I ran into something the other day in the context of a different experiment, where the aggregation of lines definitely was better than a single line.
Background:
- Dual modems, one with Verizon (regular), one with Visible.com (a Verizon MVNO)
- Single band LTE, band 13, for both lines.
- Encrypted SpeedFusion to a Balance 380 hub (on a 1 Gbps connection)
Observations:
- The download speed was quite awful - slower than the regular single line by a lot. A lot slower than simply running a speedtest with an even balance outbound policy.
- The upload speed was faster than a single line.
- The Visible.com line suffered a 150ms additional latency compared to the Verizon line. (e.g. 50ms vs. 200ms)
Test:
- Swap out the Visible.com line for a second Verizon line
Observation:
- Yowza! The two Verizon lines together outperformed the single line by a lot. (Variably so, 50+% seemed a common reading, sometimes a lot better.)
Conjecture:
Might SpeedFusion be so sensitive to latency differences that the 150ms extra for one line over the other caused the slowdown?
I have dismantled the test set-up, so this is it for now.
Cheers,
Z