Wiki

Date: June 11, 2015

Arno release retrospective material

• Cross team/project collaboration
  • The community members should continue to contribute across different projects. If there are dependencies across projects a contribution to overcome them is the best way to keep moving forward.
  • Willingness to help on-board new members (e.g. on IRC)
• Great leadership from the Functest project team who had to step forward in the past few months and manage the transition.
• Engagement with upstream communities (e.g. impact recently at the OpenStack Summit)
• Good visibility of Arno on social channels

• Arno retrospective Etherpad
• Burden on a small number of SMEs as was also highlighted in other email threads
• Should be more clear on the areas where we need resources will help our members and the broader community find the right people to have involved
• Discussed more resource needed for
  • ODL SME
    • Maybe OSGI/Karaf people to help with platform issues you see on ODL or ONOS
  • People who are in telco operations who are actively deploying applications today and can provide feedback
  • Testing competencies, with broad experience in testing for application deployment scenarios
  • Collaboration with SDOs
• More resource on training and on-boarding is needed (both for new developers & users)
  • Better project information/dashboard will be helpful for people that are new to the project

• Scope management
  • Within 2 months of project launch, there was a relatively clear idea of the Arno scope. This reduction in scope was a contributing factor in being able to deliver.
  • The communication of the Arno scope after the decision was made could have been better
  • Not having a clear idea of the resources available made it very difficult to manage the risk of the scope
• There was not a change management process for decision making during Arno and better communication could have reduced friction
  • On change management, there's a need to be wary of imposing too much top-down control/benediction of project leaders vs allowing leaders to arise organically as we grow as a community.
• A structured approach to using Jira would be helpful in coordinating community efforts
• Schedule
  • Additional effort is needed to address risk mitigation when making our plans
  • Successful time-based projects manage releases by gating early on feature readiness and gating early
  • Establishing clear milestones and being concise on the meaning of those milestones and qualifications will be important
  • Individual projects will have their own timeline that may or may not coincide with a time-based release activity. We need to be clear on the correlation between release and content.
  • Discussion that it may be acceptable to set our release date based on upstream dependencies rather than trying to stay to a completely time-based release cycle.
    • There was a discussion about release cadence, and how features will make it from upstream into a release.
    • There is not a shared understanding of what a "release" means, what release artifacts the project will create, or even what it means to be in a release
• Project resources
  • Multiple timezones require effective communications in the community.
  • Discussion on challenges with the OPNFV hardware in the LF lab. A question was raised if there is a way to automate a lot of activities on the LF hardware.

http://meetbot.opnfv.org/meetings/opnfv-meeting/2015/opnfv-meeting.2015-06-11-13.01.html