A switch uplink is a single cable: too little bandwidth and it clogs, one broken cable and it dies. Link aggregation (LACP, IEEE 802.3ad), also called port aggregation, binds two or more physical ports into one logical link — bandwidth adds up, and if one member fails the rest take over with no outage. It delivers a wider, steadier uplink, not a faster single flow.
LACP adds bandwidth (2×1G ports become a 2G uplink) and failover (a dead member shifts traffic in milliseconds to seconds).
Aggregation is multi-flow, not single-flow speed-up: one TCP session stays on one member port, so it tops out at one port's rate (e.g., 1G).
Prefer dynamic LACP over static: it negotiates with LACPDUs and isolates a misconfigured member, safer than hard-configured static that can loop or blackhole.
1. Bandwidth stacking. Two 1G ports between switches become one 2G logical uplink; four ports become 4G. That is the most direct win for video surveillance and data collection, where the uplink is the natural bottleneck.
2. Failover. Any member in the group breaks and traffic shifts to the remaining members in sub-second. A single-link design means a full segment outage on one cable break.
Link aggregation is "multi-flow stacking," not "single-flow speed-up." One TCP session takes exactly one member port, so it can never exceed one port's rate (1G). To make a single flow faster, raise the port rate — do not just add ports.
| Dimension | Static aggregation | Dynamic mode |
|---|---|---|
| Negotiation | Both ends forced, no packets | LACPDU negotiation |
| Peer not configured | May loop or blackhole | Will not pass if no negotiate — safer |
| Use | Old gear without LACP | Mainstream industrial / enterprise |
| Mode combo | None | Active + active / active + passive |
Static aggregation hard-configures both ends; if the peer forgot to enable it, you get a broadcast storm or one-way reachability. The dynamic mode probes via packets and isolates a member whose state is wrong, which is far more forgiving in the field. Default new projects to dynamic mode.
Once the group is up, the switch uses a hash to decide which member each flow takes. Common hash factors are source MAC, destination MAC, source IP, destination IP, and layer-4 port. In industrial sites, a source-IP + destination-IP + port combination hash spreads traffic most evenly across cameras and subnets.
The hash is per flow, not per byte. If the site has only a few elephant flows (a couple of large camera streams), some members sit busy while others idle. Adding ports will not fix that — look at the number of flows.
The logic is consistent across brands, even if the command words differ:
Create the aggregation group and set the mode to LACP.
Add the physical uplink ports to the group; shutting them first is safer.
Both ends' member ports must share the same speed and duplex, or negotiation fails.
Configure the aggregated port as a Trunk and allow the needed VLANs.
Save, then pull a cable to confirm failover.
Do not walk away after config — pull one member on site and watch whether traffic breaks and how long convergence takes. Many "theoretically millisecond" projects actually went dark because the peer never enabled LACP.
Warning: A mode mismatch (you active, peer static) can leave the link down or one-way. Confirm both ends run LACP and every member port shares the same rate and duplex before you trust the bundle.
Can LACP and static aggregation mix in one group? No. One end static and one end dynamic means negotiation fails — at best no link, at worst a loop. Keep the whole group in a single mode.
After bonding two 1G ports, can one camera stream reach 2G? No. One flow lands on one member port and is capped at 1G. Aggregation raises total multi-flow bandwidth and redundancy, not single-flow speed.
How fast is failover when a member drops? The protocol detects via the packet period (about 1 second by default); the remaining members take over in milliseconds to seconds depending on the peer and its timers. Pull a cable on site before go-live for the real number.
Why didn't my aggregated uplink come up? Usually a mode mismatch (you active, peer static) or a member port speed or duplex mismatch. Confirm both ends run LACP and every member port shares the same rate and duplex.
Why is one member port idle while others are busy? Hashing is per flow, not per byte. A few elephant flows can pile onto specific members. More ports will not fix it — spread the flows or rebalance the hash factors.
Port aggregation is the cost-effective way to widen and harden an uplink: bind physical ports into one logical link for added bandwidth and automatic failover. Remember it stacks multi-flow, not single-flow; pick dynamic over static; and pull a cable to verify failover before you leave.
How to Prevent Loops and a Broadcast Storm on a Managed Switch
VLAN Advanced: QinQ and Voice VLAN for Campus and Industrial Use
About the author: Sara — Sara is a Customer Manager at Rayin with over 10 years of experience in the communications field. In her free time, she enjoys badminton and swimming.
About Rayin: Shenzhen Rayin Technology Co., Ltd. — Company Profile
