Home / Tools / SnapMirror Estimator
SnapMirror Bandwidth / Catch-up Estimator
Model a baseline, daily incremental load, and recovery after an outage before a growing lag becomes a production surprise.
Workload and link
Planning result
How this is calculated
Each relationship receives an equal share of the link. Usable throughput is 75% of that share to allow for TCP/IP and protocol overhead, retransmits, latency, and normal contention.
baseline seconds = TiB × 1024⁴ bytes × 8 ÷ (link Mbps ÷ relationships × 1,000,000 × efficiency × 0.75)
daily seconds = baseline seconds × change %
catch-up seconds = outage backlog bits after efficiency ÷ (usable link bits/s − ongoing changed-data bits/s after efficiency). If the ongoing rate is at least the usable link rate, the relationship never catches up.
minimum link Mbps = daily changed bytes × 8 × relationships ÷ (2 hours × efficiency × 0.75 × 1,000,000)
24-hour transfer reference
Approximate logical TiB transferred in 24 hours at 75% link utilization. Efficiency expresses how much logical data is represented by each byte sent.
| Link | 2× efficiency | 3× efficiency | 4× efficiency |
|---|
What changes the real transfer
SnapMirror establishes a baseline from a Snapshot copy, then transfers changed blocks between later snapshots. ONTAP storage efficiency matters because SnapMirror can preserve and transfer deduplicated, compressed blocks instead of expanding every logical byte onto the wire.
The 75% utilization assumption deliberately avoids line-rate promises: TCP/IP framing, protocol work, retransmits, latency, competing traffic, and source or destination load all reduce sustained payload throughput. Where your ONTAP version and relationship support it, snapmirror modify -throttle can cap transfer bandwidth so replication does not consume the link.
snapmirror show -fields transfer-rate, ONTAP statistics, and Active IQ. Measure during representative production load before committing to an RPO or WAN purchase.