Upgrade to Pro — share decks privately, control downloads, hide ads and more …

The IT multiverse - A tourist guide with inside...

The IT multiverse - A tourist guide with insider tips

This talk was created as a keynote for a conference. Thus, it is meant rather as food for thought and offering some different point of view than concrete advice how to do things.

It starts with a journey through the "IT multiverse", visiting several types of IT organizations - first, along the evolution path of enterprise IT, then visiting startups and those who became successful and finally several bordering universes.

While the ideas found on that round flight can be used for several (often strategic) purposes, I use it here for a more limited use case, which is trying to understand where an individual is, where (s)he wants to go and how to get there.

In order to reason about those questions, I create a little model that tries to extract the general type of IT organization model (including the forces that act on it), its viability and typical evolution paths. This model can then be used to ponder the questions mentioned before (without overdoing it).

Finally, I offer a few more personal tips that might be useful on your way through the crazy, but fascinating multiverse called IT.

As always, unfortunately the voice track is missing (and there is no recording of the talk yet), but hopefully the slides in themselves also offer you a bit food for thought ... ;)

Uwe Friedrichsen

November 19, 2019
Tweet

More Decks by Uwe Friedrichsen

Other Decks in Technology

Transcript

  1. The IT multiverse A tourist guide with insider tips Uwe

    Friedrichsen – codecentric AG – 1993-2019
  2. Ad hoc • Development on demand • Peer to peer

    communication • No clear processes • No clear practices (“Whatever goes”) • Lots of manual work • Lots of DIY development (Do it yourself)
  3. Software development as … … an activity driven by a

    comprehensive software development process
  4. Process driven • Development as a division-of-labor activity • Highly

    regulated communication • Detailed processes • Hierarchical organization • Some automation, manual work defined and controlled • Standard Software and Enterprise Frameworks
  5. Outsourcing • IT not considered differentiating (“cost center”) • Expensive

    and error-prone (see chaos report) • Outsourcing to specialized IT suppliers • Expectations often not met
  6. Actual Agile • Collaborative software development • Inspect and adapt

    • Continuous delivery • Continuous feedback • Continuous learning • Continuous improvement • Usually working for selected small teams only
  7. Cargo cult “Agile” • Gaps in Scrum filled up with

    existing process and roles • Usually ignored outside the development team • Inspect and adapt cycle not implemented • No continuous delivery • No or little feedback • No or little learning • No or little improvement
  8. Scaled “Agile” • SAFe won the race • Addresses enterprise’s

    yearning for perceived certainty • Turned cargo-cult Agile into non-Agile • Process-driven development in a new disguise • Other, more agile approaches available (e.g., SWARM *) • Will remain niches at best * https://dannorth.net/2018/01/26/in-praise-of-swarming/
  9. Startup 1.0 • IT-intensive business offering • Needed to go

    fast • Used a minimal enterprise IT stack • Focus on fast feedback
  10. DevOps 1.0 • Needed to keep going fast while (hyper-)scaling

    business • DevOps on the organizational and governance level • Loosely coupled cross-functional capability teams • Microservices as supporting architecture style
  11. CrapOps / NoOps • Result of multiple misunderstandings • Mimicked

    DevOps instructions without understanding ideas • Copied technology stack without understanding drivers • Filled gaps with old corporate habits • Filled gaps with old corporate habits • “We need to control everything centrally!” • “How can we use it to save money/lay off people?” • Made companies neither faster nor better nor more scalable • IT additionally drowns in non-understood tech stack
  12. Ad hoc Outsourcing Startup 1.0 Process driven Scaled “Agile” Actual

    Agile DevOps 1.0 CrapOps/ NoOps Cargo cult “Agile”
  13. Software development as … … collaborative activity of multiple loosely

    coupled capability teams supported by platform teams
  14. DevOps 2.0 (w/ platform services) • Redundant tasks picked up

    by platform teams • Platform teams act as internal service providers • Capability teams still have decision authority • Emerging platform self-service marketplace
  15. Ad hoc Outsourcing Startup 1.0 Process driven DevOps 2.0 Scaled

    “Agile” Actual Agile DevOps 1.0 CrapOps/ NoOps Cargo cult “Agile”
  16. Startup 2.0 (NIH infested) • Wannabe Netflix syndrome • Infatuated

    with technology • Strong NIH syndrome (Not invented here) • Strong DIY syndrome (Do it yourself) • Bloated IT departments
  17. Ad hoc Outsourcing Startup 1.0 Startup 2.0 Process driven DevOps

