2026 public meeting minutes

available at TransportPCE meetings
https://lf-opendaylight.atlassian.net/wiki/display/ODL/TransportPCE+meetings

• Gilles succeeded in releasing Cr-SR1 this week

• Next step is the migration of master branch to Manganese. Already started for transportpce-models and transportpce. Stoped for transportpce due to api migration  on mdsal upstream (see test-comon)=> not the time to continue…

on transportpce-models side:

• Change related to Cr-SR1 Release.

  • Due to the introduction of models OR 17.1 in Cr, Gilles needed to split some common models for 13.1.1. and 17.1.0 on new separate bundles (common-9.1.0, device-8.0.0 and device-12.1.0).

  • Pre-release 2 OpenConfig models as well as OR R1.2.1 models have been removed  (since implementation had already been removed from transportpce)

• For OpenROADM 17.1 models, an issue was observed on otn-common-type (revision date was not updated even if the file itself was updated). Transportpce change 123175 allows solving this. The revision date of the corresponding file has been changed. But if the file common-attributes which imports the otn-common-type was updated for the import, its own revision was not updated. This could lead to inconsistency. Thus, Bala recommends to update org-openroadm-common-attributes revision.

• For Manganese we complement OC models with 5.6 models and remove OR R1.2.1 models

on transportpce side:

• Support for OC 2.0 has been backported to Cr-SR1

• Lots of relation chains are more or less in the same status as last week and are ready to be merged.

  • • topic:fixed-patchset

    • topic:datastore-read-hardening

    • topic:roadm-connection-map

    • topic:preserve-interupt-status

    • topic:tapi-fix-initial-services

    • topic:tapi-npe-guard + topic:pce-tapi-service

• Smartoptics pushed new changes in the relation chains associated with the 2 topic “Topology-builder-convergence” and “Fixed patchset”

• “Add TAPI SBI controller configurations RPCs” has been pushed by Fujitsu on behalf of NTT. This change is in the scope of developing the TAPI SBI interface. It provides models and code to configure a connection to a subtended controller.

• Gilles pushed some changes associated with the migration to Manganese. As Robert changed some mdsal APIs in Manganese, test-common fails. Gilles will probably need some support from the community to go further.

• “Add second-step-hybrid-pc-result notif to tpce-pce.yang models” introduces a new notification to handle cross-domain service creation.

• “Bump Lighty Node version” provide an update of Lightynode to support OpenConfig 5.6.

• Gilles succeeded in releasing Ti-SR3 and V-SR2 in spite of significant disruptions to the ODL infrastructure (lots of issues and disconnections with jenkins and nexus servers)

• Gilles backported some changes on the Chromium branch which is in a good shape for Cr-SR1

  • Some adaptations on transportpce-models are necessary to avoid the model duplications between several bundles due to the addition of new 17.1 models

  • backport of integration of OpenConfig 2.0 models will be applied for Cr-SR1

• Next step will be the migration of master branch to Manganese

• Changes related to the preparation of releases (Ti-SR2, V-SR2 and Cr-SR1). Very few changes backported to these branches, as there was too big a gap with the master branch

• Lots of relation chains are more or less in the same status as last week and are ready to be merged.

  • • topic:fixed-patchset

    • topic:datastore-read-hardening

    • topic:roadm-connection-map

  • Some of them have been reviewed by Olivier

    • topic:preserve-interupt-status

    • topic:tapi-fix-initial-services

    • topic:tapi-npe-guard + topic:pce-tapi-service

• “Pce Orchestration Step1” has been pushed as WIP by Olivier. This is a first draft to pave the way towards Cross-domain service creation. It introduces variables that need to be shared/passed between different classes of the PCE when path computation is performed in several steps (involving first OR PCE, and then TAPI PCE) according to the domains that the service crosses. Some of the method signatures have been adapted to handle partial path impairment calculation (mainly in PceGraph and PostAlgoPathValidator). This change is not ready for review, especially as it introduces some new empty methods, for which code will be provided at a later step.

• “Bumb lightynode version” adopts the 23.1.1.1 version of lightynode (aligned with V-SR1). It also adds a new script to  automate the installation of multiple Lightynode device simulator with support for different versions. Lightynode will now be specific per model.

• “TAPI-refactor ROADM NEP population into factory” typo has been corrected.

• 3 releases Ti-SR3, V-SR2, and Cr-SR1 need to be done as soon as possible

• Additionally, we should prepare migration to Manganese which is planned in September. OC models 5.6 could be integrated in M.

On transportpce-models:

• Changes related to the bump to Ti SR3. Robert proposed different changes to clean up our repo (notably related to tapi and openconfig where he updated some dependencies).

• From R 7.1 to 13.1 OpenROADM used some models that have been updated in R 6.1.1. which can lead to issues. The change “Split out (common, device) -6.1.1” was provided to ease integration

• Next step will be to do the same on Vanadium branch.

On transportpce:

• Changes related to the bump to Ti-SR3.

• “Add support for OR 13.1.1 and 17.1.0 models”

• The first relation chain of Olivier on tapi-sbi.

• Christophe also adjust the code to support the removal of OpenROADM 1.2.1 models.

• topic:fixed-patchset

  • Instead of parsing the XPDR number from transponder name it is done from an augmentation. At the bottom, several changes concern the refactoring of Tapi Links. This relation chain could be merged

• topic:datastore-read-hardening

  •  improves the management of the DataStore. The relation chain could be merged.

• topic:roadm-connection-map

  • adds support for ROADM connection map, and could be merged

• topic:preserve-interupt-status

  •  allows to reset the interrupt status, and can be merged

• topic:tapi-fix-initial-services

  • allows to solve some racy conditions that appear during conversion from OR to TAPI contexts

• topic:tapi-npe-guard + topic:pce-tapi-service

• Olivier’s relation chain

  • “Adjust OSNR calculation in PostAlgoValidator” & “Adjust PostAlgoValidator.getSpectrumAssignement », provide some refinement to prepare next step, introducing cross-domain service creation where intermediate impairment calculation results need to be considered.

