AWS Certified Solutions Architect COMPLETE STUDY GUIDE - Week 2/12

Understanding AWS Compute Services

Introduction to Compute

• Derek introduces the topic of compute, emphasizing its importance in processing workloads, handling app logic, and evaluating rules. He describes compute as the "brains" and processing power necessary for various applications.

Key Compute Services: EC2 and Lambda

• The main services discussed are EC2 (Elastic Compute Cloud) and Lambda. EC2 is suited for long-term use cases like web servers, while Lambda is designed for short-term tasks or functions that do not require a full server setup.

• Lambda provides on-demand compute power for executing functions quickly without provisioning an entire server, making it ideal for transient workloads.

Overview of EC2

• EC2 is described as AWS's flagship compute service that allows users to launch virtual machines (instances). Instances emulate physical computers using software, enabling multiple VMs to run on a single physical server.

• The economy of scale in AWS allows customers to share costs by provisioning their own instances on shared hardware resources.

Components of EC2

• Key components of EC2 include:

• Amazon Machine Images (AMIs): Templates for launching customized instances quickly.

• Instance Types: Various configurations based on business requirements.

• Instance Purchasing Options: Different pricing models available.

• Tenancy: Determines how instances are hosted.

• User Data & Storage Options: Custom scripts and storage configurations.

Understanding AMIs

• AMIs consist of pre-configured OS and application settings that streamline instance launches. They allow users to automate setups across multiple instances efficiently.

• Users can also purchase AMIs from the AWS Marketplace tailored to specific use cases, simplifying resource management.

Instance Types in EC2

• There are five families of instance types based on workload needs:

• General Purpose: Balanced resources suitable for most applications.

• Compute Optimized: High processing power for demanding tasks.

• Memory Optimized: For large data sets processed in memory.

• Accelerated Computing: Primarily used in machine learning scenarios.

• Storage Optimized: High read/write access needed for large datasets.

Pricing Models in EC2

• The flexibility of pricing models is highlighted:

• On-Demand Pricing: Pay-as-you-go model where users only pay for what they consume hourly without upfront costs.

Understanding EC2 Pricing and Instance Types

Reserved Instances

• Committing to reserved instances for one or three years can lead to significant savings, especially with predictable usage patterns.

• Options include standard or convertible reserved instances, both offering substantial savings compared to on-demand pricing.

• Users can share reserved instances across accounts and sell back standard instances on AWS Marketplace, enhancing cost efficiency.

Spot Instances

• Spot instances provide the highest potential savings (up to 90%) but are not recommended for critical workloads that cannot tolerate interruptions.

• Ideal for flexible workloads like batch processing jobs, spot instances can be reclaimed by AWS if needed for on-demand customers.

Scheduled Instances

• Scheduled instances allow users to pay for resources running on a defined schedule (daily, weekly, monthly), similar to reserved instances but without resale options.

• They require payment regardless of usage over a one-year term and do not have dedicated capacity reservations available.

On-Demand Capacity Reservations

• These reservations allow users to secure specific instance types in designated availability zones without time commitments.

• Flexibility is key; users can create or cancel reservations as needed, although no billing discounts apply.

EC2 Tenancy Options Explained

Types of Tenancy

• EC2 tenancy allows cost savings through shared hardware among multiple AWS accounts (shared tenancy).

• Dedicated instances run on single tenant hardware but may share the physical server with other tenants.

• Dedicated hosts provide complete control over a physical server, suitable for licensing concerns and maximum configuration flexibility.

Utilizing EC2 User Data

Automation at Boot-Up

• EC2 user data enables automation by allowing commands to run when an instance boots up, such as installing software or applying updates.

Storage Solutions in EC2

EBS Volumes and Ephemeral Storage

• Elastic Block Store (EBS) provides persistent storage volumes that can be detached from an instance while retaining data.

• Snapshots of EBS volumes can be encrypted and backed up as part of security best practices.

