What actually failed about the classic ESB, specifically
Centralising all integration logic in one platform created a genuine single point of failure and a team that became an organisational bottleneck for every change — a real, well-documented failure mode, not a fashion judgement, and it's a structural risk independent of which vendor's product embodies the pattern.
The messaging and mediation patterns themselves — many catalogued decades ago in the canonical integration patterns literature — remain sound and widely reused; it was the centralisation and governance model around them that caused the pain.
What to check in any new platform embodying these patterns
Whether the new platform recreates centralised bottleneck risk and a single owning team for all integration logic, or whether it genuinely distributes ownership and failure domains, determines if you're adopting a better architecture or the same old ESB risk with a newer logo.
The ESB pattern didn't die. It just stopped being called that. Check the new product against the old failure mode, not the old name.MNB Integration Practice
Key actions
- Evaluate new integration platforms against the classic centralisation failure mode
- Check whether failure domains are genuinely distributed, not just rebranded
- Reuse proven messaging and mediation patterns regardless of product category
- Avoid assuming a new product name implies a new risk profile