Articles
where to start? what is my approach to architecture? “do not try to boil the ocean.” dont think of everything at once. dont try to solve it at once. start high-level what is your business? what do you do? audience – who are customers? what are thy getting from you? how are they hoping to benefit by using your services? customer consumption – how are customers consuming or accessing info from your system? technical: ingestion- how is data ingested processing- how is processing occurring? storage- how is customer ready being stored need to understand pillars and their priorities pillars are: operational excellence security reliability performance efficiecy cost design will change based on prioritization of “pillars” e.g. high security system possible components: web app firewall, tokenization, encryption, cust managed keys, IP traffic monitoring and constraints — Stream processing example: 1. Kinesis – to load in social media stream real time, then trigger lambda to generate hastag trend data 2. Dynamo DB – store hashtag data, trend data for compilation of analytics reports — Weather service web app example: 1. S3 – to host front end code 2. Amazon API Gateway – make REST API call to endpopint 3. trigger lambda to retrieve weather info 4. dynamodb – store weather data used by application — IOT: 1. a farm tractor sends data to amazon through amazon API? 3. Kinesis – captures data and streams itr to lambda for processing 3. lambda runs code to detect trends in sensor data, identify anomalies, order replacement parts 4. dynamodb – to store raw signals then sends to kinesis for processing amazon examples: three tier retail/ecommerce analytics — retail site possible architectural components- 1. consumption – use APIs and front end (possibly EC2 for front end server) 2. processing front and backend storage – 1 EC2 instance for back end, then multiple databases, 1 for product info (1 amazon aurora db) 1 for trans info (1 amazon aurora db) 3. ingestion – some batch processes, real-time product loading (API) 4. if is public retail product specs perhaps heavy security not needed —
video on AWS architecture
