Fair enough. As the TSC, we try to avoid forcing collaboration where the developers decide that it's not the right idea from a technical perspective.

I'd still love to see some intermediate YANG model which is for doing MPLS-TP control that was independent of the two use cases and then have specific models built on top of it (for the given applications) or below it (to provide for control of more devices, but if there's some reason that's not possible. I wont' force it.

--Colin

On Fri, Apr 1, 2016 at 6:59 AM, 박세형 <labry@...> wrote:

Hi,

 

Thank you for agreeing with me, Weiqiang Cheng.

I understand that MPLS-TP YANG models are in your scope.

But I am guessing our MPLS-TP model is fundamentally different from MPLS-TP Solution team’s YANG model because our YANG models are designed to incorporate multiple protocols (MPLS-TP, OTN, and ROADM).

For that reason, we have extracted some sets of common features among those protocols and created a parent YANG model called “transport YANG model.”

Then, our MPLS-TP YANG model and OTN YANG model extend the transport model and add more attributes as necessary.

As much as I want to collaborate with MPLS-TP Solution team (China Mobile), I think we don't have the common grounds to work on for now.

However, I think we can have a very positive influence on each other as we share several keywords and technologies.

Regards,

Justin

 

From: Weiqiang Cheng [mailto:chengweiqiang@...]
Sent: Friday, April 01, 2016 4:43 PM
To: 박세형; 'Colin Dixon'
Cc: 박수명; 임창규; 신지수; 'Pradeeban Kathiravelu'; 'George Zhao'; 'Wang Lei'; 'lihan'; 'zhangtingtingyjy'; 'liqin'; project-proposals@...; tsc@...
Subject: 答复: Rquesting An MPLS-TP Creation Review

 

Hi，

I agree with Justin. It is better to have separate projects.

One point I want to clarify is that the MPLS-TP YANG models for NBI is also in our scope.

 

B.R.

Weiqiang Cheng

 

发件人: 박세형 [mailto:labry@...]
发送时间: 2016年4月1日 10:47
收件人: Colin Dixon; Weiqiang Cheng
抄送: 박수명; 임창규; 신지수; Pradeeban Kathiravelu; George Zhao; Wang Lei; lihan; zhangtingtingyjy; liqin; project-proposals@...; tsc@...
主题: Rquesting An MPLS-TP Creation Review

 

Hi,

 

I’m cc’ing tsc mailing list to get a creation review.

 

We carefully searched for a common ground to collaborate with China Mobile.

But as far as I understand, it seems like our common ground is limited to the keyword "MPLS-TP" only.

Our goals and models are different from each other.

China Mobile focuses on applying the MPLS-TP (PTN) solution in their wireless backhaul network using OpenFlowTTP and Qx.

We are working on bringing Optical Transport Network (OTN) and MPLS-TP in core networks together by integrating legacy EMS/NMS platforms through common YANG models.

That means we are not currently working on OpenFlowTTP and Qx, nor have any concrete plans to work on them in the near future.

 

Originally, we intended to create a project called "Transport SDN" solution to incorporate OTN, MPLS-TP, and ROADM protocols.

To think big but start small, we ended up choosing the keyword MPLS-TP instead of T-SDN.

However, now that we have a keyword conflict with China Mobile, we are willing to change our keyword from MPLS-TP to Transport SDN solution.

 

In the future, I think we can benefit and collaborate with MPLS-TP Solution team on OpenFlowTTP, Qx, and Southbound drivers.

Also, we hope China Mobile could reference our project when they want to start working on OTN.

For now, we think it would be wiser to have separate projects.

We will look forward hearing your opinions on this.

 

Regards,

 

Justin

 

From: Colin Dixon [mailto:colin@...]
Sent: Thursday, March 31, 2016 11:39 PM
To: Weiqiang Cheng
Cc: 박세형; CANPTM8GUWD=vpATEgxkyHtjUyC8zoSCG-3=zmTyBDLETTOiunw@...; 박수명; 임창규; 신지수; Pradeeban Kathiravelu; George Zhao; Wang Lei; lihan; zhangtingtingyjy; liqin; project-proposals@...
Subject: Re: 答复: [OpenDaylight TSC] MPLS-TP creation review

 

I'm cc'ing project-proposals just so this gets logged somewhere.

It sounds like common areas here would be the MPLS-TP TTP and the MPLS-TP support for OF-CONFIG. Beyond that, it sounds like we have two different applications that would use those APIs. Is that right?

 

--Colin

 

On Thu, Mar 31, 2016 at 5:33 AM, Weiqiang Cheng <chengweiqiang@China Mobile.com> wrote:

Dear Justin,

Thank you very much for your information about the MPLS-TP service project.

