2 - Compute In The Cloud Flashcards
EC2 (elastic compute cloud)
Secure, resizable compute capacity in the cloud
Multilatency
Sharing underlying hardware, secure and separate from others
EC2 process
1 launch
2 connect
3 use
Types of EC2 instances
General purpose - provide balanced resources, diverse workloads
Compute optimized - compute intensive like batch processors
Memory optimized - memory intensive like databases
Accelerated computing - calculations or data patterns
Storage optimized - data warehouse
Pricing
On demand - used in the beginning to play around, pay for duration of instance running
Savings plan - consistent amount of compute, 1/3 year, 72% savings
Reserved instances - steady workload, 1/3 year, 75% savings
Spot instances - flexible start/stop time, can withstand interruptions
Dedicated hosts - compliance requirements
Auto scaling
Remove instances when not needed
Elastic load balancer
Even distribution of work, regional construct
Tightly coupled/monolithic
One component fails causes whole app to fail
Loosely coupled/micro services
One instance fails it is isolated to not cause other failures
Simple queue service
Send store and receive messages between components without losing
Simple notification service
Send messages to services or notifications to end users
Lambda
Upload code and it’s automatically run, pay for time you consume, serverless
ECS
Containerized applications at scale, can run on top of EC2
EKS
Kubernetes
Fargate
Serverless env managed for you
