aggregation of ethernet networks
aggregation of ethernet networks
Then either what I've been taught is wrong or pfSense is calling something LACP when in actuality it's not or they've added a feature to it which you don't find in other router/switch implementations of LACP. That's a whole other ordeal and more real to spread that bandwidth across more than one client. Honestly what I'd do just to get my hands on Symmetrical Gigabit. Meanwhile I'm stuck in the middle of nowhere with 100/10 and this guy wants to turn his connection into 4.5Gbit. Feeling jealous right about now.
I wasn't anticipating this outcome because the hashing directs all traffic to a particular MAC address on the same link. Transmitting data from pfSense onto the LAN uses just one connection. But sending data from pfSense to a LAN client requires both links, or at least it can accommodate both, though this depends on how data streams are randomly allocated. I verified this using iperf3. [ID] Interval Transfer Bitrate Response [5] 0.00-10.00 sec 1.08 GBytes 930 Mbits/sec sender [5] 0.00-10.14 sec 1.08 GBytes 917 Mbits/sec receiver [7] 0.00-10.00 sec 1.08 GBytes 927 Mbits/sec 274 sender [7] 0.00-10.14 sec 1.08 GBytes 914 Mbits/sec receiver [SUM] 0.00-10.00 sec 2.16 GBytes 1.86 Gbits/sec 538 sender [SUM] 0.00-10.14 sec 2.16 GBytes 1.83 Gbits/sec receiver I’m currently testing FTTP myself; it should arrive by year-end, though I expect it to be around 1000/115. Another provider offers symmetrical speeds and I plan to switch later. Meanwhile, a friend in Texas is experiencing issues with 16/1 or similar, so I’m grateful for my 76/18 VDSL connection now—especially with the potential of 5G as well (though mobile services seem inconsistent during busy periods). It’s always entertaining to assist people here with speeds that far exceed anything I’ll have soon, and I remain fortunate to be well covered (or will be soon).
FTTP refers to the same service in different terms. It usually travels straight from the exchange or data centre, or comes through a street cabinet when the location is too distant. Sometimes it’s chosen for cost reasons—already providing sufficient bandwidth at the cabinet without needing new fiber. The asymmetry often comes from using older tech like XGPON to cut expenses. The real speed offered can vary widely, depending on how accurately they report average performance or how many users each PON supports at full capacity. Consistent speeds aren’t always reliable because the total available bandwidth for all customers is shared among them.
I find it interesting to consider whether it makes sense to suggest that even though data travels through several interfaces, the combined throughput might still be limited by one interface to one client. This feels more like a theoretical point. I don’t have personal experience with it. It’s something I haven’t tested.
Noticing that upstream aggregation only works from client to switch makes sense. It seems we've been misinterpreting the requirements all along—aggregation is one-way, and its capacity depends on the client's connection at the destination. Since the client I'm sending to uses a 10Gb port, it won't surpass that speed regardless of whether it's single or dual port. If it were also 1Gb dual, the same limitation would apply.