Implementing Large Scale DevOps

With excerpts from Industrial DevOps IT Revolution DevOps Enterprise Forum 2018

What are few problems with large systems which would hamper DevOps practices?

• Visibility of tasks and activities in the value stream decreases as the complexity of system increases.

• Most of the large systems are organized very much in siloed fashion with gated approach and numerous hand-offs.

Slow time to market, quality issues, cost overruns, mis-fit solutions

How large scale DevOps will help?

• Large scale DevOps includes various bodies of knowledge including DevOps, Lean Manufacturing, Lean Product Development, Lean startup, systems thinking, scaled Agile development

• It helps building a continuous delivery pipeline that provide a multi-domain flow of value to the users and stakeholders those deployed the systems

As per IT Revolution Enterprise DevOps Forum what are few of the recommendations?

1. Visualize and organize around the value stream.

2. Apply multiple horizons of planning.

3. Base decisions on objective evidence of system state and performance.

4. Architect for scale, modularity, and serviceability.

5. Iterate and reduce batch size.

6. Establish cadence and synchronization.

7. Integrate continuously

8. Be test driven.

Visualize and organize around value stream:

• Normally value stream should be around independent capabilities in large complex systems

• Team should remove activity based hands-off and focus on providing the specific business outcomes

Apply multiple horizons of planning:

• Product vision and major capabilities - 3 to 5 years

• Year lookout of high level functions - 1 year

• Quarterly roadmap of features - every quarter

• Sprint plans - 2-4 weeks

• Daily team plans

Base decisions on objective evidence of system state and performance

• Each increment should provide an opportunity to demonstrate objective evidence of the feasibility of solution in progress

• Reviews are done on cadence so progress can be measured by stakeholders frequently, predictably, and throughout solution development life cycle

Architect for Scale, Modularity, and Serviceability

• Design and engineering processes need to scale to address the complexity in large systems. This can be done through modularity and serviceability.

• Modularity: Component decoupling with a focus on the smallest unit of functionality.

• Serviceability: Focus on lowering cost and time required to alter functionality both pre and post production

• Architecture modularity significantly impacts DevOps goals to continuously develop, integrate, deploy, and release value

Iterate and reduce batch size

• Work in small batch sizes - both in terms of the size of the component or feature and the unit of change.

• Enables agility, regular validation and rapid feedback cycle

• Model-based systems engineering, A/B testing, and dark launches provide the ability to make changes and rapidly understand the impact of those changes.

Establish Cadence and Synchronization

• Provide rhythmic pattern to manage inherent variability in solution development

• Cadence makes wait times predictable, leading to lower transaction cost of key events

• Synchronization helps all teams working at different levels align their efforts

Integrate continuously

• Continuous integration is a planned strategy to integrate frequently, based on the economic trade-off of the transaction cost of integration versus the risk reduction of objective evidence of system performance

• Partial integration does helps address the risk even when full risk is not possible

Be test driven

• Systems built on modular, component-based architectures that communicate through well-defined interfaces are much simpler to test and verify functional behavior

• In test-driven cultures, the environment automatically runs a rich set of component tests on any change

Ref: Industrial DevOps: Applying DevOps and Continuous Delivery to Significant Cyber-Physical Systems