Biggest challenges vs on-prem:

1. Organizational change
   This was harder than the tech. Operations teams are used to “the MES server is in the plant.” Moving to cloud required new trust models, new incident response processes, and tighter collaboration between OT, IT, and security.

2. Security model shift
   Security wasn’t worse, but it was different. Zero-trust, identity-based access, certificate rotation, and auditability all required new skills. Once implemented, though, it was actually stronger than what we had on-prem.

Performance and reliability:

• For visibility, analytics, genealogy, and orchestration, cloud performance was better than on-prem.

• For real-time execution, the hybrid edge + cloud pattern worked well.

• Uptime improved overall because we weren’t dependent on plant-level hardware failures or manual patching.

Cost trade-offs:

• Short term: Custom build + cloud services + integration work isn’t cheap.

• Medium to long term: Worth it. We eliminated hardware refresh cycles, reduced vendor lock-in, scaled plants faster, and unlocked analytics/AI use cases that would have been painful on-prem.

Was it worth it?
Yes—but only because we treated it as a platform shift, not an MES re-host. If your goal is just to move an existing MES to the cloud unchanged, the value is limited. If your goal is flexibility, faster rollout, better data access, and future-proofing, a custom cloud-based approach pays off.

Key lesson:
Cloud MES works best when you accept that the edge executes, the cloud orchestrates and learns. Trying to make the cloud behave like a local server is where most projects struggle.