AWS CSA Notes '22: Part 4 of 8 - Databases

on

4 Databases

4.1 Amazon RDS (Relational Database Service )

... offers 6 familiar database engines: SQL Server, Oracle Database, MySQL, MariaDB, PostgreSQL and Aurora.

Features that AWS RDS offers to support several use-cases:

  • Multi-availability zone (AZ) deployments
  • Read replicas
  • Automatic backups
    • Default retention period is 7 days and goes as high as 35 days.
4.1.1 Amazon Aurora
  • Aurora is a PostgreSQL and MySQL compatible database.
  • Aurora is a fully managed, closed-source, proprietary database from AWS.
  • With Amazon Aurora, you don't need to provision storage in advance to handle future growth. It can auto-scale up to 128 TB per database instance in increments of 10 GB.
  • Aurora continuously scans data disks and blocks. Aurora automatically restarts the database and associated processes once it detects any error with a data disk. It can withstand the loss of up to three copies of data without impacting database read availability and up to two copies without affecting write availability in a transparent manner.
  • Aurora replicates each chunk of database volume in six ways across three AZs. 3/6 copies are required for database reads, and 4/6 copies are needed for database writes.
  • Aurora supports distributed or parallel queries using Parallel Query, which can speed up searches by up to 2 orders of magnitude while keeping transaction workloads moving at a high rate.
  • A cloud user can create Aurora cluster duplicates using Aurora cloning.
  • With Amazon Aurora Backtrack, you can rewind a database cluster to a specified point without restoring data from a back-up. The default backtrack period is 72 hours.
  • Aurora provides support for custom database endpoints for distributing and balancing workloads across different sets of database instances.
  • Aurora supports cross-region replication through logical or physical replication. It uses SSL to encrypt data in transit and KMS for data at rest.
  • The Aurora Global Database enables the creation of a single Aurora database that spans multiple AWS regions for quick disaster recovery and fast local reads.
  • A global database supports one primary region (for read/write) and replication across up to five secondary regions (for read-only), with a latency of less than 1 second. In the event of a disaster or unavailability of the primary database, Aurora switches to a database in the secondary region for read/writes.
4.1.2 Amazon Redshift

"A fully managed data warehouse service in the cloud." Some features you should be familiar with:
  • Massively parallel
  • Machine learning
  • Result caching
  • Fault tolerance
  • Federated query
Note: Redshift does not support Multi-AZ deployments.

4.2 Amazon Non-Relational Database Service (noSQL)

DynamoDB, Elastic-Cache, DocumentDB, Timestream, Neptune, QLDB

4.2.1 Amazon DynamoDB

Tips to remember:
  • Amazon DynamoDB is:
    • fully managed
    • runs on SSD to provide high I/O
    • synchronously replicates data across 3 AZs in an AWS region
  • Two modes:
    • on-demand capacity mode
    • provisioned mode - specify the number of reads and writes an application requires beforehand
  • DynamoDB creates and maintains indexes for primary key attributes.
    • Supports local and global secondary indexes.
  • Two different read modes:
    • strongly consistent
    • eventually consistent (default)
  • DynamoDB global tables provide a multi-active, multi-region, fully managed DB that delivers local, fast read and write performance for global, massively scaled applications.
  • Can set the Time-to-Live (TTL). Items are automatically deleted after the expiry.
4.2.2 Amazon ElastiCache

... an in-memory, fully managed cache and data store.

Differences and similarities between ElastiCache for Redis and ElastiCache for Memcached:
  • Latency
    • [Redis & Memcached] Provides support for sub-millisecond latency
  • Architecture
    • [Redis] Doesn't provide support for multithreaded architecture
    • [Memcached] Offers support for multithreaded architecture
  • Ease of use
    • [Redis & Memcached] Easy to use and doesn't require much code to integrate into an application.
  • Advanced data structures:
    • [Redis] Supports strings, hashes, sets, lists, hyperlog-logs, bit arrays, and sorted sets
    • [Memcached] Has no support for advanced data structures.
  • Snapshots:
    • [Redis] Provides support for point-in-time data snapshots for data recovery and archival.
    • [Memcached] Offers no support for snapshots.
  • Complex data objects:
    • [Redis] Supports complex data objects
    • [Memcached] Only supports key-value storage
  • Storage type:
    • [Redis] Persistent storage
    • [Memcached] Non-persistent storage
  • Multi-AZ:
    • [Redis] Automatic failover with Multi-AZ
    • [Memcached] Multi-AZ not supported
  • Geospatial data:
    • [Redis] Offers support for working with real-time geospatial data.
    • [Memcached] Doesn't have built-in commands or support for geospatial data analysis.
  • Scales:
    • [Redis] Scales using read replicas and highly available clusters.
    • [Memcached] Scales by using multiple nodes.
  • Backup and restore:
    • [Redis] Supports backup and restore.
    • [Memcached] Backup and restore not supported.
  • Pub/Sub messaging:
    • [Redis] Supports Pub/Sub messaging with pattern matching for real-time comment streams, server intercommunication, social media feeds, and high-performance chat applications.
    • [Memcached] No support for Pub/Sub messaging.
Summary of Database Type, Use Cases and Services.
  • Relational DB
    • Use Cases: CRM, ERP, traditional applications, e-commerce
    • AWS Services: RDS, Redshift, Aurora
  • In-memory DB
    • Use Cases: Session management, caching, geospatial applications, gaming, leaderboards
    • AWS Services: ElastiCache for Redis, ElastiCache for Memcached
  • Wide-column DB
    • Use Cases: Route optimization, fleet management, high-scale industrial applications for equipment maintenance
    • AWS Services: Amazon Keyspaces
  • Key-value DB
    • Use Cases: E-commerce systems, high-traffic web applications, gaming applications
    • AWS Services: Amazon Dynamo DB
  • Time series DB
    • Use Cases: DevOps, IoT applications, industrial telemetry
    • AWS Services: Amazon Timestream
  • Document DB
    • Use Cases: Catalogs, content management
    • AWS Services: Amazon DocumentDB
  • Graph DB
    • Use Cases: Recommendation engines, social networking, fraud detection
    • AWS Services: Amazon Neptune
  • Ledger DB
    • Use Cases: Supply chain, systems of record, banking transactions, registrations
    • AWS Services: Amazon QLDB
Glossary:
  • AES-256 = 256-bit Advanced Encryption Standard
  • AZ = Availability Zone
  • KMS = Key Management Service {AWS Key Management Service}
  • MFS = Multi-Factor Authentication

Comments

Post a Comment