  • “Change "broken-down" to "cross-domain" service” allows to use more relevant terms for this parameter in transportpce-pathDescription.yang

  • "Refine skeleton for TAPI-SBI feature” focuses on adjusting TapiSbiProvider, adding SbiRendererRPCImpl to register RPC TapiSbiServiceDelete  and TapiSbiServiceImplementationRequest and adding TapiSbiRendererOperation Interface & Implementation.

• Christophe is refactoring lighty simulators. He Tries to push an approach where we have a specific jar for each release; to solve the issue we have to support several OpenConfig models

• TransportPCE paper has been accepted for publication in the Journal of Optical Communications and Networking ! Thanks to all contributors.

• Nothing merged on tpce/models.

• Relation Chain on TAPI of Olivier reviewed by Joakim merged.

• “Remove OpenConfig 1.9.0 models” and “Remove OpenROADM 1.2.1 models” on tpce/models to remove models not supported in next future release.

• Relation chain on TAPI-SBI feature still in review and should be merged in the coming days.

• “Add support for OpenROADM 13.1.1 and 17.1.0 models” is ready for review

• “Remove OpenROADM 1.2.1 models” : Bala rebased it on top of “Add support for OpenROADM 13.1.1 and 17.1.0 models” to ease the integration and limit the merge conflicts.

• Relation Chain from Joakim solving some bugs (Top change : Fix racy future.isDone() with blocking timed reads) has been reworked by Joakim.

• Relation chain on TAPI from Smartoptics (some changes still not passing the gates) will need to be rebased.

• Question received on notification propagation. Still under study.

• TransportPCE paper seems having reached pre-production process meaning it has been accepted! Thanks to all contributors.

• 3 changes associated with OpenConfig 5.6 have been merged on tpce/models : we can now use OC R5.6 for development in tpce.

• Nothing merged on tpce.

• “Remove OpenConfig 1.9.0 models” and “Remove OpenROADM 1.2.1 models” on tpce/models to remove models not supported in next future release.

• Relation Chain on TAPI of Olivier reviewed by Joakim and split into several changes could be merged in short term.

• Relation chain on TAPI-SBI feature : Olivier addressed most of Christophe comments, and Christophe double-checked remaining comments making its own modifications. He will push modified changes (removing unnecessary files) and the relation chain could be merged.

• “Add support for OpenROADM 13.1.1 and 17.1.0 models” was failing because of race conditions, but Bala succeeded making it pass the gate. Bala addressed Joakim’s comments. This change is ready for review.

• “Remove OpenROADM 1.2.1 models” is a huge change as many modules were relying on R1.2.1 features. Christophe would like to merge it before the change of AT&T introducing R13.1.1 and R 17.1.0, because making it the opposite way would probably be very complex.

• Relation Chain from Joakim solving some bugs (Top change : Fix racy future.isDone() with blocking timed reads) has been reworked by Joakim.

• Relation chain on TAPI from Smartoptics (some changes still not passing the gates) will need to be rebased.

• Christophe started working on the evolution of Lightynode to support OC 5.6.0. He has the feeling we could support concurrently OC 5.6. and OC 2.

• Relation chain on MC-capabilities

• Change related to OpenConfig 2.0 (Migration of the code from OC Pre-release1 to Release 2.0)

• The bottom of the relation chain of Jimmy associated with Tapi Link has been reviewed by Gilles.

• “Add support for OpenROADM 13.1.1 and 17.1.1 models”: Bala changed OpenROADM model imports in port-mapping, introduced new enums, updated revision statements and version types wherever necessary. A tapi test fails probably due to a race condition (observed also by Joakim in other conditions, the test fails from time to time)

• OpenConfig: Christophe started to work on removing pre-release 1 models, and introduced R5.6 models

• Christophe also started to work on removing OpenROADM R 1.2.1 models and dependencies.

• Joakim started to review the relation chain of Olivier associated with TAPI topology Listener, transversal routing over topologies, and switching to UNIDIR links. He has no opposition with the content of these changes but started to rework the relation chain. The relation chain is restructured, but the diff shall correspond to the diff of the original relation chain.

• Relation chain associated with T-API SBI feature is independent from others and could be merged after the comments of Christophe have been addressed.

• Bala mentioned we could have an interest in updating the portMapping according to some update notifications (this is already the case for port and interface changes), when the list of Xponder changes.

• nothing

• Few changes associated with the preparation of Ti SR3.

• Relation chain associated with MC-capabilities could be merged. Christophe reviewed the changes at the bottom of the stack, but some additional review is still required for the 5 changes on top of the relation-chain.

• Olivier’s relation chain about SBI topology up to Switch to UNIDIR Links could be merged in a second step.

• Relation chain of Smartoptics associated with link refactoring: could be merged in a third step.

• Relation of NTT/Fujitsu on OC R2.0 passes the gate and could be merged.

• Relation chain of tapi-sbi features is quite independent from other changes and could be merged after Christophe’s comments have been addressed.

• Bala was able to compile his preliminary work on implementation of OpenROADM 17.1 Device models, but he experiences some issues with PortMapping710Tests. First guess of Joakim is that 1 mock would need to be updated.

• JOCN Paper: all contributors should have received a mail mentioning that JOCN paper about Tpce needs major revisions. Olivier mentions he will need help to provide some performance parameters required by second reviewer (This reviewer comment is almost at the end of the mail). The performance may be extrapolated from LOGs associated with service creation (At WDM and OTN level). If some of the TAPI functional tests would allow generating LOGs, they are based on simulators. Help his required from someone having equipment in his labs so that we can generate LOGs with real equipment. For what concerns scalability, Olivier asked Smartoptics to inquire on their ability to provide results from scalability tests.

• A discussion with ODL community was initiated by Smartopics about Cluster applications. Smartoptics try to have one active TransportPCE instance and a standby one running as secondary. They observed 2 issues : Standby instance is performing some connections to nodes which is not expected. 2^nd issue is related with storing information in Operational Data Store which can not be replicated in the Standby Instance

• nothing

• Vanadium and Scandium dependencies have been updated to respectively SR2 and SR3 on TPCE/models.

