SRE classroom: How to design a reliable application in three hours

Description

Explore the key concepts behind large system design with Jenny Liao, as she guides you through building, scaling, and provisioning a system. Apply the concepts you learn to evaluate and build systems of your own. You will be working in small groups.

Key Concepts:

• Microservices
• Failure tolerance
• CAP theorem
• Consensus in distributed systems
• SLOs
• Load balancing
• Capacity Planning

Part one: Workshop introduction: goals, expectations, and problem statement

We have a problem. Let’s solve it with software.

• Setting the tone: goals and expectations
• Present the problem statement
• Introduction of the service-level objectives (SLOs)
• Discussion of terminology and concepts so we’re using the same vocabulary

Hands-on workshop, part one (40 minutes)

• Expectation: Identify the components necessary to build a working system in a single location. Produce a sketch of this working system.
• Unified Modeling Language (UML) not required
• Afterwards: 5 minute break
• After the break, sample solution

Part two: The solution has limitations; let’s improve it.

We have identified single points of failure…because things fail. The system failed. And we lost users.

Let’s replicate this thing.

• What parts are useful to replicate?
• How do we arrange this so that work is shared efficiently?
• How do we know that these systems are doing what we expect?

We have correctness problems.

• How do we identify correctness or consistency issues?
• How do we know if we have addressed them?
• How can we apply these concepts to a real piece of software?
• What limitations does this introduce?

Hands-on workshop, part two (30 minutes)

• Expectation: Identify which components can usefully run with multiple replicas and/or in multiple locations. Produce a system that runs correctly in multiple data centers.
• Afterwards: 5 minute break
• After the break, sample solution

Part three: Provisioning, SLOs, and conclusions
Let’s provision the system based on storage, bandwidth and latency needs. Let’s evaluate our system against the SLO requirements.

We have a map of components in our system.

• How many replicas do we need for each component?
• Is the system reliable enough?
• Does the system give correct results?
• How much hardware do we need?
• Are there any bottlenecks? How can we address them?

Hands-on workshop, part three (35 minutes)

• Expectation: Identify how many machines you need for each component of the system. Determine if the SLOs are reasonably achievable with your design.
• Afterwards: 5 minute break
• After the break, sample solution

Part finale: discussion and conclusions

• What key points have we learned?
• How can we apply these key points beyond this workshop?
• Last thoughts

Hands-on workshop exercises:
For each exercise, you’ll work in small groups to make incremental progress on your solutions. Jenny will present a sample solution with distributed systems design concepts and key takeaways after each hands-on portion of the workshop.