This Page is now closed to new topics and the schedule is posted here.
(please let Ray Paik to if you want to add a new topic to the schedule)
|Topics||Time budget (e.g. x hours)||Participants||Notes||HW resource plan/requirement|
|Best Practices/Use cases/Demos|
|OPNFV Deploy & Test: Apex||2 hrs.||Targeted for newcomers to OPNFV||Lenovo pod|
|OPNFV Deploy & Test: Compass||2 hrs.||Justin chi|
|OPNFV Deploy & Test: Daisy||2 hrs.||Julien|
|OPNFV Deploy & Test: Fuel||2 hrs.||ENEA Pharos pod|
|OPNFV Deploy & Test: JOID||2 hrs.||Narinder Gupta||Lenovo pod|
|Telco host scheduled maintenance||1.5 hours|
POC demo and discussion about the use case and approach.
More in Doctor hackfest session. Session is based on this: Telco Planned Host Maintenance
|Best practices of Compass4nfv: The construction of Windriver NFVI using Compass in Chinamobile||1hr||Automatic installation of Openstack using Compass in China Mobile|
|Best practices of Daisy4nfv: The construction of TECS(ZTE) NFVI using Daisy in Chinamobile||1 hr||Xu Wang||Automatic installation of Openstack using Daisy in China Mobile|
|Automatic Integration and Testing System based on both Opensource and Standardization in China Mobile||1 hr|
China Mobile's view on automatic integration and testing
Show automation system demo
|Multi-Access Edge (Edge) project / OPNFV community demo 2.0||2 hrs||Multi-Access Edge|
|k8s in OPNFV||2 hrs.||Intel Pod.|
|ODL Container Orchestration Engine Integration upstreaming to OPNFV||3 hours|
ONAP deploy with XCI
CENGN vPOD (remote)
(remote - ssh)
|ETSI Plugtest in Jan and following co-hosted Plugtest/Plugfest||2 hrs|
|CVP and Dovetail||3 hrs||NA|
|Stress testing||3 hrs|
|VNF/NFVi characterization w/ Yardstick||3 hrs|
enabling new testcases, CGNAPT, vFW RFC2544 with Ixia/Trex with installers.
VNF Scaling tests
Testing new VNFs with Yardstick
VNF basic testing for Onboarding
|Installer External/Provider networks for testing VNFs with HW traffic generators||1 hr.|
Installer support for VNF characterization
Connecting external traffic generators to OpenStack
Adding external traffic generators to scenarios
Traffic-Generator As A Service (TGAAS) – develop common methods and tools for use in performance testing projects.
|2hrs||We would like to reuse, merge and extend existing TG methods across performance testing projects. We would also like to explore creating a new project for Traffic Generators which is an important and complex topic. Our proposed first step is to identify abstraction differences between Yardstick and VSPerf.|
|Testing Priorities in F release||3hrs|
Landing page for unified entrance of testing resources
Long duration testing
|NFVBench and VSPERF collaboration||1hrs||VSPERF and NFVBench would like to explore areas for collaboration. Other performance testing projects are encouraged to join this session to brainstorm strategies for leveraging existing and new work planned for Fraser|
VSPERF collaboration activities
VSPERF has a Plugfest planning page https://etherpad.opnfv.org/p/Plugfest-Euphrates-VSPERF
|Data Plane Performance Benchmarking in production pods with NFVbench||1 hr|
In this session we will discuss about NFVi production deployments and challenges related to benchmarking the data plane of an NFVi pod in production conditions.
NFVbench has been released for the first time with euphrates. We will see how it complements other OPNFV data plane benchmarking projects (VSPerf, VNFSample) and go through all the features that are available today.
We will do a live demo of NFVbench running on one of the OPNFV pod.
This session will be interesting for operators and for OPNFV projects that want to include NFVi full system testing on data plane (e.g. CVP, Dovetail, ONAP...).