• “Propagate node deletion cleanup features”.

• 3 changes solving bugs in Netconf Listener and NetconfTopologyListener.

• “Add a set of OpenConfig 5.6.0 models” introduces OpenConfig 5.6.0  models associated with OC platform and terminal-device (minimum required to run tpce).

• “Add a set of OpenConfig 5.6.0 models” introduces OpenConfig 5.6.0  models associated with OC system, required for system. The commit message will be adjusted to avoid confusion with previous change.

• “rework openconfig models 5.6.0” : provides some reworking (renaming files, adding revision dates, solving compilation issue associated with deprecations, changing revision of oc-extensions in oc-alarm-types…) of the models so that they are better supported by OpenDaylight and can be compiled. Christophe hopes that both R2.0 and 5.6 can be supported concurrently in the same instance of tpce.

• “Skip xponder mc-capability in spectrum assignment” : Bala has tested it partially (using config files that they have) and this change works so that according to AT&T, this change could be merged.

• Relation chain of Olivier concerning refactoring of Tapi links to Unidir : Olivier is waiting for potential review/comments from Jonas.

• 2 relation chains proposed by SmartOptics on Tapi are still in standby.

• Tapi-Sbi relation chain has been rebased and merge-conflicts associated with the introduction new 17.1 OpenRoadm models have been solved but the dependencies in the Tapisbi Pom of tapi-sbi feature also needed to be updated.

• “Add support for new OpenRoadm models” provides some updates in the renderer and the OLM to support new 17.1 models introduced recently. This change still needs a bit of reworking.

• “Support for OpenConfig R 2.0” provided by Roschan updates lighty toward lighty 22.1.0.7 which does not support OC R1.9. This leads to a failure of the tests associated with OC R1.9.0 and makes the gate failing.

• “Functional tests for oc200” provided by Roschan complements functional tests for OpenConfig R 2.0.

• nothing

• Changes concerning the maintenance of the repo and related to GHA (Github actions)

• Relation chain of Smartoptics associated with MC capabilities: the documentation has been finalized. AT&T tested it and got it working on real devices that were causing issues. AT&T will need to make complementary tests.

• “Reorganize docs into topics subfolders” Joakim reorganized the docs into topic. Some other changes related to the same topic will be pushed in a close future.

• Link refactoring relation-chain of Smartoptics:  no status change since last week.

• Relation-chain of Olivier (“Regenerate Ref test Topo with Unidir links”, “Switch to UNIDIR links in Topo conversion”, “Fix for transversal routing over tapi topos”, “SBI tapi topology Listener”, “Add SBI topo file to sample_configs”) related to SBI topology handling: the relation chain (the 5 changes of the bottom) passes the gate and is ready for review.
  Joakim will review it tomorrow. Could be merged after.

• “Abort service creation if device rendering fails” from Jonas could be also merged.

• We are very close to release Cr. TransportPCE artifacts have been released today

• gNMI connector will be included in the Chromium distribution.

on transportpce-models repo:

• OpenROADM models 17.1 were introduced. Openroadm-otn-common types has a wrong revision date, and as a result is imported with a wrong revision date. This introduces some issues. Gilles had to make some corrections on different files to import the right version of the model (not R10.0 but R 17.1). Another model has the same issue : org-openroadm-common-attributes, but doesn’t seem to lead to any integration problems, so the revision was not corrected.

on transportpce repo:

• “Bump transportpce-models to 24.0.1” allows tpce to integrate R17.1 models.

• “Use operational mode catalog of models 17.1.0” introduces a corrected version of the Json specification files.

• “Integrate the openroadm models of version 17.1.0” provides the adaptation to support OR service and network models 17.1.

• Changes related to the creation of the stable/chromium branch for the release

• Joakim has proposed a Relation chain on mc-capability issues. He has provided a “quick” fix to the issue experienced by Bala in “Skip Xponder MC capability in spectrum assignment”. Joakim asked if AT&T would have the capability to test the fix for R7.1 equipment.

• “Fix XPDR mc-capability spectrum validation for 7.1”,  “Fix silent observer skip in McCapability”, “Add min/max-edge-freq to portmapping YANG model”, “Save MC 7.1 min/max-edge-freq to PortMapping”, “Introduce FrequencyRange for 7.1 mc-capabilities”, “Filter spectrum by MC capability frequency range” & “Document PCE spectrum assignment via MC capabilities” are cleaning up or adjusting functions/methods to avoid having functions called in the case of Xponders whether they are not really relevant in that case

• “Support for OpenConfig 2.0” has been pushed by Roshan. This is a rebase of a previous change which has been completed with functional tests

• “Abort service creation if device rendering fails“ solves a bug introduced with OpenConfig support, while rendering service on devices and trying to set the OLM. The bug was identified in Smartoptics labs.

• There was no objection to the proposal of Gilles to remove the support of OpenROADM R 1.2.1 and OpenConfig R 1.9 (Pre-release-1).

• Gilles is still integrating some changes of interest for Cr.

• Changes related to the CI:  “ci: add SonarCloud GHA workflow for weekly scans” & “Update docs configuration for CR”.

• “Bump project version to 25.0.0 Snapshot”

• The 2 AT&T changes about operational modes have been merged: “Add support for proprietary operational modes” and “Bug fix in extracting xpdr operational modes”

• Gilles is currently considering the possibility to integrate “Add OpenROADM 17.1.0” pending change provided by Christophe to Cr. He is also working on adaptations on network and service models to support in tpce these new models and close to finish it.

• Olivier almost solved merge conflicts on his relation chain, but still has TAPI functional test 4 failing (at tapi feature deinstallation process). This seems not connected to code change (issue that has already been seen on the gate) but will need to be solved. After this and when the changes pass the gate, he will proceed to a clean-up removing residual trace of merge conflict that have been commented.