    2.0 Scaled “Agile” Actual Agile DevOps 1.0 CrapOps/ NoOps Cargo cult “Agile”
  18. Software development as … … collaborative activity of multiple loosely

    coupled capability teams, going cloud native all-in
  19. DevOps 3.0 (cloud native all-in) • Leveraging managed services on

    all levels • Integrating using FaaS technology • Only building differentiating logic on their own • Actively reducing vertical integration depth • Deliberately moving right on a Wardley map * • Going faster without compromising quality * https://medium.com/wardleymaps/on-being-lost-2ef5f05eb1ec
  20. Ad hoc Outsourcing Startup 1.0 Startup 2.0 Process driven DevOps

    2.0 Scaled “Agile” Actual Agile DevOps 1.0 CrapOps/ NoOps DevOps 3.0 Cargo cult “Agile”
  21. Startup 3.0 (Use over make) • Adopt cloud native all-in

    (managed services) • Go for low code/no code environments where possible • Maximize reduction of vertical IT integration depth • Focus on customer experience, not IT impressiveness
  22. Ad hoc Outsourcing Startup 1.0 Startup 3.0 Startup 2.0 Process

    driven DevOps 2.0 Scaled “Agile” Actual Agile DevOps 1.0 CrapOps/ NoOps DevOps 3.0 Cargo cult “Agile”
  23. IoT (Internet of Things) • A.k.a. the wild, wild west

    of IT • Widely heterogeneous, not yet settled landscape • Disruptive change in IT • Business models still emerging • Fascinating, but very volatile domain
  24. Ad hoc Outsourcing Startup 1.0 Startup 3.0 Startup 2.0 Process

    driven DevOps 2.0 Scaled “Agile” Actual Agile IoT DevOps 1.0 CrapOps/ NoOps DevOps 3.0 Cargo cult “Agile”
  25. Data Science (incl. AI/ML/DL) • Important complementing aspect of modern

    IT • Often a bit decoupled from normal software development • Different mindset (exploration first) • Math often frightening for average developer • Needs to be integrated better in normal IT activities
  26. Ad hoc Outsourcing Startup 1.0 Startup 3.0 Startup 2.0 Process

    driven DevOps 2.0 Scaled “Agile” Actual Agile IoT DevOps 1.0 CrapOps/ NoOps Data Science DevOps 3.0 Cargo cult “Agile”
  27. Scientific computing • We do not really understand what they

    are doing • Looks interesting, but not actually applicable for us • Quite loosely connected to commercial computing • We decide to ignore them – and hope it’s safe … ;)
  28. Ad hoc Outsourcing Startup 1.0 Startup 3.0 Startup 2.0 Process

    driven Scientific computing DevOps 2.0 Scaled “Agile” Actual Agile IoT DevOps 1.0 CrapOps/ NoOps Data Science DevOps 3.0 Cargo cult “Agile”
  29. Industrial computing • Focus on embedded computing for many years

    • IIoT acts as a game changer • Gets closer to traditional IT at high speed • Multi-billion dollar/euro market emerges • Still mostly unnoticed/ignored by traditional IT
  30. Ad hoc Industrial computing Outsourcing Startup 1.0 Startup 3.0 Startup

    2.0 Process driven Scientific computing DevOps 2.0 Scaled “Agile” Actual Agile IoT DevOps 1.0 CrapOps/ NoOps Data Science DevOps 3.0 Cargo cult “Agile”
  31. Ad hoc Industrial computing Outsourcing Startup 1.0 Startup 3.0 Startup

    2.0 Process driven Scientific computing DevOps 2.0 Scaled “Agile” Actual Agile IoT DevOps 1.0 CrapOps/ NoOps Data Science DevOps 3.0 Cargo cult “Agile”
  32. … but let us stop here and try to use

    the knowledge we have discovered so far
  33. Ad hoc Industrial computing Outsourcing Startup 1.0 Startup 3.0 Startup

