Understanding RVDC Bandwidth Caps and Throughput

Overview

When an RVDC bandwidth limit is configured (for example, 100 Mbps, 250 Mbps, or 500 Mbps), it represents the maximum rate at which traffic is allowed to traverse the RVDC connection.

It is important to understand that this is not the same as the application throughput you will see within Zerto or during replication.

Why doesn't a 250 Mbps cap equal 250 Mbps of usable throughput?

The configured bandwidth limit applies to all network traffic, not just the replication payload.

Several factors consume part of the available bandwidth:

• TCP/IP protocol headers

• Ethernet overhead

• Packet acknowledgements (ACKs)

• Packet retransmissions

• Rate limiting behaviour

• Encryption overhead (where applicable)

These are all normal parts of network communication.

As a result, the amount of actual replication data transferred will always be lower than the configured bandwidth cap.

Example

A customer upgraded from 100 Mbps to 250 Mbps and observed:

• Configured RVDC cap: 250 Mbps

• Measured application throughput: ~192 Mbps (24 MB/s)

After adjusting the burst policy on the rate limiter:

• Application throughput increased to ~26 MB/s

• Equivalent to approximately 208 Mbps

• Around 84% efficiency of the configured bandwidth cap

This is considered healthy and within expected operating parameters.

Why does this happen?

The RVDC enforces the configured bandwidth limit by rate limiting traffic.

When traffic reaches the configured cap:

• Packets may queue

• TCP may retransmit packets

• Additional acknowledgements are exchanged

• Network overhead increases

These mechanisms ensure reliable delivery but reduce the amount of bandwidth available for actual application data.

MB/s vs Mbps

A common point of confusion is the difference between megabytes per second (MB/s) and megabits per second (Mbps).

The theoretical maximum assumes zero protocol overhead, which is not achievable on a real network.

Can customers expect the full configured bandwidth?

No.

A configured cap of 250 Mbps means:

• The connection is permitted to transmit up to 250 Mbps of network traffic

• It does not guarantee 250 Mbps of usable replication throughput

Depending on network conditions, protocol overhead, and retransmissions, customers should expect usable throughput to be somewhat lower.

Throughput in the region of 80–90% of the configured bandwidth is typically considered normal for sustained TCP replication traffic.

How can throughput be improved?

If replication is consistently falling behind:

• Increase the RVDC bandwidth allocation.

• Reduce competing network traffic where possible.

• Ensure latency and packet loss are within acceptable limits.

• Review replication workload and change rates.

Increasing the bandwidth cap provides more capacity for both application data and the unavoidable protocol overhead.

Key Takeaways

• The RVDC bandwidth cap limits total network traffic, not just replication data.

• Protocol overhead and TCP behaviour reduce the usable application throughput.

• Seeing less than the configured bandwidth is expected and does not necessarily indicate a fault.

• Achieving approximately 80–90% of the configured cap as usable throughput is generally considered healthy.

• If replication cannot keep up with change rates, increasing the bandwidth cap is the recommended solution.