• “Fix mc capabilities in Xpdr advertizement” was proposed by Joakim to address the issue Bala encounters with the default value of 1 being set for Max-slot if no max slot is announced. It will however probably be abandoned. An alternative solution that is not to consider the default value of 1 for transponder will be investigated and proposed by Smartoptics.

• Gilles mentions that in OpenROADM 17.1 models, org-openroadm-otn-common-types the revision is not correct (2021-09-24 in place of 2025-01-10)

• Integration of different changes for integration in Cr still ongoing…

• End of Chromium migration (“Modernize DataTreeChangeListener dispatches”…)

• Pending changes related to the CI (update of upstreams, docs, maven in tox…)

• Changes related to the consolidation of Topo Uuid use

• OpenConfig end2end functional tests (“Add E2E functional tests for OC services creation”) was fixed and merged

• Joakim 11 bottom changes of the main relation chain concerning Tapi Link Refactoring (reviewed by Olivier) have been merged

• Top of the relation chain of Smartoptics concerning Tapi Link Refactoring

• Other relation chain of Smartoptics concerning Tapi Refactoring

• “Improve DataStore Transaction error handling” could be merged

• “Fix racy Future.isDone() with blocking timed reads” could be merged

• ”Bug fix in extracting xpdr operational-mode” introduced a regression on OpenConfig side. Gilles has modified it so to solve this regression

• “Add support for proprietary operational mode”: need a Jira Ticket

• Olivier has rebased his relation chain. He asks for temporizing a bit the merge of the rest of the relation chain of Smartoptics, to make sure we will be able to rebase this relation chain on the remaining change of Smartoptics relation chain when they will be merged

• “Fix mc-capabilities in XPDR advertisements” is a new change provided by Bala. It tries to solve some issues we have with used default values for MC capabilities when not all the values are populated.

• Still in progress on the path to Cr release. All other MRI projects are ready for Cr. Gilles stated during the TSC that he needs time to integrate all needed changes.

• Cr migration is done for tpce models. For tpce “Clean up Maven project files” and another change, still not pushed (addressing deprecation of some upstream methods), need to be merged.

• Gilles mentions that the relation chain of Joakim on Tapi Links and the relation chain on the use of Tapi Uuid of Olivier could be rebased to get visibility, but Gilles proposes to wait until Monday evening (after he has merged 2 additional changes) for final rebase.

• Joakim is pursuing testing his relation-chain regenerating the Tapi topology file used in PCE Unit tests, using the specific python test built for this purpose.

• “Add support for proprietary operational modes”: Renderer test was failing because the order of the interface creation in the test was not the right one. Bala corrected this as well another failure cause and the tests now pass. He also modified the commit message. Bala will solve the merge-conflict and the change will be ready for review.

• Bala is working on finalizing the JOCN paper focusing on OpenROADM and hope finishing it next week. UTD is also planning to finish his paper next week.

• OpenDaylight distribution for V-SR1 was released on April 22.

• Gilles proposed Joakim, to add him to the team of committers of the project. Joakim will provide an answer on whether he is OK with that, so that we can launch the process if this is the case.

• “ci enable Gerrit voting and add all tox envs” was reverted.

• "Add openROADM 17.1.0 models” has been pushed by Christophe. This allows to use device and common models of OpenROADM R17.1, while still relying on 13.1.1 models for Network and Services. Gilles identified an issue with this dual dependency to 13.1.1 and 17.1 models. We may need to migrate the entire stack of the models.

• Migration to Cr

  • it seems all other projects have migrated to Cr.

  • On transportpce-models side:

    • master branch is already aligned to Cr.

    • “Add openROADM 17.1 Models”: we have some issues that come from the OpenRoadm catalog that we load at the initialization of transportPCE (servicehandler bundle). The Catalog json files needs to be updated to be aligned with 17.1 models. Unfortunately, api-doc files need to be corrected, and probably the PostAlgoValidator code will also need to be updated to fit with new data structure. Thus, Gilles proposed to postpone the integration of 17.1 models in Cr-SR1 instead of Cr-GA.

  • on transportpce side:

    • “Bump upstream dependencies to V-SR1”,

    • “set the scale for Decimal64 based values” makes some corrections on fraction-digit for Decimal64, since Decimal64 implementation of Yangtools has evolved to integrate the fraction-digit field coming from yang models. However, Robert is also investigating some additional implementations in Yangtools associated with the treatment of floats

    • “Resolve IntervalCollection unit tests” & “Resolve ValidInput unit test”: Joakim repaired some test that were disabled by Gilles to pass the gate. Failures were associated with the new implementation of Decimal64 in Yangtools and some issues in the Mocks.

    • “test : strengthen ValidInput and ValidSlot tests” makes sure that all mocks are called in associated tests.

• “CI : update maven version used by tox-verify job” reestablishes the gate normal operation.

• Gilles will merge the 3 remaining changes of Olivier associated with topo Uuid handling, as well as the change proposed by Christophe for OpenConfig.

• Joakim relation chain can be reviewed up to “Topology based endpoints resolution in OR listener”. Joakim will check that regenerating the topology with the python test file provided by Olivier, does not impact the tests relying on it, which will demonstrate that the relation chain does not introduce any regression.

• “Hybrid path computation” is the first of a suite of changes that will be dedicated to handling of service request that span across several domains. It introduces some basic function required for PCE operation in “hybrid” mode. It helps through comments,  understanding how the PCE algo will be modified in the following changes to support this new mode of operation.

• “launch-tests : improve invalid test name errors” & ”launch tests : add -stop at option and usage hello” provide function that help to launch test and provide information about their location and execution. These changes are ready to be merged.

• Smartoptics relation-chain:

  • “Fail fast if TAPI service already exists”, “TAPI - apply OpenROADM topology updates to NEPs” have been reviewed by Orange (Gilles and/or Olivier)

  • Joakim changed to Private status the other changes of the relation chain. Gilles highlighted that

• Olivier just updated his relation chain to solve some pylint issue. We however have some issues with the Daexim dependency. Daexim code would need to be integrated into NETCONF so that it is integrated in an upstream dependency. Robert is working on the issue but need a formal approval form Daexim developers to do so, most of them not working anymore on the project

• V-SR1 shall be released very soon.