    2.0 Process driven Scientific computing DevOps 2.0 Scaled “Agile” Actual Agile IoT DevOps 1.0 CrapOps/ NoOps Data Science DevOps 3.0 Cargo cult “Agile”
  34. Therefore let us look at the universes from a bit

    different angle drivers – forces – viability - paths
  35. Formal part of value creation Solution: machine Dynamic part of

    value creation Solution: man sluggishness/low dynamic high dynamic high dynamic The historical course of market dynamics and the recent rise of highly dynamic and complex markets The dominance of high dynamics and complexity is neither good nor bad. It‘s a historical fact. t 1970/80 today Age of crafts manu- facturing Age of tayloristic industry Age of global markets 1850/1900 Spacious markets, little competition Local markets, high customi- zation Outperformers exercise market pressure over conventional companies We call the graph shown here the “Taylor Bathtub”. The “bathtub” curve Source: BetaCodex Network Associates, “Organize for complexity”, BetaCodex Network White Paper 12 & 13
  36. Key drivers Pre-industrial era • No clear driver • “Whatever

    pays the bill” Industrial era • Cost-efficiency • Scalability • Repeatability • Stability • Efficiency & scale Post-industrial era • Cycle times • Adaptability • Flexibility • Resilience • Effectiveness & speed
  37. Properties Pre-industrial Industrial Post-industrial Market Emerging market Wide & slow

    market Narrow & fast market Strategic goal None (“Get the bill paid!”) Cost-efficiency Short cycle times System theory Obvious Complicated Complex Work model Ad hoc Processes Collaboration Organization model Peer to peer Hierarchies Autonomous teams Practices “Whatever goes” ITIL, V-Modell, SAFe DevOps, Kanban, XP, Agile Individuals Experts (Cube-shaped) Specialists (I-shaped) Generalists (T-Shaped) Automation Low Medium-high Very high Manual work Ad hoc Defined and controlled Avoided Programming model Do it yourself Standard Software, Enterprise Frameworks Cloud-native, Serverless, Managed Services
  38. Good Fast Cheap Optimizing for quality and cycle times will

    result in higher costs Optimizing for quality and costs will result in long cycle times Optimizing for cycle times and costs will result in reduced quality
  39. Pre-industrial No clear driver Good Fast Cheap Industrial Deliver large

    batches at minimized costs towards slow contexts Post-industrial Quickly adapt to ever-changing needs of dynamic, fast-moving contexts Startup Test hypotheses and pivot as fast as possible to discover a product-context fit
  40. Now we are ready to look at the universes from

    a bit different angle drivers – forces – viability - paths
  41. Ad hoc • Pre-Industrial • Suitable for companies with little

    IT demand • Truck factor typically < 3 • Does not scale • Usually develops into process-driven IT
  42. Process driven • Industrial • Predominant in larger companies •

    Self-reinforcing system • Strong attractor – requires energy not to drift into • Requires huge amount of energy to escape from • Usually terminal state • Sometimes disguised as “Agile” (à scaled “Agile”) or alike
  43. Outsourcing • Industrial • Very fashionable IT strategy 10 -

    20 years ago • Valid option for non-differentiating parts of IT • Okay, if lead time for changes are irrelevant • Yet, often a result of a general misperception of IT by decision makers who do not really understand IT • As a general IT strategy, usually not sustainable
  44. Actual Agile • Post-industrial • Addresses needs of highly dynamic

    environments • Rather a developer ideal than IT reality • Needs protected environment to survive • E.g., strong leader shielding the team from outer forces • Usually assimilated by process-driven IT (cargo cult “Agile”)
  45. Cargo cult “Agile” • Industrial • Result of corporate IT

    assimilating actual Agile • Implementing rituals without understanding ideas • Often just "group therapy for frustrated developers" • Overall usually no real improvements realized • Will eventually drift into process-driven IT or scaled "Agile"
  46. Scaled “Agile” • Industrial • Result of scaling cargo cult

    “Agile” to corporate level • Can help larger corporates to add a tiny dose of flexibility into their overly rigid processes (there is value in doing this) • Not agile, but process-driven IT in a disguise • Usually terminal state • Sometimes drifts back to pure process-driven IT
  47. Startup 1.0 • Startup • Vertical IT integration depth needs

    to be considered • Will be superseded by startup 3.0 as future default • Still, will remain as a niche • With growth, usually develops • either into process-driven IT • or into DevOps 1.0
  48. Startup 2.0 (NIH infested) • Startup • Very deep vertical

