential in the Inter-Provider environment.
At least, the SRLG can be allocated appropriately within a provider's network according to the provider's policy, and multi-paths with the SRLG diversity can be correctly established within a provider's network. Thus, it is practical to establish the working path and the corresponding backup path through the same sequence of provider's networks in the case of Inter-Provider LSP establishment.
Anyway, if both the working path and the backup path have damage simultaneously, dynamic re-routing is necessary.

As you pointed out, it is required to develop the SRLG encoding that can indicate the SRLG level (area, duct, cable, fiber, etc.). Each SP can establish a pair of working path and backup path with the appropriate SRLG diversity and satsfying the SLA when he/she can recognise the level of each SRLG.

Best regards,

Nagao Ogino


At 10:44 04/07/28, Tomohiro Otani wrote: 
Hi Adrian,

Thank you for your comments.

Some of SPs including ourselves have a plan to introduce GMPLS control plane
technologies for a NGN, and even NGNs should be interconnected with each other,
as SONET/SDH networks are as well as IP/MPLS networks are.

Your supposed model is another example to reveal the necessity of Inter AS TE.
I agree that a crankback mechanism is one of the ways to achieve this.

In terms of the amount of computation by the ASBR, we have to investigate
the impact on the load to the ASBR, hearing from the routing (EGP: BGP4?) expert.
That's why, in the draft, we described that the TE information may be statically or
dynamically exchanged. Operators would like to know the route of a GMPLS LSP
within an AS and the pair of ASBR in advance, if the LSP is established across
the inter AS, especially multiple ISPs, resulting in the implementation of TE extensions.

The network resiliency is getting significant more and more in a SP environment.
There was a big discussion about SRLGs among us and so far, as you pointed out,
we do not have any solutions for this. Altouhgh each SP believes that these TE links
are not a SRLG because these are belonging to a different SP, fibers accommodating
such TE links may be laid in the same duct otherwise the same tunnel under the road.
Our conclusion is that the level of SRLG is another issue.....(one issue is indeed how
to assign SRLG IDs to each SP)

Regards,

tomo




At 18:41 04/07/27, Adrian Farrel wrote: 
Hi,

A nice short draft. Congratulations!

Also, it is very good to see SPs bringing forward requirement drafts. Thank you.

I understand the point you are making in the draft, but I am not sure it is specific to GMPLS and switching capabilities.

For example, suppose we modify the bandwidth in your MPLS example network to:

+----+ +----+ | +----+ +----+
| A1 |----//----| A3 |---------| B1 |----//----| B3 |
+----+ 10G +----+ 10G +----+ 2.5G +----+
| | | | |
=2.5G =10G | =2.5G =2.5G
| | | | |
+----+ +----+ | +----+ +----+
| A2 |----//----| A4 |---------| B2 |----//----| B4 |
+----+ 2.5G +----+ 10G +----+ 10G +----+
|
MPLS AS 1 | MPLS AS 2
Now, set up a 10G service from A1 to B4.
AS 1 is going to select the ASBR pair A3/B1 as the shortest path out of AS 1.
But B1 will fail the setup.
We must rely on crankback or a wider view (PCE, TE aggregation, etc.) in order to be successful.


I think what your draft points out, however, is that the complexity for GMPLS is increased considerably (perhaps to a power of three). It is further worth noting that if we added some speculative future routing constraints (such as source-based lambda selection, optical impairments, etc.) the problem would get even more complex.

Essentially, however, the problem is the same: IP route aggregation is not sufficient to enable inter-domain TE and some other solution is needed. Your proposal for EGP extensions to general reachability information is certainly one option.

The concern that I have heard voiced is that there may be significant churn in this information, and this would result in a significant amount of aggregation computation by the ASBRs. My view is that:
- in a non-PSC GMPLS network the rate of change is
not likely to be significant
- we should, in any case, specify damping of computation
and updates if we proceed with this approach.
But I would be interested to hear this debated further especially by the EGP experts.



Your point about SRLGs is very valid. Currently, however, (as I understand it) we don't have a satisfactory encoding for SRLG IDs to allow an ID to be globally unique *and* to allow an SRLG to span ASs.

Thanks,
Adrian

******************************
Nagao Ogino, Dr. Eng.
KDDI R&D Laboratories Inc.
2-1-15 Ohara, Kamifukuoka-shi,
Saitama 356-8502 Japan
TEL +81 49 278 7860
FAX +81 49 278 7811
E-mail ogino@kddilabs.jp
******************************