Serverless

Front 1

What use case is serverless trying to solve?

Back 1

• Underutilised computer capacity
• Having to patch and deploy
  

Front 2

What is the essence of how it works?

Back 2

You only write the functional part of the app, not the server itself

Front 3

When would I NOT use Lambda?

Back 3

• When the goal is explicitly to avoid making any changes to anapplication - Lambdas need some changes.
• When you have code/infrastructure that won't work when the implementation details are changed (think about loss of control when using managed services)
• When an on-premises application hasn’t been migrated to the publiccloud.

Front 4

How it a serverless application typically split into components?

Back 4

Lambdas are like your use cases, doing the logic and the other components such as message queues and databases are separate to this.

Front 5

How does data come into a lambda?

Back 5

Via events

Front 6

What is AWS SAM and what does it do?

Back 6

Serverless Application Model - an open spec for specifying the serverless app and its components

Front 7

How does it flow out?

Back 7

Through streams and data sources

Front 8

What is the purpose of orchestration with serverless?

Back 8

Chaining together actions that are short lived into something more useful

Front 9

What limitations of stateless does orchestration cover us for?

Back 9

• Lambdas themselves are stateless but we might want two consecutive lambda to pass state around
• A work flow outside of the lambda itself may run for longer than any single lambda
• Error handling might be better handled with orchestration code.
  

Front 10

What are edge locations in serverless?

Back 10

Locations where the end users access AWS services. Usually fronted by a CDN and located in major cities

Front 11

What 2 technologies does Serverless have for Orchestration?

Back 11

AWS Step Functions and CloudWatch Events

Front 12

How is deployment of Lambdas done?

Back 12

Cloudformation

Front 13

How is monitoring done?

Back 13

CloudWatch

Front 14

Which technologies handle interprocess messaging?

Back 14

SNS and SQS

Front 15

Which streaming technologies are there?

Back 15

Kinesis

Front 16

Which storage mechanisms are there?

Back 16

S3 and Dynamo DB

Front 17

What does the compute itself?

Back 17

AWS Lambda or Lambda Edge

Front 18

How is it that Lambda scales perfectly with workload?

Back 18

Because each event just runs another copy of the function, which is stateless. The event is the data, the function is the code.

Front 19

What is the definition of Function as a Service, or FaaS?

Back 19

The function is the unit of deployment and execution

Front 20

What is Serverless Faas?

Back 20

Where the programming model never specifies machines or containers