Unobvious on-premise costs
- Labor
- broken disks
- patching hosts
- servers going offline
- Network
- Bandwidth needs (peak/average)
- Capacity/Utilization
- Cost of overprovisioning
- How much buffer capacity
- what to do when exceeded
- Availability
- Do you have DR
- Power
- What is peak/average power requirements
- HVAC
- 2N redundancy (everything is duplicated)
- Space
- How much can you grow
Migration Bubble
- Temporarily increased costs
- Planning and assessment
- Duplicate environments
- Staff Training
- Migration Consulting
- 3rd Party Tooling
- Lease penalties
- Due to early termination of contracts
Methodology
- Evaluate
- Migration Readiness Assessment (MRE)
- Are we ready?
- Do we have all resources?
- Migration Readiness Assessment (MRE)
- Plan
- MRE and Planning Engagement
- Design
- Migrate
- Execution
- Discovery
- What do I have
- Grouping / Mapping
- Servers into application stacks
- Always migrate verticals
- Horizontal move (e.g. DB first) unrealistic
- Migration Strategy
- See below (6R Approach )
- Data Migration
- Landing Zone
- Server Migrations
- Database Migrations
- Discovery
- Execution
- Optimize
Migration Strategy (6R Approach)
- Retain (Revisit)
- Do not touch
- e.g. mainframe that must stay
- Do not touch
- Remove
- Decomission
- Rehost
- aka Lift&Shift
- Minimal changes to application
- typically configuration only
- Storage migration needed
- e.g.
- Import VM as-is
- Replatform
- Up-version of OS
- Data Conversion
- e.g. MysQL -> RDS (PostgreSQL)
- Refactor
- Change to use Cloud native mechanisms, e.g.
- multi-region
- Auto-Scaling
- Change to use Cloud native mechanisms, e.g.
- Rearchitecture
- Complete rewrite
- e.g. Server -> Serverless
- Complete rewrite
Difficulties in migration
- Coordination
- Lack of testing
- Server downtime during cutover
References
No comments:
Post a Comment