Preferred time slot: anytime on Wednesday or Thursday
|High availability testing cases discussion||1 hr|
HA test cases planning for F release and future releases. We will update with the community about the recent progress of the HA testcases, and have a discussion about future test cases that we proposed.
|Release Automation||1 hour|
|Docker Infrastructure & Builds||2 hour||Trevor Bramwell|
|Scenario documentation||1h||Review of the requirements. Discussion on how we document, store and work with the scenario descriptions.|
Current status, what's next, establishing pipelines for Test (and other) OPNFV Projects, which other communities we are reaching out.
CI pipeline for long duration test
More details TBD
|Experiences and thoughts on PDF/IDF and way forward with SDF||2h|
Session to have conversation around PDF/IDF so the people who worked with them can share their experiences with others. Relation of SDF can be part of the conversation.
Please note that this is not a planning or endless conversation session like the ones we had before. We are in a state that majority of the installers already adapted PDF/IDF and it is already happening. Apart from that, we seriously need to finalize the official template as it is a requirement for LaaS/other Infra Initiatives and PDF/IDF will be prerequisite for Installers to benefit from LaaS.
|Using OpenStack Kolla containers for XCI|
1h (before Thursday PM)
|The XCI project currently tests upstream OpenStack using PackStack off of master and tests the test projects. Adding a stable version of Kolla to test the test tools could ease the RCA of failures in the test tools.|
|Testing hardware/software interactions brainstorming (testing/infra cross-topic)||1h||Tim Irnich|
We are missing a way to create tests of the OPNFV software platform with the Pharos lab hardware without making tests lab-specific, which would limit usability of such tests and be in conflict with the concept of dynamic CI. Examples are some of the currently considered high-availability test cases that target e.g. switching fabric failure conditions, configuration of L2GW HW-VTEP on ToR switches, SMartNIC integration testing, DC-GW integration testing etc.
We will discuss how to address this through a set of coordinated actions towards Pharos spec, test frameworks, infra/releng, and PDF.
|Project team hacking/discussions|
|Doctor Hackfest||1 hrs x2|
Doctor Hackfest Sessions:
|Clover (Cloud native NFV)||1 hour||Tuesday||NA|
|SNAPS-OO E enhancements F planning||2 hrs (not on Friday)||Steve Pisarski||Overview of the latest features added to SNAPS-OO during E and F feature planning. This could be split into two separate sessions||NA|
|Yardstick F-release planning||2hrs|
|Bottlenecks Load Manager Discussion||1 hour||Yang (Gabriel) Yu||Registration center, service discovery, API category, job scheduler, monitoring and analytic results/visualization|
|Daisy F-release planning||1 hour||zhou ya|
|OPNFV Docs Hacking||2hrs|
|Kolla's Openstack, ODL, OVS-DPDK containers for XCI||1 hour|
|Moving to stable/latest using kolla built containers|
|Masakari integration into OPNFV releases – discussion with XCI & compass projects||1 hour|
|The HA project is working with the openstack masakari project to integrate the proposed HA API into Queen release. We would like to organize a joint discussion with the Compass team and XCI team to see if we can integrate the masakari project in F release.|
|Fraser release overview||2 hours||Tim Irnich||Wrap up and present the Fraser release overview|
|Improving OPNFV user support||2 hours||What can we do to improve usability of our release artifacts - explain our release artifacts better, download page improvements, other resources, linking, documentation quality and structure, how does the dev community support users, etc.|
|Scenario consolidation - how to make it happen?||2 hours||Tim Irnich||Reviving the scenario consolidation discussion, cf. the presentation from Beijing summit. Target is to drive specific features into the more "generic" scenarios in order to foster cross/feature integration and achieve a richer set of functionality users have available after installation of a given scenario.|
|Euphrates Retrospective||2 hours||David McBride||Wrap up discussion of went well and what could be improved, based on our collective experience during the Euphrates release cycle.|
|Project releases outside of the standard OPNFV release cycle.||1 hour||David McBride||Some projects, such as NFVBench would like to have releases outside of the standard release cycle. Let's discuss additional release models that would accommodate this.|
|Installer topics||1 hour||David McBride Tim Irnich||Discuss issues related to installers. In particular, can we improve how we work with upstream components (i.e., fixed tags vs. tip of branch)|
|TSC Composition||2 hours||Discussion on merit-based TSC composition for OPNFV|