 

Base on your description, I think the requirements are different. You hope to base on the OTN baseline to support MPLS-TP.

 

China Mobile hope to use MPLS-TP as the basic solution for Packet transport network(PTN). We hope to use ODL to implement both super controller and domain controller. I also hope to details the overview of MPLS-TP solution.

 

The contribution will include:

- Yang Model with restconf solution for the interface between super controller and domain controller.

- Yang Mode and intent interface for the interface between Application and controller

- Openflow TTP with MPLS-TP extension

- OF-config with MPLS-TP extension

 

Currently, we have setup a test environments based on our prototype. More than 7 PTN equipment vendors including Huawei, Alcatel-Lucent, ZTE, Fiberhome, GWTT, Raisecom, Shanshui Optoelectronic… are involved in the test environments. And we also developed the MPLS-TP Openflow TTP which currently has been supported by several chip vendors such as Broadcom, PMC, Centec, and we are running the interoperation test in parallel.

 

The test result shows that the models and mechanism work well and China mobile planed to do field trial recently.

 

The willing committers are also increasing. The developers from Ericsson, GWTT, Raisecom, ZTE, Broadcom are planning to contribute to the project.

 

B.R.

Weiqiang Cheng

 

发件人: 박세형 [mailto:labry@...]
发送时间: 2016年3月29日 9:28
收件人: CANPTM8GUWD=vpATEgxkyHtjUyC8zoSCG-3=zmTyBDLETTOiunw@...
抄送: Colin Dixon; 박수명; 임창규; 신지수; 박세형; kk.pradeeban@...; chengweiqiang@China Mobile.com; George.Y.Zhao@...
主题: [OpenDaylight TSC] MPLS-TP creation review

 

Greetings,

 

This is Justin from ETRI.

Thank you for giving us a chance to do a creation review as an email Q&A.

Most of the details of our project can be found at the following link.

https://wiki.opendaylight.org/view/Project_Proposals:MPLS-TP_Service

 

However, let me briefly introduce you to the status of our solution.

MPLS-TP Service Team has put intensive efforts into this project and has almost finished the prototype controller and the plugins.

We received our requirements from local carriers (SKT and KT) and have successfully demonstrated the creation/deletion of multiple services over two different vendors (Woorinet and ETRI devices).

The Telefield (a local vendor) MPLS-TP devices are expected to be added to our compatible vendor's list after tiny adjustments.

We are willing to share our YANG models and controller source code, as well as one of Southbound plugins to ODL community.

 

<Figure 1: T-SDN Platform based on ODL>

 

Regarding collaboration, we want to know more about MPLS-TP Solution Team’s project. We want to know which controller uses ODL, super, domain, or both? We also want to know if your super controller is equivalent to our super controller in the figure 1. I've looked at the architecture of MPLS-TP managers from MPLS-TP Solution Team, but the architecture seems very different from ours.

MPLS-TP Solution Team (China Mobile) seems like their main focus in on southbound drivers including Qx and OpenFlowTTP, whereas we are focusing on the northbound interface and using YANG models as a communication mechanism between the managers and the plugins. Currently, we are working on brownfield networks and bringing legacy devices into SDN realm, but we are not considering Qx nor OpenFlowTTP.

 

<Figure 2: T-SDN Platform>

 

To keep you informed, we are also working on OTN (Optical Transport Networking) over ODL, which will eventually evolve into a POTN solution, refer to figure 2.

The prototype OTN controller is almost ready, but we are still working on the internal details of OTN Southbound plugins and its counterpart EMS.

We are confident in saying that our prototypes (MPLS-TP and OTN controllers and plugins) prove that our approach is very promising. ODL has saved us enormous time and effort in bringing all these together.

We would welcome anyone who is willing to make contributions to our project or learn from us.

If you have any questions related to the project, I'd be glad to answer them..

 

•       Project Dependencies

–      Controller

–      ODLparent

–      Yangtool

•       Deliverable (Boron offset 1)

–      Yang Models (MPLS-TP)

–      MPLS-TP Controller (Serivce, Inventory, and Topology Managers)

–      Southbound Plugin(s)

•       Willing Committers (10+)

–      ETRI: Justin Park, Jisoo Shin, Jin Won Kang, Chang-Gyu Lim, Soomyung Park (5)

–      SKT:  Jongyoon Shin, Junhee Lee (2)

–      Woorinet: Moonho Yoo (1)

–      Mobigen: Moonkuk Park, Hyungcheol Park (2)

•       Planning to migrate to Beryllium and beyond

•       Project Lead – Justin Park

•       Document Contact – Justin Park

•       Test Contact – Justin Park

 

Regards,

 

Justin