Ephemeral Storage Characteristics

Understanding EC2 Storage Options

Types of Storage in EC2

• Data stored on ephemeral storage is not recommended for critical data as it can be lost. For persistent storage, use Elastic Block Store (EBS).

• Ephemeral storage is faster than EBS since it resides directly on the instance, making it suitable for temporary data.

Security Considerations in EC2

• EC2 security encompasses three main areas: Security Group configuration, key pairs (public and private keys), and regular operating system patching.

• Security Groups control traffic at the instance level; understanding their rules is crucial for managing network access effectively.

Key Pairs and Instance Access

Authentication Mechanism

• Key pairs are essential for accessing instances via RDP or SSH; they authenticate users to connect securely.

• When creating an instance, a public key is generated and associated with the instance while the user retains the private key for authentication.

Auto Scaling Groups and Load Balancers

Enhancing Scalability

• Auto Scaling Groups automatically adjust the number of instances based on performance requirements, allowing dynamic scaling during traffic spikes.

• Load balancers distribute incoming traffic across multiple instances to optimize latency and improve user experience.

Types of Load Balancers

• There are three types of Elastic Load Balancers:

• Application Load Balancer: Handles HTTP/HTTPS traffic by distributing requests to targets like EC2 instances or containers.

• Network Load Balancer: Designed for high performance, capable of handling millions of requests per second using TCP traffic.

• Gateway Load Balancer: Distributes traffic across third-party virtual appliances.

Benefits of Auto Scaling

Advantages of Automation

• Auto Scaling provides automation that enhances customer satisfaction by adjusting resources based on real-time demand without manual intervention.

• Cost reduction occurs as unused resources are automatically removed when demand decreases, ensuring efficient resource management.

Capacity Management

Auto Scaling and Load Balancers in AWS

Understanding Auto Scaling Components

• Auto scaling components of EC2 include launch configurations or templates, with templates being more user-friendly.

• Key considerations for auto scaling include AMIs, instance types, public IP usage, user data, storage volume configuration, and security group settings.

• Auto scaling groups define the minimum, desired, and maximum number of EC2 instances to maintain; they can specify resource scaling across availability zones.

Types of Auto Scaling Policies

• There are four types of auto scaling policies: manual (rarely used), dynamic (main method), scheduled, and predictive.

• Dynamic scaling uses step scaling (adding/removing instances based on metrics like CPU usage via CloudWatch) or target tracking (automatically adjusting to meet demand).

• Scheduled scaling adjusts resources based on time parameters while predictive scaling anticipates traffic changes to optimize resource allocation.

Role of Load Balancers in Architecture

• Load balancers work with auto scaling to create resilient architectures by distributing traffic among multiple EC2 instances.

• The primary function is managing inbound request flow to targets such as EC2 instances or Lambda functions while preventing overload on any single instance.

Components of Load Balancers

• Key components include listeners (defining how connections are routed), target groups (resource collections for routing), and rules that dictate traffic management actions.

• Each listener can have multiple rules; conditions lead to specific actions for routing traffic effectively.

Health Checks and Load Balancer Types

• Health checks ensure load balancers only send traffic to healthy targets by monitoring responses from specified protocols.

• If a target fails health checks, it is marked unhealthy to prevent poor user experiences due to unresponsive instances.

Load Balancers and Their Types

Overview of Load Balancers

• Load balancers can only serve requests originating from within your VPCs, highlighting their restricted access compared to other services.

• Each availability zone requires a node for the Elastic Load Balancer (ELB), which is essential for cross-zone load balancing.

Application Load Balancer

• The Application Load Balancer supports HTTP and HTTPS traffic at Layer 7, as well as SMTP and NFS protocols. Cross-zone load balancing is always enabled.

Network and Gateway Load Balancers

• Network Load Balancers are designed for TCP/UDP network traffic, capable of handling millions of requests per second with low latency.

• Gateway Load Balancers balance inbound and outbound traffic between virtual appliances running third-party software in your VPC.

Elastic Container Service (ECS)

Introduction to ECS

• ECS supports Docker-enabled applications packaged as containers across a cluster of EC2 instances, simplifying application deployment.

Containers Explained

• A container encapsulates everything an application needs to run except the OS, making it appealing due to its decoupled nature. AWS Fargate manages these containers.

Monitoring and Scaling

• Containers can be monitored using CloudWatch for events like scaling. ECS clusters aggregate resources but can only scale within a single region.

Elastic Container Registry (ECR)

ECR Components

• ECR provides a secure location for storing Docker images, including components like the registry itself, authorization tokens, repositories, repository policies, and images used to build containers.

Elastic Kubernetes Service (EKS)

Overview of EKS

• EKS integrates with Kubernetes to automate deployment and management of containerized applications while abstracting control plane management from users.

Control Plane vs Worker Nodes

• The control plane schedules containers onto nodes and tracks Kubernetes object states; users manage worker nodes that comprise the Kubernetes clusters using specific AMIs.

AWS Elastic Beanstalk

Functionality of Elastic Beanstalk

• AWS Elastic Beanstalk automates resource provisioning needed for web applications using EC2 Auto Scaling and load balancing. While the service itself is free, provisioned resources incur costs.

Core Components

• Key components include application versions (deployable code), environment configurations (parameters/settings), environments (web server workers), configuration templates (baseline settings), and overall application collections.

AWS Batch

Purpose of AWS Batch

• AWS Batch manages batch computing workloads requiring significant compute power across resource clusters. It works effectively with spot instances for cost efficiency.

Job Management in AWS Batch

AWS Outposts and Serverless Computing Overview

Understanding AWS Outposts

• AWS Outposts are designed for hybrid cloud environments, allowing companies to run applications on-premises while utilizing native AWS services.

• There are two deployment options: VMware on AWS for native VM infrastructure or a native AWS variant that uses the same API and management tools as standard AWS.

• Outposts consist of physical hardware delivered to the data center, enabling seamless integration with existing systems.

Introduction to AWS Lambda

• AWS Lambda is the primary serverless computing service from Amazon, allowing users to run workloads without managing servers.

• Functions in Lambda can be written in various programming languages (e.g., Python, C++) and can be invoked manually or automatically based on events.

• Users only pay for compute time when functions are running, making it cost-effective; monitoring is facilitated through CloudWatch.

EC2 Instance Types and Purchase Options

EC2 Instance Types

• Amazon Machine Images (AMIs) serve as templates for launching new instances quickly.

• There are five main instance types: General Purpose, Compute Optimized, Memory Optimized, Accelerated Computing, and Storage Optimized—each tailored for specific workloads.

EC2 Purchase Options

• On-Demand Instances allow flexible usage with pay-as-you-go pricing suitable for irregular workloads.

• Spot Instances offer significant cost savings but are best suited for interruptible workloads; Reserved Instances require upfront payment commitments over one or three years.

EC2 Tenancy and Security Features

EC2 Tenancy Options

• Three tenancy types exist: Shared (default), Dedicated Instances (single tenant hardware), and Dedicated Hosts (physical server dedicated entirely to you).

EC2 Security Measures

• Key pairs encrypt credentials used to access instances; public keys encrypt data while private keys decrypt it. The same key pair can be reused across multiple instances.

Auto Scaling and Load Balancing in EC2

Auto Scaling Features

• Auto Scaling allows automatic adjustment of EC2 instances based on application demand; scaling up increases instance size while scaling out adds more instances.

Elastic Load Balancing

Compute Overview and Transition to Storage

Summary of Compute Section

• The compute section was extensive, comprising approximately 30 slides, indicating a thorough exploration of the topic.

• Emphasis on practical application through labs is encouraged for better understanding and hands-on experience.

• A summary reference is suggested for any concepts that may be unclear or unfamiliar to the audience.

• The speaker expresses anticipation for the next session focused on storage solutions, marking a transition in topics.