Section 23 - Disaster Recovery & Migrations

Disaster Recovery Strategies

• Backup and Restore
• Pilot Light
  • Small version of app is always running in the cloud
  • Faster than backup and restore as critical systems are already up
• Warm Standby
  • Full system is up and running but at a minimum size
  • Can scale to production load when a disaster occurs
• Hot Site / Multi Site Approach
  • Low recovery time objective (seconds or minutes) but very expensive
  • Full production scale running on AWS and on-premises

Disaster Recovery Tips

• Backup
  • Snapshots
  • Regular pushes to S3, S3 IA, Glacier, lifecycle policies, cross region replication
  • Snowball or Storage Gateway if on-premises
• High availability
  • Route53 to migrate DNS over from region to region
  • RDS multi-az, ElastiCache multi-az, EFS, S3
  • Site-to-Site VPN as a recovery from Direct Connect
• Replication
  • RDS Replication (cross region), Aurora + Global Databases
  • Database replication frm on-premises to RDS
  • Storage Gateway
• Automation
  • CloudFormation or Elastic Beanstalk to re-create a whole new environment
  • Recover and reboot EC2 instances with CloudWatch if alarms fail
  • Lambda functions for customized automation
• Chaos
  • Netflix has a "simian-army" randomly terminating EC2

Database Migration Service

• Quickly and securely migrate databases to AWS
• Resilient and self-healing
• Source database still available during migration
• Supports:
  • Homogeneous migrations (ex. Oracle to Oracle)
  • Heterogeneous migrations (ex. MSQL Server to Aurora)
• Continuous data replication using CDC
  • Must create an EC2 instance to perform replication tasks
• Sources
  • On-premises and EC2 instance databases: Oracle, MSQL Server, MySQL, MariaDB, PostgreSQL, MongoDB, SAP, DB2
  • Azure SQL DB
  • RDS
  • S3
  • DocumentDB
• Targets
  • On-premises and EC2 instance databases: Oracle, MSQL Server, MySQL, MariaDB, PostgreSQL, SAP
  • RDS
  • Redshift, DynamoDB, S3
  • OpenSearch
  • Kinesis Data Streams
  • Apache Kafka
  • DocumentDB, Neptune
  • Redis, Babelfish

AWS Schema Conversion Tool

• Convert DB schema from one engine to another
• Prefer compute-intensive instances to optimize data conversions
• Don't need SCT if migrating from the same DB engine (ex. on-premise PostgreSQL to RDS PostgreSQL)

DMS Multi-AZ Deployment

• If multi-az is enabled, DMS provisions and maintains a synchronously stand replica in different AZ
• Benefits:
  • Data redundancy
  • Eliminates I/O freezes
  • Minimizes latency spikes

On-Premises Strategy with AWS

• Ability to download Amazon Linux 2 AMI as a VM (.iso format)
• VM import / export
  • Migrate existing apps to EC2
  • Create a DR repository strategy for on-premises VMs
  • Can export back the VMs from EC2 to on-premises
• AWS Application Discovery Service
  • Gather information about on-premises servers to plan a migration
  • Server utilization and dependency mappings
  • Track with AWS Migration Hub
• AWS Database Migration Service
  • Replicate databases
  • Works with various DB technologies
• AWS Server Migration Service
  • Incremental replication of on-premises live servers to AWS

AWS Backup

• A fully managed service to centrally manage and automate backups across AWS services
• Don't need custom scripts and manual processes
• Supports:
  • EC2, EBS
  • S3
  • RDS (all DB engines), Aurora, DynamoDB
  • DocumentDB, Neptune
  • EFS, FSx (Lustre and Windows File Server)
  • Storage Gateway (Volume Gateway)
• Supports cross-region and cross-account backups
• Supports PITR for supported services
• Backups can be scheduled or on-demand
• Tag-based backup policies
• Create backup policies known as backup plans
  • Backup frequency (every 12 hours, daily, weekly, monthly, cron expression)
  • Backup window
  • Transition to cold storage (never, days, weeks, months, years)
  • Retention period (always, days, weeks, months, years)

AWS Backup Vault Lock

• Enforce Write Once Read Many state for all backups stored in Backup Vault
• An additional layer of defense to protect backups against inadvertent/malicious delete operations or updates that shorten/alter retention periods
• Root users can't delete backups when enabled

Application Discovery Service

• Plan migration projects by gathering information about on-premises data centres
  • Sever utilization data and dependency mappings are important for migrations
• Agentless Discovery (AWS Agentless Discovery Connector)
  • VM inventory, configuration, performance history (CPU, memory, disk usage)
• Agent-based Discovery (AWS Application Discovery Agent)
  • System configuration, system performance, running processes, details of network connections between systems
• The data can be viewed within AWS Migration Hub

Application Migration Service

• A lift-and-shift solution which simplifies migrating apps to AWS
• Converts physical, virtual and cloud-based servers to run natively on AWS
• Supports a wide range of platforms, OS and DBs
• Minimal downtime and reduced costs

VMware Cloud on AWS

• Enables customers to manage on-premises data centres and extend its capacity to AWS but maintain using VMware Cloud software
• Useful for migrating your VMware vSphere-based workloads to AWS, running production workloads across private, public, hybrid environments and having a disaster recovery strategy

Data Transfer

• Context: Transfer 200TB of data in the cloud. The connection speed 100 Mbps
• Over internet / Site-to-Site VPN
  • Immediate setup
  • Convert 200TB to Mb, it will take 185 days (16 million seconds)
• Over direct connect 1 Gbps
  • Long for one-time setup (at least a month)
  • Convert 200TB to Mb, it will take 18.5 days (1.6 million seconds)
• Over Snowball
  • Takes 2 to 3 snowballs in parallel
  • About 1 week for end-to-end transfer
  • Can be combined with DMS
• On-going replication / transfers
  • Site-to-Site VPN or DX with DMS or DataSync