AWS Study Group - Chapter 10 - Matching Supply and Demand [Solution Architect Associate Guide]
0 likes209 views
This chapter discusses how to match computing resource supply and demand on AWS. It covers Elastic Load Balancing (ELB) and its three types - classic, application, and network load balancers. It also discusses AWS Auto Scaling, which allows automatically scaling computing resources up or down based on demand. Key attributes of ELB like stateless/stateful, internet-facing/internal-facing, and cross-zone load balancing are explained.
AWS Study Group - Chapter 10 - Matching Supply and Demand [Solution Architect Associate Guide]
- 1. Study Group: AWS SAA Guide Chapter 10 - Matching Supply and Demand Sean 2020.Jun
- 2. AWS Certified Solutions Architect Associate Guide Chapter 10 Matching Supply and Demand
- 3. Topics ● System Reliability ● Elastic Load Balancing and Types ● ELB Attributes ● AWS Auto Scaling
- 4. System Reliability Reliability is the characteristic of a system to acquire compute resources when patterns in traffic change and the ability to acquire resources dynamically to match the demand.
- 5. AWS ELB (Elastic Load Balancing) 1. A Reserver Proxy Service 2. Decoupling Applications 3. Scaling Horizontally (Scale-up) with High availability 4. Health Checks
- 6. Benefits 1. Highly Available (高可用性) 2. Secure (安全) 3. Elastic (彈性) 4. Flexible (靈活) 5. Robust Monitoring & Auditing (監控及稽核功能) 6. Hybrid Load Balancing (混合負載平衡)
- 7. The Elastic Load Balancing service provides with three different load balancers for the general and specific purpose: ● Classic Load Balancer ● Application Load Balancer ● Network Load Balancer ELB Types
- 8. Classic Load Balancer Layer 4 or Layer 7 Load Balancing. CLB (Classic Load Balancer) is intended for applications that were built within the EC2-Classic network.
- 9. Classic Load Balancer ● Support for EC2-Classic ● Support for TCP and SSL listeners ● Support for sticky sessions using application-generated cookies
- 10. Network Load Balancer Only Layer 4 Load Balancing. For extreme performance/low latency applications are recommended using Network Load Balancer
- 11. Network Load Balancer NLB works at layer 4 with the ability to scale up to millions of requests per second and they can are compliant with the WebSockets protocol
- 12. Application Load Balancer If applications only need to load balance HTTP requests, Application Load Balancer is recommended.
- 13. Application Load Balancer Protocol (HTTPS, HTTP) Path ( /orders/*, /img/* ) Port (8080 for frontend, 8081 for backend)
- 14. ● Stateless versus Stateful ● Internet-facing versus internal-facing ● TCP Passthrough ● Cross-zone load balancing ● Connection draining ELB Attributes
- 15. Stateless versus Stateful Robust scalable and fault-tolerant architectures are designed using stateless servers。Services like DynamoDB and ElastiCache are great options to offload temporary session data to an external service
- 16. Stateless versus Stateful Stateful applications, on the other hand, do manage conversational state with the end user and are not fault tolerant。 To enable sticky sessions and send a session cookie to rout all future request to the same server No sticky sessions Sticky sessions
- 17. Internet-facing versus internal-facing Internet-Facing An internet-facing load balancer will listen for requests coming from the internet. Internal-Facing internal-facing load balancer will only route requests from private IP addresses across the VPC. Internet-Facing Internal-Facing
- 18. TCP Passthrough This connection mode will take the request and pass it through as is without adding any additional header information, this is a good fit for SSL certificates that are out of the scope of the ELB, for example, CloudFront incoming requests.
- 19. Cross-Zone Load Balancing When enabled will balance the number of instances evenly across AZs.
- 20. Connection Draining “Graceful Shutdown” A graceful method to remove in service instances that have in-flight requests so end users have a good experience when rolling out new versions of code or scale out activities are performed.
- 21. Auto Scaling Lab ● Create AMI Image ● Create ALB ● Create Launch Configuration ● Create Auto Scaling Group