DevOpsDays NZ 2019 – Day 2 – Session 3

Everett Toews – Is GitOps worthy of the [BuzzWord]Ops moniker?

  • Usual Git workflow
  • But it takes some action
  • Applying desired state from Git
  • Example: Infrastructure as code
    • DNS
    • Onboarding and offboarding
  • Git is now a SPOF
  • Change Management Dept is now a barrier
  • Integrate with ITSM
  • Benefits: Self-service, Compiience

Joel Wirāmu Pauling – Why Bare Metal still maters

  • Cloud Native Dev doesn’t exist as a closer system
  • IoT is all hardware
  • AI/ML is using special hardware
  • Networks is all hardware offloads
  • FPGAs and ASICS need more standard open way to access
  • You’ll always have weird stuffs on your network
  • Virtualization has abstracted away the real
  • We care able vendor lockin with cloud APIs and Aus electricity isn’t all that green

Steven Ensslen – Do you have a data quality problem?

  • What is data ops and why do we want it?
  • People think they have a data quality problem but they don’t actually measure it to see how bad.
  • Causes all sorts of problems.
  • 3 Easy steps to fix data quaility
  • 1 – Document data charactersistics and train people to know them
  • 2 – Monitor data as if it is infrastructure
    • Test data like it is code
  • 3 – Professionalize your support of data professionals
    • Bring in the spreadsheet experts
    • Support reporting and analytics people too

Mandi Buswell – What are Kubernetes Operators and Why do I care

  • Like an App Store on your kubernetes cluster
  • Like a like Kubernetes robot doing that hard work for you. Lifecycle management
  • Operators run as microservices on the kubernetes cluster
  • operatorhub.io
  • Work on any kubernetes cluster
  • You can even write your own

Laura Bell – Securing the systems of the future

  • Fear and Lothing
    • It is an old problem because “People are Jerks”
  • All organization try either Fight, Flight, Freeze
  • Trying to protect: Confidentiality, Integrity, Availbality
  • Protect, Detect, Respond
  • Monolith
    • A big wall around
    • Layered defense is better but not the final solution
    • Defensive software architecture is not just prevention
    • Castles had lots of layers of defenses. Some prevention, Some Detection, Some response
  • MIcroservices
    • Look at something in the middle of a star and erase it
    • Push malicious code into deployment pipelines
  • Avoid scar tissue, stuff put in just to avoid specific previous problems. Make you feel safe but without any real evidence.
  • Fearless security patterns and approaches
  • Technology is changing but the basics are still the same
  • Lots of techniques in computer security.
  • Prevention and Detection are interchangeable
  • Batman vs Meercat model
  • Be Aware and challenge your own bubble
  • Supply Chains are vulnerable: Integrations, dependencies, Data Sources
  • Determinate threat vs Dynamic Threat
    • Can’t predicts which steps in which order are going to get the result
    • Comprimise the data then the engine will return bad results
  • Plug for opensecurity.nz

Share