Let's not talk about how many times I had to respell Architecting.
Most Secure, High performing, resilient, efficient
Consistent for evaluating architectures.
Quicker Build and Deploy.
Stop guessing, start automating.
Lower or Mitigate Risks
- What are they?
- Do something about it.
- Cause and Effect.
- Not Affect.
Check ‘em out at aws.amazon.com/architecture/well-architected
- Monitor things, deliver better value.
- Operations as code (scripting, automation)
- Automate documentation (Wow then why are humans here again?)
- Tiny changes for tiny reversals if needed.
- Refine processes
- You will fail. It’s okay. But you can mitigate or remove where you might fail.
- Most important is that you learn from it.
Protect Systems, information, and assets.
- Strong identity foundation
- Least privilege
- Seperation of duties.
- authorize sections of your AWS resources properly
- audit, audit, audit.
- And logs.
- And metrics.
- In-depth security layering, not just the outer wall, so to speak.
- Automate security (lot of automation)
- Is your data encrypted at move and at rest? It should be. Also, don’t let people touch it. No one should be editing it at the home base.
- Run incident response sims.
Is it up? And if it’s down, how long will it take to be up again? What happened to data in the process?
- Test in the cloud to validate recovery procedures.
- Monitor KPIs to trigger automation and auto notification about failures and to kick in auto recovery processes.
- Horizontal scaling.
- Out of many, One. But instead, Kill one to make many for better redundancy.
- Don’t guess capacity. Monitor the demand and system resources used.
- Let automation change your infrastructure.
- Deployment. Mass Deployment.
- Take programming (I THINK it’s programming, it mentions NoSQL and Databases) into services.
- Go serverless (Though we know there re still servers)
- Have fun and experiment in a non-work enviroment of course.
- Pay as you go, so use as much or as little as you can afford.
- Test it out first. Analyze what you need.
- Stop them when you’re not using them.
- Measure efficiency.
- Don’t pay for data centers