Software architecture evolution, what to expect? Panel Discussion
summary
Argues there is no silver bullet and separates essential from accidental complexity, recommends scenario-based architectural evaluation and threat modeling as underused, field-proven techniques, and urges quick prototype-and-measure experiments. Gives concrete surprises from empirical tests of NoSQL databases (table locks that kill write throughput; replication modes that silently lose updates), elevates 99th-percentile latency as the right metric for cloud services, and promotes observability, A/B deployment, visualized technical-debt maps, and Kafka-style messaging for high-scale systems.
AI-written summary · about
chapters (43)
- 0:00Introduction to panel
- 0:22No silver bullet in software
- 2:30Essential vs accidental complexity
- 4:17Ideas that sound good on paper
- 6:26Learning from internal case studies
- 7:50Vendor marketing pitfalls
- 9:41Prototyping to validate tech
- 11:25Sharing tech experiences
- 11:47General vs specific principles
- 13:35Distributed systems challenges
- 14:39Balancing generality and specificity
- 16:45Engineering ethics in architecture
- 17:48Metrics and observability
- 19:59History of software architecture
- 21:07Principles across environments
- 22:10Evolution of system demands
- 24:40Architecture in DevOps era
- 26:05Modularity and 1971 paper
- 27:29Rediscovering old knowledge
- 28:35Repeating past mistakes
- 30:21Fragmentation of knowledge
- 31:24Fundamental architecture principles
- 32:07Conway's law in practice
- 33:30Refresh on Conway's law
- 34:13Observability as principle
- 35:19Synchronous vs asynchronous trade-offs
- 36:00Evolving architecture with A/B tests
- 37:10Evergreen delivery and technical debt
- 38:14Technical debt as metaphor
- 40:21Business feedback on system health
- 40:45Evolving legacy systems
- 41:28Natural selection in software
- 42:30Systems growth and complexity
- 43:13Simplifying toolchains
- 44:18Radical simplification in design
- 45:01Software as a growing system
- 46:06Lehman's laws of software evolution
- 47:10Critique of garden analogy
- 48:15Software vs building analogies
- 49:58Quantification in architecture
- 51:02Organizational change inertia
- 51:44Automation in software process
- 53:08Closing remarks and thanks
related talks
57mSoftware Architecture: The Hard Parts
50mTowards an Evolutionary Architecture
58mThe past, present and future of Architecture
08mLightning Talk: Breaking the Mold: Unveiling Anti-Architectural Patterns in...
08mLightning talks – 101 Things Software Never Learnt
01h 00mEvolutionary Architectures
42mSoftware Architecture in the Age of Agility and DevOps
49mGetting a handle on software architecture