• Some discussion ongoing regarding the management of the CI in the context of moving from Jenkins to Github actions.

• on transportpce-models, “Refresh Upstream”: preparation of Cr.

• “Bump lfit/releng-reusable-workflows/.github/workflows/gerrit-required-info-yaml-verify.yaml”: Change for migrating to Github actions.

• “CI: enable Gerrit voting and all tox envs” pushed by LFN has been merged to show that it was breaking the CI and that Github actions was not compatible with our functional test that requires not only python, but also maven. The change need to be reverted (revert not merged yet).

• “Revert CI: enable Gerrit voting and all tox envs” : to be merged if we want to re-establish normal operation of the CI.

• “Add E2E functional tests for OC service creation” was failing (only on the gate) due to the parallelism of tests and the fact that the ROADM simulators were needed through different tests running at the same time. Gilles changed the ROADM type in the simulators to solve this conflict.

• Olivier’s relation chain about the usage of Tapi Topo Uuid has been rebased by Gilles and is almost ready to be merged, but part of the changes are not compatible with some of the changes performed by Smartoptics on LinkId.  Gilles proposed to merge this relation chain so that Smartoptics can rebase his relation chain on top of it.

• “Build Skeleton for TAPI-SBI feature”: Gilles made some corrections so that this feature can be activated.

• “Speedup TAPI device change notifications tests” : in this change Joakim proposes to optimize the running of Tapi functional test3, checking when some changes have been applied by the controller rather than relying on time constant. As a result, it allows reducing the time needed to perform the tests. Gilles suggests building generic functions that can be reused through the different tapi functional tests.

• “Add support for proprietary operational mode” : Gilles did not have time to review but will do it in the coming weeks.

• Smartoptics pushed some new changes completing the relation chain regarding tapi refactoring (mainly concerns links). Olivier has some concerns with change 122254 “TAPI link remove dependency on hardcoded ROADM" that he explained in his comment of this change : TAPI PCE needs to route over links and nodes, which might be affected by this relation-chain and we can’t see the real impact of these changes on the behavior of the PCE because the Unit tests of the TAPI PCE runs over a preloaded topology. Thus, to see the real impact of these changes in this relation chain, we would need to regenerate the topology used for some of the unit tests (pce -> src/test/resources/topologyData/refTopoTapiFull.xml) after all the changes have been applied, and check that Junit test based on it still pass… Olivier will provide this functional test to Joakim.

• Olivier’s relation chain regarding transversal routing over different tapi domains provides corrections in the PCE and Tapi topology. The 2 changes on the top are related to Daexim feature activation, the last providing the test which was used to debug and make all corrections provided in the other changes of the relation chain. These 2 top changes shall not be merged. The rest of the relation chain will be reworked so that all changes pass the gate.

• Next week meeting is canceled

• Ti-SR2 has been approved but the distribution has still not been released

• Part of the Smartoptics refactoring activity on tapi has been merged

• A change to solve Github CI from Christophe, “Update pre-commit config file” (concerns AutoPep 8 and versions)

• Update concerning the CI to be in line with other project of the community.

• The augmentation of the documentations with Diagrams and PlantUml from Smartoptics “Add Plant uml sequence diagrams in documentation”.

• Change regarding the consolidation of TopoUuid have been rebased by Gilles.

• Smartoptics pushed some new changes completing the relation chain regarding tapi refactoring (mainly concerns links).

• Olivier’s relation chain regarding transversal routing over different tapi domains provides corrections in the PCE and Tapi topology. The 2 changes on the top are related to Daexim feature activation, the last providing the test which was used to debug and make all corrections provided in the other changes of the relation chain (The 2 top changes were initially at the bottom of the relation chain, but have been moved on top because we can’t merge them in the current situation (Daexim handled as a MSI project). Gilles mentions it is probably the right time now to start developments on the SBI, so that we have a different way to populate a SBI topology in tpce. This should start with the development of a server that mimics the TAPI-NBI of a NMS.

• Support for OpenConfig 2.0 does not pass the gate because the latest models available in transportpce/models are not in line with this change.

• “add E2E functional tests for OC Service creation” should be merged before the migration to Cr but we need to understand why it fails on the gate.

• Olivier briefly introduced his expectation for the review of JOCN paper dedicated to tpce in the OpenROADM special edition..

• Ti-SR2: a release candidate is underway for approval by TSC.

• V-SR1: still in a pending status.

• “TAPI - OpenROADM TP to TAPI ONEP spectrum mapping”, “Refactor RDM link conversion to use topology”, and “Add OpenROADM link tp resolution support” have been merged

• "Add openROADM 17.1.0 models” has been pushed by Christophe. This allows to use device and common models of OpenROADM R17.1, while still relying on 13.1.1 models for Network and Services. Gilles identified an issue with this dual dependency to 13.1.1 and 17.1 models. We may need to migrate the entire stack of the models.

• Smartoptics relation-chain:

  • “Fail fast if TAPI service already exists”, “TAPI - apply OpenROADM topology updates to NEPs” have been reviewed by Orange (Gilles and/or Olivier)

  • Joakim changed to Private status the other changes of the relation chain. Gilles highlighted that

• "Update pre-commit config file”: is an attempt to solve the issue we currently have with Github action. However, this does not solve the issue.

• Olivier just updated his relation chain to solve some pylint issue. We however have some issues with the Daexim dependency. Daexim code would need to be integrated into NETCONF so that it is integrated in an upstream dependency. Robert is working on the issue but need a formal approval form Daexim developers to do so, most of them not working anymore on the project

• Andrea made a summary of the demonstration they run during OFC 2026.

• Ti-SR2: Still in a pending status since we see some issues with the integration test results of the BGPCEP project.

• V-SR1: still in a pending status.

• The bottom of SmartOptics relation chain up to “Copy OpenROADM service to TAPI” has been merged.

