When planning redundancy, especially in firewall deployments, you typically have two options: Active-Standby or Active-Active. Selecting the right model depends on understanding when each design is appropriate and where it makes the most sense to use it.
Active-Standby: The Comfort Design
One device is working, the other waits.
Simple, predictable, troubleshooting friendly.
Split-brain scenarios can occur where both devices become active, often due to software bugs or heartbeat communication failures. That's why it’s best practice to connect the heartbeat links directly to prevent issues specifically related to heartbeat failures, avoiding reliance on intermediate devices that could interrupt those signals.
Active-Active: The Performance Design
Both devices work at the same time.
More throughput, Better utilization, Unpredictable forwarding path, troubleshooting nightmare.
Asymmetric routing issues can occur, there for carefully designing is needed. Shared control plane dependencies can make maintenance harder, One node can fail silently as well in this mode.
Common Use Cases
Active-Standby is the most common choice for enterprise networks.
Examples are:
- Enterprise Campuses / Corporate HQs (1-10Gbps internet edges, common enterprise apps)
- Environments with Heavy Stateful Inspection (NAT, DPI, SSL Decrpt, IPS Inspected)
- Organizations with Small Network Teams
Customers would love to have Active-Active designs as it seems both the devices they are paying for are utilized, but Active-Active can be justified in certain situations only in enterprise networks.
Active-Active becomes necessary when performance or continuity requirements are strict.
Examples are:
- High-Throughput Data Centers (40G/100G Links, East-West traffic, Spine-leaf Archs)
- Financial Trading Platforms (microseconds matter)
- Service Providers (Massive traffic, Future horizontal scaling)
As you can see, Active-Active is typically deployed in environments where traffic symmetry is built into the design, or any asymmetry is intentionally engineered and tightly controlled. These setups commonly leverage ECMP, require horizontal scaling for future growth, and are supported by teams with the expertise to manage the added complexity.
