The future is already here Mastering the challenges of the coming years Uwe Friedrichsen – codecentric AG – 2008-2023

Uwe Friedrichsen Works @ codecentric https://twitter.com/ufried https://www.speakerdeck.com/ufried https://ufried.com/

3 challenges

1 The world has gone VUCA

Volatility Uncertainty Complexity Ambiguity

Volatility – Things change dynamically Uncertainty – Lack of predictability Complexity – No long-term upfront planning Ambiguity – Unclear meaning of signals

Drivers of VUCA Mostly unpredictable market behavior Exception: Non-efficient markets Future effects very hard to predict Consequences of adverse events unpredictable Future effects unpredictable Production processes, supply chain, … From a traffic jam to COVID Chip crisis, supply chain crisis, … Post-industrial markets Supply >> Demand Highly dynamic Geo-political developments Power structures shift Economic Political Over-optimization Very narrow tolerance ranges Extremely vulnerable to deviations from the norm Very long recovery times Climate crisis (Finally) becoming a palpable factor Unchecked complexity

Planning is dead. Convictions are lethal. Best practices do not exist. Not even good ones.

Probe-Sense-Respond

2 IT has become indispensable

Digital transformation made IT indispensable

Business functions Business processes Customer interactions Business models 1960 1980 1990 2000 2010 2020 1970 IT support IT dominance

Business is IT. IT is business. Business and IT are the same side of the same coin. The other side are the market and your customers.

IT has become indispensable for our business and private lives. Tendency: Rising.

3 IT complexity is growing uncontrolled

Complexity drivers in IT Ongoing digital transformation CV-driven software development Trying to solve non-tech problems with technology Project-centric thinking Local, short-term optimizations Recurring “cleanup” initiatives Never completed, only adding more complexity Never retiring old technologies Falling for FOMO Chasing the next technology hype Efficiency obsession Yearly planning cycles Company self-absorbedness Neglecting effectiveness …

Nobody knows what our IT landscapes do. We can only guess why they are running. We can only guess how to evolve them. To solve the problem, we add more complexity.

3 responses

Indispensability Sustainability VUCA Complexity

Sustainability is much more than (just) ecological sustainability

Indispensability Ecological Going green in IT Social Sustainable for everyone involved Technical Creating dependable systems Sustainability Economical Surviving in the long run VUCA Complexity

VUCA Indispensability Ecological Economical Social Technical Resilience Sustainability Capabilities Complexity Stability Maximize MTBF (before resilience) Robustness Ability to handle expected adverse events (partial resilience) Surprises Ability to handle unexpected adverse events (actual resilience) Anti-fragility Improve and thrive from adverse events (advanced resilience)

VUCA Indispensability Complexity Ecological Economical Social Technical Responses Withstand Recover quickly Adapt Transform Resilience Sustainability Scope IT Systems IT Organization Division Company … Capabilities Robustness Surprises Anti-fragility Stability

VUCA Indispensability Complexity Ecological Economical Social Technical Responses Withstand Recover quickly Adapt Transform Address at all levels Resilience Sustainability Simplification Accept essential complexity Avoid accidental complexity Scope IT Systems IT Organization Division Company … Capabilities Robustness Surprises Anti-fragility Stability Reduce overall complexity Watch the sum, not the summand Market Governance Organization Projects Process models Requirements Architecture Technology Implementation Improvement initiatives Optimize over time Take future effects into account

VUCA Indispensability Complexity Ecological Economical Social Technical Responses Withstand Recover quickly Adapt Transform Address at all levels Resilience Sustainability Simplification Responsible IT Avoid accidental complexity Scope IT Systems IT Organization Division Company … Capabilities Robustness Surprises Anti-fragility Stability Reduce overall complexity Optimize over time

How to get started

There are a lot of things we can do …

Compliance as Code Resilient Software Design Resilience Automate everything IaC CI/CD … Batch Size 1 DevOps (incl. Team Topologies) Learning culture Preserve slack Observability Continuous training (e.g. fire drills) Chaos Engineering SRE Sustainability Green IT CSR/ESG Focus on effectiveness (before considering efficiency) Accelerate Hypothesis-driven software development Understand post-industrial markets Foster dependability Use boring technology (resist FOMO, leave IT rat race) Avoid overengineering Understand the business domain Rethink incentives Abandon projects Beware local optimizations Improve design skills Simplification Abandon cargo cults Construction for deconstruction (Easy to remove) Changeability (Easy to adapt) Usability (Easy to use) Enterprise “Agile” Enterprise “DevOps” Abandon big cleanup initiatives Help people understand software Stop “solving” non-tech problems with technology

… but the hardest part is to get going

Deeply ingrained habits stand in our way

Deeply ingrained habits Ongoing digital transformation CV-driven software development Trying to solve non-tech problems with technology Project-centric thinking Local, short-term optimizations Recurring “cleanup” initiatives Never completed, only adding more complexity Never retiring old technologies Falling for FOMO Chasing the next technology hype Efficiency obsession Yearly planning cycles Company self-absorbedness Neglecting effectiveness …

The thought alone is killing me

But what is the alternative?

After all, we are in for an endless game bearing growing responsibility in an ever less predictable world And no, AI will not magically solve this challenge

How to get started

You are here You want to get there Create awareness Provide transparency Offer training Implement pilot Rollout Continuous improvement … … … Start here … … and here A (very) simple change process

Create awareness • Requires a topic evangelist or alike • Must convey conviction and enthusiasm • Must be willing to communicate the message multiple times • Creates initial awareness • Creates initial group of supporters • Overall effect limited but required to get started

Provide transparency • Make status and progress measurable • Numbers are harder to ignore or argue away • Closes missing feedback loop (nobody wants to be “last”) • Often much more powerful than any talk • (Still, an evangelist is needed first) • Measure outcome, not output • Focus on business-relevant metrics • Start with metrics that allow for short-term progress • Make sure relevant stakeholders can shine with progress

Potential metrics (1/2) • Value chain mapping • Feature usage • Feature energy consumption • Dependability index for applications • SLO/SLI • DORA metrics • Incident frequency

Potential metrics (2/2) • Code complexity correlated with, e.g., change failure rate and developer satisfaction • Service startup times • Self-healing score • Time to restore service • Application (remote) dependency score ... and many more

Being an architect • Architects need to be good communicators • Good prerequisite to create initial awareness • Architects need to be good software designers • Good prerequisite to provide a metric • Architects should know how to code • Good prerequisite to implement a metric • Architects need to collaborate with operations • Good prerequisite to collect actual data

Architects have the ideal preconditions to start a movement

So, maybe you want to start a movement … ;)

It's easier to ask forgiveness than it is to get permission. ― Grace Hopper

But isn’t that old wine in new skins?

Yes and No

The pieces already exist …

… but the combination is different

The future is already here – it's just not evenly distributed. ― William Gibson

Uwe Friedrichsen Works @ codecentric https://twitter.com/ufried https://www.speakerdeck.com/ufried https://ufried.com/