Application Design Framework (ADF)
High-level application design from use case to code
This example focuses on establishing application boundary and mapping architecture to code. Requirements are out of scope. Note: It is not necessarily the actual Amazon EC2 design.
Working backwards
Company wants to build Amazon EC2 product. The first use case is to allow customers launch instances.
Stories and flows
First story is “As a developer, I want to launch an instance”. Flow describes developer experience launching an instance using AWS Management Console.
Architecture
Application boundaries
Context
We need to identify application boundaries by describing stories and flows on architecture level.
Decision
The decision is to create two applications: EC2 Instances Console and EC2 Instances Control Plane.
EC2 Instances Console components
Context
We need to identify EC2 Instances Console components.
Decision
EC2 Instances Console application includes Toolchain and Service components. Toolchain includes Deployment Pipeline and Pull Request Build components. Service includes Network, Ingress, Compute, WebApp, Database and Monitoring components. Components use infrastructure services and runtime code to provide functionality. Application (metadata), Toolchain and Service resources deploy as a stack each.
EC2 Instances Control Plane is out of scope.
Code structure
EC2 Instances Console
The example uses AWS Cloud Development Kit (AWS CDK) pseudo code for infrastructure services. It should be possible to use the same approach with other infrastructure automation tools.
service/
webapp/
Dockerfile
app.py
instances.py
compute.py
class Compute(Construct):
ec2.SecurityGroup
ecs.Cluster
ecs.Service
ecs.TaskDefinition
ecr_assets.DockerImageAsset
database.py
class Database(Construct):
ec2.SecurityGroup
elasticache.CacheCluster
ingress.py
class Ingress(Construct):
ec2.SecurityGroup
elasticloadbalancingv2.NetworkLoadBalancer
elasticloadbalancingv2.TargetGroup
certificatemanager.Certificate
route53.HostedZone
route53.CNAMERecord
waf.WebACL
monitoring.py
class Monitoring(Construct):
cloudwatch.Metric
cloudwatch.Alarm
cloudwatch.Dashboard
xray.Group
network.py
class Network:
ec2.VPC
ec2.Subnet
ec2.RouteTable
service_stack.py
class ServiceStack(Stack):
compute.Compute
database.Database
ingress.Ingress
monitoring.Monitoring
network.Network
toolchain/
deployment_pipeline.py
class DeploymentPipeline(Construct):
pipelines.CodePipeline
pull_request_build.py
class PullRequestBuild(Construct):
codebuild.Project
toolchain_stack.py
class ToolchainStack(Stack):
toolchain.deployment_pipeline.DeploymentPipeline
toolchain.pull_request_build.PullRequestBuild
app.py
service.service_stack.ServiceStack("EC2InstancesConsole-Service-Sandbox")
toolchain.toolchain_stack.ToolchainStack("EC2InstancesConsole-Toolchain-Sandbox")
toolchain.toolchain_stack.ToolchainStack("EC2InstancesConsole-Toolchain-Management")