• Smartoptics relation-chain:

  • Gilles started the review of changes on top of “Copy OpenROADM service to TAPI”

  • it has been complemented with around 30 additional changes.

  • The huge size of this relation chain leads to some issue, especially when the relation chain is rebased, since this affects other members of the community saturating the CI.

• 2 Changes from the LFN need to be merged.

• “Add E2E functional tests for OC service creation” has been pushed by Christophe but does not pass the gate for still unidentified reasons (Functional tests failing seems unrelated to the changes made).

• Olivier published 2 new changes.

  • The first, “Fix for transversal routing over tapi topos” provides some fixes so that the PCE can route over several TAPI domains (different topology Uuids) interconnected together through interdomain links.

  • Tapi Functional Test06 allows testing it, uploading a SBI Topology through Daexim features and creating dynamically an OpenROADM topology, brideged to the first through 2 interdomain links. As Robert mentioned that the way Daexim feature is activated will lead to some issues, we will need to find a way to activate this feature in a different way to run this test.

  • The second “Switch to UNIDIR links in Topo conversion”:  Switches from BIDIR to UNIDIR link in TapiLinkImpl createTapiLink  keeping BIDIRECTIONNAL Connections (for OTN Link). This is made to Improves stability in Uuid of created links, and better fits with OR link creation rationale.

• Several changes from github Gerrit are also under review…

• We plan to remove support for OR 1.2.1 in Cr.  We will need to check as AT&T plans to use release 17.1 for devices how we should add 17.1 models including device, but also common, while avoiding potential conflicts with service and network models which also rely on common models. We may need to update also Service and Network models to the same release. No disruption was identified from R13.1 to 17.1 for what concerns Network and service models. However, the structure of the operational mode catalog has been updated and we may need to adjust the code and the catalog we upload at controller initialization.

• Ti-SR2: All projects were released end of last week, and Gilles released tpce this week

• V-SR1: All projects were released end of last week, and Gilles released tpce this week. Main features concerns contention-less Management and Tapi Topology bug fixing & developments, OpenROADM Service and Network Models evolution to 13.1.

• Today was the official date for Cr release, but this will be delayed.

• Christophe has removed OC models 1.9 from models for migration to Cr. We still need to remove the 1.2.1 models and add 17.1 OR models.

• On tpce side we will need to remove the code associated with OC 1.9 and OR 1.2.1

• “Bump Upstream dependencies to V SR2” (error in message, concerns in fact SR1), “Bump Upstream dependencies to Ti SR2” and other associated changes have been merged to handle the migration to these releases.

• Smartoptics relation-chain:

  • Olivier reviewed SmartOptics relation chain up to “Copy OpenROADM service to TAPI”. All of the changes have been considered as OK (+1)

  • Changes on Top of this ones focus on TP and NEP Management starting with  “TAPI-OpenROADM TP to TAPI ONEP spectrum mapping”, a huge change.

• “add E2E functional tests for OC service creation ”, pushed by Christophe is still not fully complete but tests service creation for OC Release 1.9.

• Robert reviewed the change associated with the add of Daexim dependency, highlighting an issue : Daexim is a downstream MSI project on which we can not rely as a dependency : transportPCE can only rely on upstream dependencies. We will ask the question on how to solve this during the TSC.

• Last meeting was canceled following a technical issue on Orange internal Network

• Ti-SR2: we are still waiting for NETCONF to be released.

• Changes associated with OpenConfig R2.0.0 models easing to set up the demonstration for OFC on transportpce-models

• A lot of changes were provided by Anil to solve issues with the CI.

• The changes updating Maven version pushed by Gilles to solve gate issues :  “CI: update maven version used by tox-verify job”

• “ci: disable Jenkins tox-verify for transportpce master” : Gilles reviewed it and requested to not merge it, since he did not see equivalent feature available on Github Actions (GHA) and until now Jenkins provides the only way to run functional tests of TransportPCE on the gate.

• IPoWDM use case: Bala has published all the code associated with IPoWDM use case, except code associated with specific-operational-mode changes which was pushed on a specific private branch to test IPoWDM ATT implementation.

• OFC Demo: The LOG associated with the issue setting temporary services are on a sever shipped to OFC Exhibition. Thus UTD can not send them before they arrive in LA and start this server.

• Ti SR2 : still waiting for NETCONF R 9.0.2 to be able to release transportpce (transportpce/models has been released)

• Annual election for several leadership positions at the LFN TAC (Technical Advisory Council). For some of them, need to be a committer of a LFN project, for others not.

• The bottom of the stack dedicated to TAPI refactoring up to “Improve readability by creating helper methods”.

• “Service creation with OC-XPDR and OR ROADM” provided by Fujitsu to support OpenConfig 1.9.

• Gilles will continue reviewing the changes of the stack dedicated to TAPI refactoring starting from “Improve readability by creating helper methods”.

• Olivier’s relation chains:

  • Topo Uuid usage consolidation: the 3 changes are ready for review. An additional change is forecasted on top of it to allow for setting externally 2 variables: 1 to set the way the OpenROADM Topology is translated in the TAPI Datastore (Full versus Abstracted), and the second to set the behavior for saving services in TAPI datastore (all service saved versus only services created from TAPI NBI). For this Oliviers ask for help. He has locally created a change, but some deeper refactoring of corresponding classes will be required to be able to handle this.

  • Transversal TAPI topology consolidation: the change “SBI tapi topology Listener” has been completed to improve the management of Inter-domain link. It is still WIP, but changes below could be reviewed.

• “NetConf Proxy Tool” change has been provided by Fujitsu and corresponds to the translation of the existing Proxy Tool we had based on Pearl into a Java version, which will be very useful and will ease the maintenance.

• “Add support for proprietary operational modes”: an update has been provided, but the change is still WIP as some tests fail.

• Gilles made an update on bump for TI SR2, but is still waiting for the update of the NETCONF project

• OFC Demonstration will be based on OC Release 2.0. This implementation shall be provided by Fujitsu in a short time as we merged the end to end service implementation for release 1.9. (which was a pre-requisite from Fujitsu)

• Andrea mentions that all equipment will need to be shipped in a 2 weeks’ time-frame, which means OpenLabs needs the code asap, and at the latest, early next week. Christophe will send a reminder and explain the situation to Roshan.

