Does All‑Optical Bandwidth Aggregation Simply Sum Three Broadband Lines? What Can and Cannot Be Aggregated

On acceptance day, the business owner put forward a clear request: after connecting three broadband lines, run a speed test from one PC and show the aggregated total bandwidth at full speed.
The engineer did not rush to plug in cables. Instead, he prepared two checklists: one for bandwidth, one for concurrent sessions. The phrase “three lines added together” needs clarification. Two resources can be aggregated, while one cannot. A single download task on one PC falls into the latter category.

I. Aggregatable Item 1: Concurrent Aggregate Throughput

The first aggregatable resource is total bandwidth, more precisely the aggregate throughput under multi‑user, multi‑task concurrency.
Imagine multiple water pipes feeding one reservoir. The total inflow rises as all pipes run together. When hundreds of staff browse the internet, dozens of surveillance streams upload simultaneously, office downloads and guest traffic run in parallel — these consist of numerous independent sessions. The gateway distributes sessions across uplinks, so the overall throughput can approach the sum of individual line speeds.
The more users and mixed workloads you have, the greater the benefit of aggregation. This matches the typical traffic pattern for most enterprise egresses.

II. Aggregatable Item 2: Concurrent Session Capacity

The second aggregatable resource is concurrent session capacity, often overlooked.
Each uplink comes with its own NAT session table and public egress resources. N uplinks deliver N sets of concurrent session capacity. A single endpoint may maintain dozens of sessions; hundreds of terminals can easily generate tens of thousands of sessions, which quickly exhausts the session table on a single broadband link.
AINOPOL all‑optical multi‑link aggregation balances sessions evenly across uplinks and reserves headroom for each, preventing premature exhaustion on any one link.
Balancing purely by bandwidth ratio causes failures: if the scheduler only tracks remaining Mbps and ignores session limits, cheaper broadband links hit NAT table ceilings first. New requests routed to them time out. Symptoms include existing web pages loading fine, new pages failing, and occasional POS freezes — while bandwidth tests show spare capacity. This is not a bandwidth shortage; it is session exhaustion.
Aggregation covers two resources: bandwidth (Mbps) and concurrent sessions. Neglecting either means you are not fully leveraging all uplinks.

III. What Cannot Be Aggregated: Instant Speed of a Single Session

What cannot be combined is the instantaneous speed of one TCP session, such as a single‑file download, one video stream, or one online banking login.
One session stays on one uplink for its entire lifecycle. Packets cannot be split across different carrier egresses. If split, the remote server sees different source IPs, rejects the flow, and drops the connection.
Using the highway analogy: uplinks are parallel lanes. Total traffic volume rises with more cars, but one car cannot be split into two pieces to travel on separate lanes.
Multiple links do not accelerate a single task. They provide extra capacity for concurrent users and workloads.

IV. Three Conditions for Valid Aggregated Speed Tests, and Dual‑Resource Orchestration

Can a speed test show near‑sum throughput? Yes, only when all three conditions are met:

1. All uplinks are online and pass health checks;
2. The policy distributes traffic by session;
3. The speed test uses multi‑thread mode and selects a nearby test node.
   When all three conditions hold, multi‑thread tests can approach the aggregated value. If any condition fails, do not demonstrate aggregated speed on site. Single‑thread downloads remain limited to one link’s speed.

“Three broadband lines added together” is only half correct. Aggregatable resources are total throughput and total session capacity. Single‑session speed cannot be aggregated.
The water‑reservoir and highway analogies describe concurrent bulk traffic. The “one car cannot be split” rule defines the boundary for individual connections.
AINOPOL all‑optical multi‑link aggregation combines bandwidth and session capacity with session‑based distribution. Multi‑thread speed tests only reflect aggregated performance when all prerequisites are satisfied. Clarifying what can and cannot be aggregated avoids disputes during acceptance.

FAQ

Q: Can downloads from one PC become three times faster with three broadband lines?
A: No. One session stays on one uplink, so single‑thread downloads are capped at the speed of one line. Aggregation improves total throughput under multi‑user multi‑task concurrency and expands overall session capacity.

Q: What exactly does session aggregation combine?
A: Each uplink has its own NAT session table. N uplinks provide N independent pools of concurrent sessions. Sessions are spread evenly to reserve headroom and avoid mass timeouts when one link’s session table fills up.

Q: Why do multi‑thread speed tests sometimes fail to reach the aggregated total?
A: Near‑sum throughput requires all three conditions: healthy online uplinks, session‑based scheduling, and multi‑thread testing against a nearby node. Skip this demo if any condition is missing.