- Compute
- Storage
- Database
- Network
- Management
- Auto Scaling (more EC2 instance)
- PaaS for container
- Simple API to start/stop container
- Check container status
- stateless
- trigger by lamda API
- auto scale
- only need to specify memory
- CPU、network、disk I/O would auto scale
- failover
- S3
- Glacier
- Storage Gateway
- scale、failover
- version
-
Bucket & Object
- Access through HTTP
- http://songlin.s3.amazonaws.com/photos/ggininder.jpg
- Object
- data + metadata(modified time、 Content-Type ...)
- Connect local with cloud (backup)
- Volume-based
- Type-based
- Backend would store to S3/Glacier
- Relational database
- Compatible to MySQL、MariaDB、Oracle、MS SQL Server、 PgSQL、Amazon Aurora (5x faster than MySQL)
- Auto patch
- Backup & Snapshot
- Simple API to scale DB
- vertical (hardware scale out or storage scale up)
- horizontal (partition by your self)
- NoSQL
- Auto spread data to mutilple instance
- Auto divide table across mutiple partitions as independent
- Auto replicate data across to other facilities to achieve availability
- For cache
- Supprt Memcached & Redis
- Auto failover
- Scale out (more instance)
- e.g. work with memcached auto discovery
- VPC
- Route 53
- Load balancer
- Virtual private cloud
- Just like in the same server room
- DNS service
- Traffic Flow (Policy)
- depend on delay、GEO、server status to find optimize way
- API to configure
- Like haproxy
- Support HTTP/HTTPS/TCP/SSL
- Auto scaling
- e.g. if specify EC2 instance > 2, auto start instance
- Combine with router 53 for more flexiable
- health check
- CloudWatch
- CloudFormation
- OpsWorks
- Collect log
- Set Alert
- Monitor status & resource usage & performance
- Monitor DB
- Trigger
- Recipient for creating aws service
- Use Chef to setup your environment
- Route 53
- Load balancer
- EC2 instance
- ElastiCache
- S3 storage
- RDS
- CloudWatch
- CodeCommit
- Trusted Advisor
- WAF
- Directory Service
- ElasticSearch
- kinesis
- API Gateway
- IoT
- ...
Made with Slides.com