Building High Assurance Systems with SAFe 4.0.

Building High Assurance Systems with SAFe 4.0.

SAFe Resources SAFe Updates

Hi Folks,

thumb_building_high_assurancesystemsSome of you already know that one of my technical pursuits over the last few years has been to better understand and describe the way in which enterprises can apply Agile methods to building big and important systems, specifically “high assurance” systems that have an unacceptably high social or economic cost of failure. We see these systems in our everyday lives—automotive and aeronautic systems, defense systems, medical devices, systems that control our financial security.

In much of my first career, I worked primarily in the medical device industry, building computer-based systems for healthcare. In that environment, a bug could have catastrophic consequences. Fortunately, to the best of my knowledge, I didn’t leave any such bugs in the field and I eventually moved on to other pursuits. But my interest in super high quality software has remained. Now, as Lean-Agile development crosses the chasm to the enterprise, it’s a good time to understand and rethink how these methods can accelerate quality, as well as speed. In support of this, we built some additional hooks into SAFe 4.0.

My talk this week at Agile2016 was on this topic. The packed room was a good barometer that this is fast becoming a new focal point for many in Agile development, so I’m sharing the slides below for everyone interested:

Anyone working with systems that are subject to regulation or industry compliance requirements, including Verification and Validation, should find this useful, especially if they are engaged in or considering a SAFe implementation.

I hope to turn this into a more detailed guidance article in the next few months, so your comments now on this topic are most welcome.

Stay SAFe!
—Dean

Author Info

Dean Leffingwell

Recognized as the one of the world’s foremost authorities on Lean-Agile best practices, Dean Leffingwell is an author, entrepreneur, and software development methodologist.

comment (3)

  1. Andy Elliis

    31 Aug 2016 - 7:37 am

    Dear Dean,

    Many thanks for driving SAFe forwards into the High Assurance Systems context.

    There are close parallels between SAFe 4.0 and the Agile approach that we used to design and build a large scale and assured high availability information infrastructure within a defense context. Our taxonomy and terms were different but the core concepts were much the same.

    Learning points for me on that programme included:
    – All stakeholders, including senior management and programme delivery, need to understand and support delivery as an Agile programme, not Waterfall, even though many aspects of a building a high assurance system using Agile may look like traditional Waterfall. In particular managers whose prior experience is predominantly Waterfall may well need education and guidance to understand the differences and to adapt their management style.

    – For large, rapidly changing and complex systems whole-system baselines are not likely to be realistic, so a System of Systems approach is likely to be more appropriate with component systems being baselined. Dependencies between the component systems, and alignment to value delivery milestones, need to be tightly managed. It is important to understand which dependencies demand synchronised delivery, and which can be asynchronous, to avoid unnecessary delays.

    – The delivery system should be set up to provide a choice over the pace of delivery for individual capabilities and features. Relatively simple changes can then benefit from an accelerated path to Live so that they are not held back unnecessarily by more complex changes. This means having different paths to live to suit different needs, such as hotfixes, minor changes to third party products, break-fixes, as well as the main value streams of features and enablers.

    – Agile needs even more discipline than Waterfall. Control theory tells us that to be agile and manoeuvrable a system has to be inherently unstable (as illustrated by the difference between an aerobatic display aircraft (unstable) and a passenger airliner (stable)), with continuous finely judged feedback and control to keep it on track. In this programme we adopted and used an MBSE approach across the whole lifecycle to support that discipline, and provide traceability to the value objectives. This made it easier to associate detailed delivery process and review point data to a programme-wide Kanban which we used to inform rapid learning cycles and accelerate service delivery.

    Thanks for bringing this high assurance systems discussion to a wider audience!

    Kind regards, Andy

    • Dean Leffingwell

      Dean Leffingwell

      03 Sep 2016 - 11:15 am

      Thanks Andy,
      This add a lot of color to some of the similar insights we have with respect to integration points, and they way in which they control the ability to objectively assess increments.

  2. Peter Pedross

    01 Aug 2016 - 4:28 am

    Dear Dean

    Thanks for sharing this information with us!
    We at PEDCO are convinced that in context with high assurance systems, agile, complex and regulated environments it is necessary to have processes and specific practices in place (e.g to manage a SRS with the help of ‘Skeleton Requirements’). In order to achieve sustained success, an organization needs strong leadership and clear strategic direction. They need to develop and improve their people, partnerships and processes to deliver value-adding products and services to their customers. I think that processes serve as a competitive advantage between competitors. As we see for example in the automotive industry with new competitors like Tesla on the market, the competitive pressure between companies is even more increased. This asks for high innovation levels and high efficiency, still while maintaining a low time-to-market. With Industry 4.0, Lean Start up, Agile and Cyber-physical systems the inter-system dependency and complexity increases dramatically. To make this job for companies even harder, the regulatory and organisational environment is becoming ever more demanding. Therefore, processes and environments must ensure that all the relevant regulations are complied with.
    As we have seen in the development of Applied SAFe; scaled agility and high assurance systems are something that goes along very well. SAFe has the concepts needed to bring such an endeavor to live. We love to support the exchange of knowledge and here’s a list of methods and processes for compliance in high assurance systems as we actually use:
    – Pilot, train and verify and processes improvements and process changes.
    – Learn as a system.
    – Share available knowledge.
    – Use defined metrics.
    – Work with proven best practices.
    – Ensure that knowledge doesn’t get lost
    – Inspire people to question the status quo.
    – Ensure feedback cycles and structured improvement of PAL.

    Thanks for bringing this topic into a broad audience!

    Kind regards, Peter

Leave a Reply