Migrated from FBB to Launtel due to 85Mbps issue
Posted: Tue Nov 03, 2020 1:12 pm
I've been waiting for my internet and VoIP to churn out before posting about my experience.
<tl;dr/>
Long story short, I've churned my household's internet (NBN HFC) from Future Broadband to Launtel due to the 85Mbps issue.
The Problem
I didn't have the time or the inclination to pull apart my network, particularly while my wife and I work from home on weekdays, to troubleshoot something that is, in my eyes, clearly an upstream issue (likely TPG). My connection seemed to fall foul of the ~85Mbps speed throttling/capping issue (with the occasional flex up to 90Mbps). Outside of a speed test, this was also apparent when downloading games via Steam (rarely would I see speeds above 10MBps (that's megabytes per second, so 80Mbps).
There have been theories about the cause of this throttling/shaping of traffic but we're none the wiser as to the actual cause.
Post-Churn Experience
This past Saturday, Launtel took over the internet connection and the only configuration items I had to change in the router were the IPv4 and IPv6 settings (changing from static IP back to DHCP) and removing the VLAN tagging from the WAN interface on the router. Off the bat, I was downloading at 170Mbps and uploading at 23Mbps (on a 250/25 plan) on the upstairs Wi-Fi access point (behind an Ethernet-over-Power connection - might have to give the EoP adapters a power cycle) while the downstairs access point was netting me 273/23.
FBB vs Launtel Raw Data
Anyway, I've attached half-hour results for both FBB and Launtel - all of these results were collated using the Speedtest command-line interface tool on a Windows VM with identical network settings within the VM. I let Speedtest CLI choose the best Speedtest POP rather than forcing a specific to remove any personal bias (although I know Telstra in Sydney almost always yielded the best results). Unsurprisingly, the Speedtest results after churning to Launtel have usually selected the Launtel Sydney POP as the preferred server.
Also, to be fair to FBB, I filtered out the worst 5% of results based on download speeds and treated them as outliers.
Future Broadband Results It's as clear as day something is going on here with averages by Speedtest POP ranging between 85.1 to 85.6Mbps. There are some max values that are in the 88Mbps range (Telstra and AARNet) but everyone else mad max speed results in the 86Mbps band.
Launtel Results Please note, I haven't found a "typical evening" download speed for Launtel's 250/25 service so the target line is missing from the first graph. That said, the dotted yellow line representing the six-hour rolling average (12 data points taken at each half-hour) has, for the most part, remained above 250Mbps. Over-provisioning is well and truly in play as well, with the fastest observed result being just shy of 276.1Mbps (unsurprisingly on the Launtel Sydney Speedtest POP).
Conclusion
While it has been many years since I have worked in internet tech support, I know when the upstream network is at fault versus a fault on my local network. Yes, I'm that customer that has done all the troubleshooting stuff before I call up and I'd really just like you to log a fault with the upstream provider. I really liked what FBB was offering but if I can't get the stated 92Mbps (peak) or 94Mbps (outside peak) speeds then I don't think explaining this phenomenon away under the guise of ensuring low jitter is justification enough for not delivering the speeds advertised on the tin.
In my opinion, a provider (wholesale or otherwise) should be going in to bat for its customers rather than passing on the burden of proof when much in the way of data has already been provided. If there is congestion or technical issues that are negatively impacting service then let your customers know or arm them with the information to see what's going on. Launtel has a leg up here in terms of CVC status and providing extra information about NBN outages in its portal. FBB was good in terms of getting various technical issues/requests sorted (initial dataflow issues due to upstream provider error, IPv6 provisioning, VoIP cutover, etc) but they didn't handle the curly ~85Mbps issue all that well.
I'm also absolutely under no illusion that HFC is a shared network medium but the speeds I am getting on Launtel take that off the table as a potential issue and eliminate my network kit as the cause. I'm also not going to expend time and energy dragging the TIO into the situation (which would kick them both in the pants and in the wallet), even though I reckon I would have a pretty good case - voting with my feet seemed to be the most efficient and effective means of addressing the issue. I also disliked the idea of compensation for and acceptance of a degraded service (which might be $5 - $10 per month), I want the service delivered as paid for and specified on the website.
