{{ summaryTitle }} {{ summaryValue }} {{ summaryLine }} {{ badge.label }}{{ badge.value }}

Use the configured member count before the failure case is applied.
links
LACP members in one aggregator normally use the same speed and duplex settings.
Gbps
Model the present outage or the failure count the design must survive.
links
Compare a measured or forecast peak, not the sum of both traffic directions.
Gbps
Aggregate spare capacity does not let one normal hashed flow exceed one member.
Gbps
Lower the estimate for few elephant flows or limited hash entropy; validate with interface counters.
%
Keep this assumption explicit and avoid double-counting overhead already present in traffic measurements.
%
Automatic is neutral; this setting affects presentation only.
MeasureValuePlanning meaningCopy
{{ row.measure }}{{ row.value }}{{ row.meaning }}
CheckpointFailedActiveUsable capacityHeadroomCopy
{{ row.checkpoint }}{{ row.failed }}{{ row.active }}{{ row.capacity }}{{ row.headroom }}
CheckSignalEvidenceNext actionCopy
{{ row.check }}{{ row.signal }}{{ row.evidence }}{{ row.action }}
ECMP Path Capacity Calculator Estimate ECMP route capacity from healthy paths and hash efficiency, then test reserve headroom plus member-failure and single-flow limits.
Ethernet Frame Efficiency Calculator Calculate Ethernet payload efficiency and goodput from frame size plus VLAN and wire-time overhead, then check whether a target rate fits.
Network Throughput Calculator Estimate useful network throughput and transfer time from link capacity and RTT with TCP window, loss and packet-rate bottleneck checks.
PoE Budget Calculator Plan a PoE switch from endpoint demand and cable loss, then check reserve capacity against powered ports, startup surge and shedding priority.

A four-link 10 Gbps aggregation is often called a 40 Gbps link, yet that label describes gross member line rate rather than guaranteed usable throughput. A failed member immediately removes 10 Gbps. Operational reserve removes another share, and imperfect traffic distribution can leave one link busy while another still has room.

Link Aggregation Control Protocol (LACP) coordinates compatible Ethernet member links so a Link Aggregation Group (LAG) can be treated as one logical link. Aggregation improves total bandwidth across many conversations and adds resilience when a member fails. Traffic is still assigned to physical members according to a distribution policy, commonly a hash built from address and protocol fields.

Aggregate capacity
The combined throughput available across active members after reserve and distribution assumptions.
Single-flow ceiling
The throughput one ordinary hashed connection can receive from one member. Adding members does not multiply this ceiling.
Hash efficiency
A planning discount for uneven distribution across members. It is an assumption to validate with per-interface counters, not an LACP setting.

Workloads with many similarly sized flows usually spread better than workloads dominated by a few “elephant” flows. Hash fields and entropy also matter: several large flows with identical endpoint fields can select the same member. The bundle can therefore have positive aggregate headroom while one member drops packets or reaches line rate.

Capacity should be checked at the failure count the design must survive, not only in the healthy state. The direction and measurement point must remain consistent too. Full-duplex links can carry traffic in both directions, but adding both directions into one demand value can double-count a measurement that already represents the constrained direction.

How to Use This Tool:

Model the configured bundle and the exact failure condition against a measured or forecast peak in one traffic direction.

  1. Enter Member links and the common Member line rate. Members in one active LACP aggregator normally need compatible speed and duplex settings.
  2. Set Failed members to the present outage or the failure count the design must tolerate. The value may equal the member count, but that leaves no forwarding links.
  3. Enter Planned aggregate demand and the Largest single flow. Use the same Gbps basis and avoid summing both directions unless that matches the measured bottleneck.
  4. Set Hash distribution efficiency from observed member balance and choose an Operational reserve that is not already included in the traffic measurement.
  5. Read Capacity snapshot for current headroom, then check the next-failure row and single-flow result. Validate the assumption with member counters under representative traffic.

Interpreting Results:

Usable capacity is the active aggregate after reserve and hash-efficiency discounts. Demand headroom is positive when planned aggregate demand fits and negative when the model shows a shortfall.

  • A ready aggregate can still fail the Single-flow ceiling. One ordinary hashed connection cannot borrow capacity from several members.
  • Next failure headroom answers whether one more member loss still carries the same demand. Zero or negative margin warrants a failover or load-shedding plan.
  • Additional failures safe counts further member losses that retain aggregate demand under the same reserve and distribution assumptions.
  • Recheck per-member utilization, errors, drops, LACP state, and hash policy. A favorable estimate cannot detect a misbundled port or traffic skew.

Technical Details:

The model applies the selected failure count first. Every surviving member contributes its line rate, reduced by one reserve factor and one hash-distribution factor. The single-flow check omits the hash factor because it tests one member's reserved line rate rather than average aggregate spread.

Formula Core:

Aggregate and single-flow capacity are calculated separately so spare bundle bandwidth cannot conceal a flow that is larger than one member.

A =N−F Cusable =AS(1−r100)h100 Cflow =S(1−r100) H =Cusable−D
SymbolMeaningUnit
N, F, ATotal, failed, and active member linkscount
SLine rate of one memberGbps
rOperational reservepercent
hHash distribution efficiencypercent
D, HPlanned aggregate demand and resulting headroomGbps

The next-failure capacity uses max(0, active members − 1) with the same usable capacity per member. Safe failed members are found by checking every failure count from zero through the total member count and retaining the largest count whose usable capacity is greater than or equal to demand.

Status Rules:

StatusExact boundaryInterpretation
ReadyDemand utilization < 70%Modeled aggregate headroom exceeds 30%.
WatchUtilization ≥ 70% and < 90%Track skew and failure margin closely.
Near limitUtilization ≥ 90% with non-negative headroomDemand still fits, but little modeled margin remains.
ShortfallDemand headroom < 0 GbpsPlanned aggregate demand exceeds usable capacity.
Single flow near limitLargest flow ≥ 90% of its ceilingOne member is close to its reserved line rate.
Single flow too largeSingle-flow headroom < 0 GbpsThe flow exceeds one member's modeled ceiling.

Limitations:

Hash efficiency and operational reserve are user-supplied planning discounts, not guarantees from LACP. The model assumes equal-speed members and a stable active aggregator. It does not simulate a specific vendor hash, flow-size distribution, asymmetric routing, receive-side behavior, link negotiation, minimum-links policy, convergence time, oversubscription elsewhere, packet-size overhead, or microbursts. Confirm forwarding members and distribution with device counters and a failure test.

Worked Examples:

One failed member in a four-link bundle

Four 10 Gbps members with one failure leave 30 Gbps of active line rate. A 3% reserve and 78% hash efficiency reduce usable aggregate capacity to 22.698 Gbps. Demand of 18 Gbps leaves 4.698 Gbps headroom and uses about 79.3%, which falls in the watch band. A 6 Gbps largest flow fits below the 9.7 Gbps single-flow ceiling, but a second member failure would reduce aggregate capacity to 15.132 Gbps and create a 2.868 Gbps shortfall.

References: