Listen to update pitch of SAFe 4.5
Started at 5 Reasons to Update
Improve big picture usability Too complex Create more white space and reduce duplication
Menu structure improved Why SAFe ← from SAFe distilled
New name - SAFe for lean enterprises To incorporate whole business
Configurable framework Full SAFe - portfolio and large solutions Portfolio SAFe - multiple solutions but not a lot of dependencies (default) Large solution SAFe - don't need portfolio - e.g. DoD Essential SAFe - start here - sometimes it's all you need
Spanning palette icon is smaller set than rest You don't have to have metrics for example But not essential Can still use of course
4 phases of the continuous delivery pipeline Continuous exploration Continuous integration Continuous deployment Release on demand Can use parts of it - don't have do it all at once
Drive this through hypothesis, build, measure, learn
Epics are big Chance you need all of it are slim Use lean startup cycle to determine Met MVP or keep going
Changed epic value statement Now have hypothesis statement Includes intermediate measures associated with the anticipated outcomes Being clear “we don't know” upfront
Epic Then continuous exploration
Too much focus story level and this is not just about technical problem Need to integrate across whole solution And this is culture as well
Decouple deployment from release Want to deploy to production Use feature toggles etc Get feedback, check you haven't broken anything yet Turn on when necessary Prevent production emergencies
Should be release incrementally Idea of “streamlet” - part of a value stream E.g. Security updates vs release update
New program Kanban Want to go beyond getting ready for implementation So includes steps as it is released as well - more part of total value stream
No standard definition of DevOps So have SAFe definition Bridge gaps between dev and ops so deliver value faster and more reliable Reduce large batch
CALMR approach to DevOps Culture of shared responsibility Automate everything Lean flow accelerates delivery Measure everything - application telemetry (business a/b testing, but also the system) Recovery - enable low risk solutions (canary releases etc)
Culture first - interesting not culture last of rest of SAFe Both have to be equally responsible Need input from operations Need tolerance to risk - we have good safety nets (e.g. Test first) Need to share discoveries - learning across silos
Automation still important but need culture as well Match staging environments to production Match development environments to extent the same way Automate deployment - everything under version control, auto build environments, automate the actual deployment process
Lean flow Need to identify bottlenecks, and balance WIP (not through over wall to ops), reduce batch size, manage and reduce queue length
Measure everything Data collected on business, application, infrastructure, client layers Different telemetry for different stakeholders Compare to event - when did we deploy and when did we find the problem
Recovery Stop the line Plan for and Rehearse the failure
Smaller cycle for features Innovation happens at every level of SAFe Feature now have benefit hypothesis As don't really know if we are getting the benefits
Toolkit executive workshop for tipping point Some minor things as well - get more information Just documenting process that we know works Shows training
Implementing 1-2-3 has been deprecated
SPC is part of big picture Can bring experience from external
Foundation is now SPCs Values Principles Implementing SAFe Lean agile leadership
Train of agile trains Large complex systems And a lot of dependencies across the trains And need to be aligned Trains build solutions themselves - this is only large solutions
Compliance Part of solution intent Schedule and manage specific compliance activities Stored in solution intent
Team level Better align with scrum guide Iteration review (vs team demo) Dev team is “anyone contributes to the solution”
Terminology changes Note MVP is part of SAFe