01
Understand
Decision comparison
| Factor | Edge or on-site may fit | Cloud may fit |
|---|---|---|
| Latency | Fast local response is necessary | Delay is operationally acceptable |
| Connectivity | The function must continue through WAN loss | Buffered or delayed operation is acceptable |
| Scope | One machine, line, or site | Cross-site aggregation and collaboration |
| Data volume | Filter or aggregate near the source | Central retention or analysis of selected data |
| Maintenance | Local skills/vendor support are available | Managed platform reduces selected operational burden |
| Cost | Stable local capacity and long asset life | Elastic or shared services improve comparable economics |
03
Apply
Hybrid failure design
Document what continues locally, what queues, how long data can be buffered, how duplicate replay is prevented, what users see during an outage, and how records reconcile after reconnection.
04
Review
Questions before placement
01What decision does the workload support and how quickly?
Record the decision, responsible owner, supporting evidence, and follow-up action.
02What happens when the site or provider is unreachable?
Record the decision, responsible owner, supporting evidence, and follow-up action.
03Which data may leave the site and under whose approval?
Record the decision, responsible owner, supporting evidence, and follow-up action.
04Who patches, monitors, backs up, and restores each layer?
Record the decision, responsible owner, supporting evidence, and follow-up action.
05Are licensing, egress, support, and hardware replacement included?
Record the decision, responsible owner, supporting evidence, and follow-up action.