• Ti SR2 : still waiting for NETCONF R 9.0.2 to be able to release transportpce (transportpce/models has been released)

• Upstreams regarding TiSR2 (TPCE models) : “Bump Upstreams to Ti-SR2” & “Bump project version to 22.2.0 SNAPSHOT”

• “CI: Update maven version used by tox verify job”

• “Adapt serialize context Test from test-common”

• “Fix a bug in functional test script”

• “Fixup Javadoc-based violations in OLM”

• “Bump project version to 12.1.0-SNAPSHOT"

• “TAPI - Refactor populateNepsForRdmNode readability”

• “TAPI - Unambiguous available spectrum”

• “TAPI converting SRG PP used wavelength”

• “Support for complete public OpenConfig 2.0.0 models will probably been abandoned in favor of the one that only uses the modules necessary to run TransportPCE and the one that adds an additional revision common to the same release (for future developments to be able to simultaneously support multiple releases of openconfig models).

• “Add a set of OpenConfig 2.0.0 models” will be merged on Chromium, whereas pre-rel 1.9 models will still remain on Ti and V.  

Note : we plan to abandon OR Device 1.2.1 support in Cr.

• “Service creation with OC-XPDR and OR-ROADM” change was pushed by Fujitsu on behalf of NTT. Christophe has reviewed the change. It is based on 1.9 models of OpenConfig. We plan to merge (on master and backport it on stable Vanadium) this change asap, even if we don’t have corresponding functional end to end functional test, since this code is to be used for OFC demo. Fujitsu will then push code associated with OC 2.0 which at the end shall be merged on Cr.  

• “SBI tapi topology Listener" : provides a TAPI SBI topology listener to reflect changes occurring in TAPI SBI topology in the OpenROADM node that abstracts this topo. TapiProvider was adjusted to register the listener at tapi feature installation as well as a functional test that verifies appropriate TPs are populated in TAPI-SBI-ABS node after a SBI Topology is loaded in the DS.

• ”Initial transversal Topo func test" : checks PCE behavior across several TAPI topologies of different Uuid.

• “Add SBI topo file to sample configs” : a new topology file that mimics the topology retrieved from TAPI South Bound Interface that still needs to be developed, is provided, to be copied to karaf/target/assembly/daexim folder.

• “Add Daexim dependencies to odl-tpce Pom” : adds Daexim dependencies in common so that the feature is activated launching karaf.

• Smartoptics relation chain: Christophe merged the 3 bottom changes. Gilles made some additional comments on some of the changes. Joakim rebased the relation chain and addressed comments from Gilles, but it seems there is an issue with the gate, which makes them fail.

• “Add provision mode for B100G OCH interfaces” which provides Profile based provisioning was abandoned. Akash and Bala are still working on it, and Bala will provide updated changes. 

• Aparaadjitha mentions that trying to create Degree NMC interface with Lighty node (for OpenROADM Version 7.1) we get “operation-fail” answer.

• “Bump upstreams to Ti-SR2” has been merged on transportpce-models.

• “Bump the lightynode version”: this version of lighty (22.1.0.6) uses Ti SR1 upstream dependencies, to be able to use transportpce-models compiled in Ti-SR1. A newer version (22.1.0.7) supporting OC models 2.0 is also available on gitlab.com public repository.

• “Bug in functional tests scripts”: when only renderer functional test was selected the tests were not correctly launched.

• Some changes are associated with the preparation of Ti-SR2:

  • “Bump upstream dependencies to Ti-SR2”: this change is still based on 9.0.2 SNAPSHOT version of Netconf.

  • “Fixup javadoc-based violations in OLM”: new one related to the new version of checkstyle

  • ”Adapt serializeContextTest from test-common”: backported from master

  • “CI: update maven version used by tox-verify job”: backported from master

•   Smartoptics: relation chain could be merged up to TAPI- Refactor OpenROADM TP datastore reads

  • “TAPI OpenROADM TP to TAPI ONEP Spectrum Mapping” provides in in package-info relevant information on associated mechanisms

  • “Rest listener state on all exit paths” has been removed from the relation chain since it was leading to some issues. Some deeper investigation is required to correct the change.

•   Other changes :

  • “Modify validation process in PceTapiLink”  and “consolidate topoUuid usage step 1 & 2”: Olivier should add the Jira ticket and reviewers

  • We are waiting for a contribution from Fujitsu to support OpenConfig R2.0. Fujitsu successfully tested LightyNode 22.1.0.7. They made it run with the developments they made. They plan to push some code by the 13^th of February. As OC 1.9.0 is no longer needed. We will probably abandon the support of 1.9.0 and keep only 2.0.0 release support in Cr. Fujitsu is currently evaluating how we can support in the proxy-hijacker the modification of hello messages. For Cr targeted for march we will remove support 1.9.0 and will rely on modified version of 2.0 models.

• For Chromium release, targeted for march, we will remove support of OC 1.9.0  and will rely on the stric necessary subset of openconfig models in version 2.0.0. Gilles suggests that we also remove the support for OpenROADM R 1.2.1 in Chromium . Bala mentions that AT&T may provide code to support OR 17.1 in transportpce.

• Vanadium has been approved and is available since the 26 of January.

• Ti SR2 is still not available since we are waiting for Netconf 9.0.2 availability.

• “Bump upstreams to Ti SR2” and “Bump project version to 22.2.0-SNAPSHOT” have been merged on TransportPCE models.

• “Add a set of OpenConfig 2.0.0 models” and “Update openconfig models” : For OpenConfig we will keep only the required model since we have some limitation with Yangtools for what concerns include statement . So the models pushed have been adjusted and only corresponds to a limited set of modules.