I think the best we can do is warn others about the situation if they want 100/40 speeds on HFC NBN in an affected state so they go in with eyes wide open.
<tl;dr/>
Long story short, I've churned my household's internet (NBN HFC) from Future Broadband to Launtel due to the 85Mbps issue.
The Problem
I didn't have the time or the inclination to pull apart my network, particularly while my wife and I work from home on weekdays, to troubleshoot something that is, in my eyes, clearly an upstream issue (likely TPG). My connection seemed to fall foul of the ~85Mbps speed throttling/capping issue (with the occasional flex up to 90Mbps). Outside of a speed test, this was also apparent when downloading games via Steam (rarely would I see speeds above 10MBps (that's megabytes per second, so 80Mbps).
There have been theories about the cause of this throttling/shaping of traffic but we're none the wiser as to the actual cause.
Post-Churn Experience
This past Saturday, Launtel took over the internet connection and the only configuration items I had to change in the router were the IPv4 and IPv6 settings (changing from static IP back to DHCP) and removing the VLAN tagging from the WAN interface on the router. Off the bat, I was downloading at 170Mbps and uploading at 23Mbps (on a 250/25 plan) on the upstairs Wi-Fi access point (behind an Ethernet-over-Power connection - might have to give the EoP adapters a power cycle) while the downstairs access point was netting me 273/23.
FBB vs Launtel Raw Data
Anyway, I've attached half-hour results for both FBB and Launtel - all of these results were collated using the Speedtest command-line interface tool on a Windows VM with identical network settings within the VM. I let Speedtest CLI choose the best Speedtest POP rather than forcing a specific to remove any personal bias (although I know Telstra in Sydney almost always yielded the best results). Unsurprisingly, the Speedtest results after churning to Launtel have usually selected the Launtel Sydney POP as the preferred server.
Also, to be fair to FBB, I filtered out the worst 5% of results based on download speeds and treated them as outliers.
Future Broadband Results It's as clear as day something is going on here with averages by Speedtest POP ranging between 85.1 to 85.6Mbps. There are some max values that are in the 88Mbps range (Telstra and AARNet) but everyone else mad max speed results in the 86Mbps band.
Launtel Results Please note, I haven't found a "typical evening" download speed for Launtel's 250/25 service so the target line is missing from the first graph. That said, the dotted yellow line representing the six-hour rolling average (12 data points taken at each half-hour) has, for the most part, remained above 250Mbps. Over-provisioning is well and truly in play as well, with the fastest observed result being just shy of 276.1Mbps (unsurprisingly on the Launtel Sydney Speedtest POP).
Conclusion
While it has been many years since I have worked in internet tech support, I know when the upstream network is at fault versus a fault on my local network. Yes, I'm that customer that has done all the troubleshooting stuff before I call up and I'd really just like you to log a fault with the upstream provider. I really liked what FBB was offering but if I can't get the stated 92Mbps (peak) or 94Mbps (outside peak) speeds then I don't think explaining this phenomenon away under the guise of ensuring low jitter is justification enough for not delivering the speeds advertised on the tin.
In my opinion, a provider (wholesale or otherwise) should be going in to bat for its customers rather than passing on the burden of proof when much in the way of data has already been provided. If there is congestion or technical issues that are negatively impacting service then let your customers know or arm them with the information to see what's going on. Launtel has a leg up here in terms of CVC status and providing extra information about NBN outages in its portal. FBB was good in terms of getting various technical issues/requests sorted (initial dataflow issues due to upstream provider error, IPv6 provisioning, VoIP cutover, etc) but they didn't handle the curly ~85Mbps issue all that well.
I'm also absolutely under no illusion that HFC is a shared network medium but the speeds I am getting on Launtel take that off the table as a potential issue and eliminate my network kit as the cause. I'm also not going to expend time and energy dragging the TIO into the situation (which would kick them both in the pants and in the wallet), even though I reckon I would have a pretty good case - voting with my feet seemed to be the most efficient and effective means of addressing the issue. I also disliked the idea of compensation for and acceptance of a degraded service (which might be $5 - $10 per month), I want the service delivered as paid for and specified on the website.
I think the best we can do is warn others about the situation if they want 100/40 speeds on HFC NBN in an affected state so they go in with eyes wide open.