Let’s clarify what VMware means by “Global” here… VMware vSAN ESA “Global Deduplication” refers to:
- Cluster-wide deduplication domain
- No longer limited to disk groups (like OSA)
- Data is deduplicated across all hosts within the same cluster
So yes, it is more “global” than before, but only within the boundaries of a single cluster.
What it does NOT mean (biggest misconceptions)?
- “Global” does NOT mean:
- Cross-cluster deduplication
- Cross-site (stretched cluster sites included)
- Cross-region or multi-cloud
- Federation across multiple vSAN environments
Each cluster is still its own dedupe island.
Why this naming trips people up (myself included)…
- As architects/engineers, we hear “global” and think immediately:
- Universal scope
- Shared efficiency across environments
- Massive storage savings at scale
But in this case, “global” is really “Expanded scope compared to the old architecture”, not “Global across your infrastructure”.
Why VMware designed it this way (and to be fair, there’s a very good reason):
- Performance First
- ESA is optimized for high-performance NVMe architectures
- Deduplication is post-process, not inline
- Keeps dedupe off the write path
- Simplicity & Scalability
- Cluster-level boundary means predictable performance
- Avoids complexity of cross-cluster metadata coordination
- Easier failure domains (no single disk causing an entire disk group failure!)
I’ll be honest, this one got me when I first heard the announcement during VMware Explore too.
When I first heard “Global Deduplication,” I immediately thought broader scope, bigger efficiency and more reach. It wasn’t until I dug deeper that I realized what it actually meant which is exactly why I’m sharing this article to clarify for those that made the same assumptions as myself.
