Published

(compare with Model_ AWS Cloud Value Framework) (compare with Model_ 4 benefits of cloud computing) Cloud Computing Benefits variable expenses (vs upfront expenses) cost optimization (don’t invest in maintaining data centers) capacity (peak use vs rest of year) economies of scale (AWS buys a lot, can get that cheaper) Speed and agility (time between wanting resources vs having resources) Global in minutes (1) variable expenses upfront expenses - invest before using variable expenses - pay what you use (2) cost optimization invest...

cloud deployment on-premise deployment hybrid deployment (1) cloud deployment run full app in the cloud migrate to the cloud build new apps (2) on-premise deployment use virtualization and resource management tools to deploy resources use app management and virtualization to increase resource usage (3) hybrid deployment connect cloud with on-premise integrate cloud with legacy

Data Ownership Data Access Data Ownership the owner is the one who writes to the table Model_ 3 data ownership patterns (in microservice architecture) Data Access Model_ 4 data access patterns (in microservice architecture)

what if we simplify each to binary: we get 8 permutations (1) Sagas dimensions of change Model_ 8 types of sagas hybrid sagas (2) sharding example: concert tickets example: sysops experts (1) Sagas dimensions of change Any change to any of these 3 forces, changes everything… treat as binary for simplicity then we get: Model_ 8 types of sagas epic saga fantasy fiction saga fairy tale saga parallel saga phone tag saga horror story saga time travel saga anthology saga !...

if the semantics of your problem change then everything else needs to change prevent implementation coupling, that makes your architecture brittle contracts: strict vs loose method calls “schema” serialized objects RPC remote procedure call graphql (aggregation framework, not a invocation framework) value-driven contracts json KVP key value pairs example: brittleness not brittle: changing the type of country, does not break the contract for wishlist support all versions forever?...

Model_ 5 service disintegrators Model_ 3 granularity integrators

What is async? async implies events or messages Distinction_ events vs messages Why async? improves responsiveness (not performance - Distinction_ responsiveness vs performance) async usually works well in happy path, but get’s more complicated in the unhappy path… happy path -> works well unhappy path: eg.“comment uses a word that is not allowed” new user flow: send email? notification to user? automate correction censor that word *****

orchestration: conductor at musical orchestra choreography: dancers Orchestration single point of failure services depend on orchestrator lots of communication order placement needs to be highly available orchestrator needs to be highly available order placements needs to be highly available smaller orchestrator (vs Enterprise Service Bus) eg. Order Placement Orchestrator, Order Cancellation Orchestrator logic shifts to the orchestrator error handling becomes a design decision in the orchestrator? in the service?...

Architecture: The Hard Parts How do I determine the appropriate level of service granularity? service granularity How do I achieve high semantic coupling and low implementation coupling? semantic vs implementation coupling How do I create bounded contexts and access data I don’t own? bounded contexts and access data I don’t own How do I decide between synchronous versus asynchronous communication? synchronous vs asynchronous communication How do I choose between choreography and orchestration for my workflow?...

Model_ AWS Cloud Value Framework 1. Cost Savings Migration Challenges migration costs shifting to opex cloud readiness entrenched IT org 2. Staff Productivity 3. Operational Resilience 4. Business Agility code throughput (1) Lead Time (2) Deployment Frequency system stability (3) Change Failure Rate (4) Mean Time To Recover