What is Scaled Agile Framework
The Scaled Agile Framework?or SAFe is a template for scaling Agile Principles and Tools to large Organizations. SAFe has been implementing in both Private and Public sectors industry banks, governments etc. In this article, we'd talk about what SAFe is all about and how it works.
SAFe is a framework meant to cover the entire organization and it operates on 3 levels.
Team
The team level works very much like a standard scrum, also the team can work in Kanban. At this level, we have a team that is cross-functional including developers and testers who work together to deliver working software every two weeks which are called iterations. The content for the iteration is determined by the Product Owner who's in charge of the team Backlog which is an ordered list of user stories. The iteration starts with a planning meeting where the team decides what user stories they can deliver by the end of the iteration. Each day the team meets at the daily scrum to discuss their progress and at the end of the iteration, the team gets together with the product owner to demo what they have done to ensure they've delivered what has been decided upfront. Then they get together to retrospect what they can improve for the next sprint before starting their cycle again with a new planning meeting. All of these are guided by Scrum Master who make sure the team works smoothly within the process and that it keeps improving. You might wonder why I spend so much time writing on basic scrum which most of you might know. The reason is the program level which is the heart of SAFe is built exactly on the same elements, only scaled up nevertheless.
Program
At the program level, we have a team comprised of multiple teams working to deliver something together. This can range from 50-125 people. These team will also time work their efforts in one program increment also called pi which lasts 8-12 weeks and together deliver bigger working software. The main metaphor for this team of teams is the agile release train. Imagine you're flying to Paris from San Fransisco, we need to book the flight ticket in advance, verify the time and reserve a seat. In contrast, when we take the subway we don't plan ahead, we just go and catch the train. WHY? Because trains are so frequent and ride at different intervals, we know that if we miss one train then we board the next one. The same is true of agile release train or art. Because it is constant and relatively short it's easy to explain to stakeholders that if we missed this train we just wait for the next one. pi will provide most of the content to the team backlogs; at this level, content is called features which is either functional or architectural in nature. The train is governed by RTE or Release Train Engineer. Who acts as the trained Scrum Master ensuring it runs smoothly and remains on track. Each pi starts with a planning meeting in which all the people on the train and all the members of the team get together to hit a vision and road map of the train and goals for the upcoming pi. Each team then breaks out to plan what they can do in this pi while communicating with all teams on the train whom they might need to collaborate.
In addition, the team capture dependencies between themselves to make sure the train is headed in the right direction. Program coordinate dependencies through sync meetings. Scrum masters gather together on a weekly basis to show visibility on progress and impediments they're facing. This meeting facilitated by RTE is called Scrum of Scrums. Product managers and Product Owners have their own meeting to discuss visibility into progress and priority adjustments and it's called the PO sync. Every two weeks a system demo takes place to review the new features and open discussion for feedback. This demo offers real measures of the system value, velocity and progress to learn and adjust. Note that the time should be taken dedicated to preparing the pi. Because SAFe is a framework to relentlessly improve and adapt. The pi retrospective takes place during the inspect and adapt event. Its goal is to identify systemic issues that need to be addressed.
领英推荐
Eventually, since innovation is a key safe success factor an innovation and planning iteration event is organized to stimulate innovation hackathons, infrastructure improvement and provide cadence based planning. The last thing to know at this level is the architectural runway since the train tries to run fast as we can. We need to provide good architecture and infrastructure. Each pi gives consisted in adding technology work and growing architectural runway to support future features. We build the technical items we think that we'll need in order to achieve our goals in the following pi these items become the architectural features and user stories in the backlog and are the responsibilities of the train architects.
Portfolio
The portfolio level is slightly different and similar to the previous levels. At this level, the programme portfolio management organizes the portfolio around value fund investments and deals with agile portfolio operations and lean governance. It allocates budget on chosen investment themes and creates a backlog of portfolio-level epics which need needs to be addressed by trains product management in the upcoming pis. To ensure cohesiveness at this level we use a kanban system to limit the amount of portfolio level initiates and make sure all trains are focused on finishing those initiatives together rather than starting new initiatives.
This was a quick summary/tour of what safe is and how it works. It's important to know that the safe is a template for starting the Agile implementation and the organizations will likely customize it to their situation but it does provide a clear path of how to start scaling agile outside the teams.