ISYS Cloud Architecture Principles
- Auto-scaling – Servers are created and removed automatically so performance keeps up with demand.
- Load balancing – Redundant load balancers distribute workload between the scaling group of servers.
- Built to fail – failure of one or more components provides little degradation of service. The service recovers or rebuilds failed components.
- Data is replicated to multiple datastores for redundancy.
- Stateless web services – Session data is centrally managed for scaling.
- Application and Database servers are on secure isolated private subnets.
- DNS, CDN, Object Storage and load balancers have full redundancy
