11. Serverless II

Other Serverless AWS Services

AWS SQS

AWS SQS(Simple Queue Service)

• fully-maanged message queueing service that stores messages
• messages are stored(but will expire) until they are processed and deleted
• act as a buffer btw senders and receivers
• securely share SQS queues anonymously or with specific accounts
  • queue sharing can also be restricted by IP address and time of day
• msgs are stored in encrypted form, and SQS decrypts msg only when they are sent to an authorized consumer

Benefits of Amazon SQS

• requires no administrative overhead and little configuration for message queueing
• decouple processing steps from compute steps and postprocessing steps
  • using asynchronouse processing -> separate producer logic & consumer logic
• makes the system more resilient with high volume capacities
  • acts as a buffer to absorb spikes in traffic
  • gives your application additional time to complete scale-out actions
• built-in fault tolerance
  • if exception or transaction failure -> order processing can be retried
  • after the maximum number of retries, SQS can redirect the msg to SQS dead-letter queue where you can reprocess or debug it later

Achieving Loose Coupling with SQS

• Use asynchronous processing to get responses from each step quickly
• Handle performace and requirements by increasing the number of job instances (worker instances)
• Messages will remain in the queue

Amazon SQS Queue Types

• Standard Queue
  • at-least-once delivery: occasionally more than one copy of a msg is delivered
  • best-effort ordering: occasionally msgs might be delivered in out-of-order
  • nearly ultimated number of tx per second (TPS) per API action
• FIFO Queue
  • guarantee that msgs are processed exactly once, in the exact order that they are sent
  • support up to 300 msgs per second
  • 순서가 중요하고 중복이 허용되지 않는 상황에서 사용

Use Cases

• Work queues
  • decouple components of a distributed aplplications
• Buffering and batch operations
  • add scalability and reliability to your architecture
  • smooth out temporary volume spikes without losing msgs or increasing latency
• Request offloading
  • enqueue slower erquests
• Auto scaling
  • determine the load on an application
  • can combined with auto scaling

 

AWS SNS

AWS SNS (Simple Notification Service)

• web service that makes it easy to set up, operate, and send notifications from the cloud
• follows the publish-subscribe messaging paradigm, with notifications beling delivered using a push mechanism
• create a topic and control access to it by defining policies that determine which publishers and subscribers can communicate with the topic
  • each topic has a unique name that identifies that SNS endpoint
  • subscribers receive all msgs published to the topics to which they subscribe

Amazon SNS Subscription Types

• Email
  • Email-JSON sends notifications as JSON object, while email sends text-based email
• Mobile text messaging (SMS)
  • sent to phone numbers as SMS text msg
• HTTP/HTTPS
  • notifications are delivered through an HTTPS POST to the subscriber's specified URL
• Amazon SQS
  • enqueue a notification msg to specified SQS standard queue
• AWS Lambda
  • msgs can be delivered to AWS Lambda functions

Amazon SNS Characteristics

• A single published msg
  • notification msg contain a single published msg
• No recall options available
  • when msg delivered successfully, no way to recall it
• Retries take place with HTTP or HTTPS
  • SNS delivery policy can control the retry pattern
  • provide access control mechanism to ensure topic and msgs are secured against unauthorized access
• Order and delivery isn't guaranteed
  • to prevent msgs from being lost, all msgs published to SNS are stored redundantly across multiple servers and data centers

Use Cases

• receive an immediate notification when an event occurs
• push targeted news headlines to subscribers
• notifications to app indicating that an update is available

Amazon SNS and SQS Difference -표 15p

Features	Amazon SNS	Amazon SQS
Message persistence	No	Yes
Delivery mechanism	Push(passive)	Pull(active)
Producer and consumer	Publisher and subscriber	Send or receive
Distribution mode	one-to-many	one-to-one

 

Serverless Characteristics of S3

Serverless Characteristics of S3

• Pay for value : pay for what you use
• Elastic storage: storage capacity grows and shrinks automatically
• Continuous scaling: fully scalable
• Fault tolerance: high data durability of objects, high availability
• No server management

 

Lambda@Edge

Lambda@Edge

• feature of Amazon CloudFront that lets you run code closer to the users of your application, which improve performance and reduces latency
  • you don't have to provision or manage infrastructure in multiple locations
• pay only for the compute time that you consume
• making your application globally distributed and improving their performance
• zero server administration

 

API Gateway

API Gateway

• With API Gateway, you can create, publish, maintain, monitor, and secure APIs
• API Gateway allows you to create APIs that act as "front door" for your applications
  • can handle up to hundreds of thousands of concurrent API calls
  • prevent exposing endpoints
  • integrates with AWS Lambda
  • integrates with private VPCs
  • create and distribute API keys

 

AWS Step Functions

AWS Step Function

• serverless function orchestrator that you can use to sequence AWS Lambda functions and multiple AWS services into business-critical applications
• low-code, visual workflow service that used to build distributed applications, automate IT business process, machine learning pipelines
• coordinate individual tasks into a visual workflow
• developers can focus on logic
• change and add steps without writing code

Benefits of AWS Step Function

• Build and deploy rapidly
  • get started quickly with Workflow Studio, a drag-and-drop interface
  • express complex business logic as low-code, event-driven workflows
• Write less integration code
  • Step Functions can integrate with many other AWS services
• Build fault-tolerant and stateful workflow
  • managing state, checkpoint, restart
• Designed for reliability and scale
  • standard workflow: manage long-running workloads
  • express workflows: support high-volume event processing workload

Features and Functionality of Step Function

• define and manage the workflow independently from its business logic
• AWS service integrations
• can coordinate any apps that can make an HTTPS connection
• workflow abstraction
• built-in error handling
• history of each job
• visual monitoring
• compliance Health Insurance Portability and Accountability ACT (HIPAA) eligible

Step Functions States

• State: elements in your state machine
  • referred to by its name, must be unique within the scope of the entire state machine
• individual states can make decisions based on their input, perform actions and pass output to other states
• Step Function console providese a graphical representation of state machine

State Machine

• State machine is an object that has a set number of operating conditions that depend on its previous condition to determine output
• you can use Step Functions to create and automate your own state machines
• Step Functions uses JSON-based Amazon States Language
• 여러 단계 또는 상태를 가진 프로세스를 표현, 제어 하는 데 사용