• “Bump the Lightynode version": Adopts the 22.1.0.6 version of lightynode (aligned with Titanium-SR1)

• “Modify validation process in PceTapiLink”  and “consolidate topoUuid usage step 1 & 2” : Olivier should add the Jira ticket and reviewers

• “Service creation with OC XPDR and OR ROADMs” has been pushed by Fujitsu but is still not reviewed. Christophe will review it beginning of next week.

• Smartoptics provided a relation chain to refactor some of the TAPI code to improve reusability and added some new functionalities in the 5 Top level changes. These last allow listening for topology update in OR when a service is created. TAPI owned node edge points are then updated with their spectrum capabilities in the TAPI topology. Creation of connection still needs to be investigated.

  • “TAPI - apply OpenROADM topology updates to NEPs”

  • “TAPI - add OpenROADM TP mapping and NEP PACs”

  • “Copy OpenROADM service to TAPI”

  • “Send OpenROADM topology update notification”

  • “Refactor spectrum extraction and remove wrappers”

  • “TAPI add SpectrumRanges-based spectrum extraction”

  • “TAPI - Add empty() factory method for Range”

  • “TAPI - round THz->Hz conversion in TeraHertz”

  • “TAPI - Normalize SortedRange by merging intervals”

  • “Cleanup DefaultOpenRoadmSpectrumRangeExtractor”

  • “TAPI - Extract OpenROADM spectrum range parsing”

  • “TAPI - extract SpectrumCapabilityPac factory”

  • “TAPI - Refactor photonic NEP spectrum construction”

  • “TAPI - Refactor OpenROADM TP datastore reads”

  • “Simplify CEP-in-topology checks in TapiContext”

  • “Prevent NPE processing links during OTN creation”

  • “Reset listener state on all exit paths”

  • “TAPI - Reduce log detail and log file size”

  • “Improved readability by creating helper methods”

  • “TAPI - Refactor finding avail/used freq”

  • “Deduplicate TTP C-band available freq map logic”

  • “TAPI - Value object for OpenROADM topology node id”

  • “TAPI - Refactor populateNepsForRdmNode readability”

  • “TAPI - Unambiguous available spectrum”

  • “TAPI converting SRG PP used wave length”

• Dr Fumagalli made an update of OFC 2026 ongoing work to prepare the demos in UTD. From the discussion they have with equipment manufacturers it seems that tpce was not tried with some of the devices that will be provided : in some of the showcases some equipment will host R2.0 of OpenConfig models. Christophe made an update on the way we will manage OpenConfig in TransportPCE

  • We will keep only a subset of about 15 of 130 models in tpceModels.

  • The models will be adjusted (Revision fixed to an arbitrary date across all models, some files imported rather than included)

  • LightyNode will be customized, and we will have specific version for each release to avoid conflict when having different releases cohabiting, which is impossible to manage with both ODL and Lighty.

• We will use the proxy-hijacker to remove/adjust capability statements coming from devices.

• Jimmy mentions the fact that OpenConfig is using both the revision and oc-ext:openconfig-version, a specific key to handle version.

• Status on Vanadium: TSC is close to deliver the V release. A release candidate is going to be proposed to the TSC for vote..

• No changes merged this week.

• “Bump the lightynode version" adopts the 22.1.0.2 version of lightynode (aligned with Titanium-SR1). This will allow to manage next version of tpce-models, handling several releases of OpenConfig in the same instance of tpce. OpenConfig models R.2.0 will need to be backported in Ti Release.

• “Consolidate topoUuid usage step 1 & 2": through both of these changes, we introduce topoUuid as an input parameters of some TAPI dedicated methods, to make sure they can be used whatever is the considered TAPI topology (Multi-Layer, Full Muli-layer, Alien, SBI) and try to remove hard setting of the TopoUuid in the classes where this might not be relevant. The goal is to be able to handle the different topologies that will cohabitate in the dataStore whereas until now we focused on Full Muli-layer topology.

• ”Modify validation process in PceTapiLink": this change applies the same approach as the one used for Pce Nodes, when we refactored it to address some comments provided by Smartoptics. Olivier will need to complete his commit messages.

• “Service creation with OC-XPDR and OR-ROADM”: Portmapping and Renderer have been complemented with additional methods to provide end to end service provisioning with OpenConfig Transponders. Test will need to be complemented (either in OC dedicated tests or through hybrid tests).

• “Added available frequencies”: Jimmy implemented both scenarios 1 & 2 as proposed. This change is ready for review and if no specific issues is identified, it could be merged.

• Smartoptics big relation chain needs to be rebased on master since it does not pass the gate.

• Bala mentions that they start having equipment in their labs with OpenRodam models in version 17.1. These models (device an common) will need to be included in tpce-models. Some bugs have been identified with some corrections applied on the devices that won’t be backported to a 17.1.1 model. Thus, some models might be corrected and may be slightly different from currently official 17.1 models.

• Gilles and Olivier will work on consolidating backlogs, so that if we have time during next meeting, we can launch a new development sprint.

• Status on Vanadium: Vanadium GA has still not been submitted to the TSC vote.

• “CI: update Maven version used by tox-verify job” makes the CI functional again. Need to rebase current relation chain to make this fix effective.

• Smartoptics relation chain: Joakim has associated tickets to his changes. Tapi Listener on OpenROADM topology updated is still in Work in Progress. Changes at the bottom of the relation chain are open for review.

• “Added available frequencies” has been removed from the relation chain. Olivier was fine with the 2 proposals made.

• “WIP : support for OpenConfig V2.0.0”: this change allows considering the new reduced set of OpenConfig models introduced in transportpce-models. Its works in a local environment and allows having several OpenConfig Releases’ models supported concurrently. It is not passing the gate, since his related change on transportpce-models is not merged yet.

• “Add a set of OpenConfig 2.0.0 models” proposes to include only absolutely required yang models. Christophe needed to slightly modify the openconfig-platform model, to directly include into it groupings coming from another yang file managed as an “include”. This means this model diverges from the original OpenConfig model. The fact that this model has been modified must be mentioned in the commit message as well as in a specific file under the resource directory. Revision have also been added everywhere.