    IT integration depth • Okay for unicorns with unique problems • Not sustainable for everyone else • Either evolution to Startup 3.0 or extinction over time • With growth, usually develops • either into process-driven IT • or into DevOps 1.0/2.0
  49. Startup 3.0 (Use over make) • Startup • Most up-to-date

    implementation of Startup • Enables extremely short cycle times at very low cost • With growth, usually develops • either into process-driven IT • or into DevOps 3.0
  50. DevOps 1.0 • Post-industrial • Effective, but not very efficient

    • Lots of duplicate work in different teams • Usually evolves into DevOps 2.0/3.0 over time
  51. DevOps 2.0 (w/ platform services) • Post-industrial • Brings a

    healthy dose of efficiency back into DevOps • Must not be confused with Ops 0.5 • Platform services as marketplace, not as prescriptions • Sustainable model • Still, often energy required to fend off industrial forces • Drift towards DevOps 3.0 should be considered
  52. DevOps 3.0 (cloud native all-in) • Post-industrial • Most up-to-date

    implementation of DevOps • Enables very short cycle times without compromising quality • Sustainable model • Still, often energy required to fend off industrial forces
  53. CrapOps / NoOps • Industrial • Cargo cult – implementing

    rituals only • Often augmented with lots of new technology • Overstrained teams as typical effect • Less speed, more errors • Not sustainable • Usually drifts back to process-driven IT or scaled "Agile"
  54. IoT (Internet of Things) • Often pre-industrial or startup •

    Complementing universe • Currently brings more volatility to the connected universe • Will become more stable with settling of IoT landscape • Will cease to exist as a distinct universe over time • Instead will become integral part of other universes
  55. Data Science (incl. AI/ML/DL) • Pre-industrial most of the time

    • Complementing universe • Brings more trial and error attitude to connected universes • Thus, often experiences repulsive force from industrial IT • Will cease to exist as a distinct universe over time • Instead will become integral part of other universes
  56. Scientific computing • Not really connected to the commercial IT

    universes • Only accidental connections from and to this universe
  57. Industrial computing • Industrial • Target IT market with growing

    relevance • Widespread conservative technology adoption patterns • Very affine to standard software solutions • Especially post-industrial IT will need to learn to respond to it • Sustainable model • Will at least partially merge with other universes over time
  58. Key Model Extra details Sustainability Model Extra details Sustainability Pre-industrial

    Industrial Post-industrial Startup Sustainable Partially sustainable Sustainable only in niches Not sustainable Usually drifts into different universe over time Terminal state Not viable in the long run Will merge with other universe Unstable (will drift forth and back) Requires protection Mostly unknown
  59. Ad hoc Industrial computing Outsourcing Startup 1.0 Startup 3.0 Startup

    2.0 Process driven Scientific computing DevOps 2.0 Scaled “Agile” Actual Agile IoT DevOps 1.0 CrapOps/ NoOps Data Science DevOps 3.0 Cargo cult “Agile”
  60. Moving away from process driven • Extremely hard to overcome

    the inertia • It’s not the individuals, it’s systemic reinforcement • Easier to change the job than to change the company • Sometimes you can establish a protected niche • Still requires continuous efforts to protect your niche from the company’s immune system that will fight it all the time • Usually only works with upper level management support • Focus on culture, not processes or tools
  61. Evolving from startup • Organizational growth leads to new challenges

    • Typical response: Process, hierarchy, bureaucracy • Inevitably leads to process-driven IT • Alternative response: Beta organization • Leads to DevOps • The second path is a lot harder (but worth the effort)
  62. How do I become actually Agile? • Aim for DevOps

    (in its original meaning!) • DevOps provides better framing for IT than, e.g., Scrum • Creates relentless focus on value creation, delivery, feedback (from where value is realized) and continuous improvement • Working Agile environment emerges as a “side effect”
  63. Ad hoc Industrial computing Outsourcing Startup 1.0 Startup 3.0 Startup

    2.0 Process driven Scientific computing DevOps 2.0 Scaled “Agile” Actual Agile IoT DevOps 1.0 CrapOps/ NoOps Data Science DevOps 3.0 Cargo cult “Agile” This is only a model
  64. Technology • Go cloud native – use managed services •

    DIY is fun, but takes too long and does not create value • Focus on the frontend • That is where the battle for the customer is won today • Mind the IoT train • Accept data as the currency of the 21st century • Do not forget security
  65. Take it easy Go your own way and do not

    take things too seriously