Im also having issues with this (cant post stats since Im at work)

Had to use 4G last night to raid since it was so bad (had someone else also on Spark VDSL having to do the same thing)

MrRex:

 

Im also having issues with this (cant post stats since Im at work)

 

Had to use 4G last night to raid since it was so bad (had someone else also on Spark VDSL having to do the same thing)

 

Been happening to me for over a week now, pretty over it.
Weird that it seems Vodafone isn't having these issues :s

here is from Vodafone VDSL

Out of interest, what's the packet loss like?

Cheers - N

0% Packet loss on www.telstra.com.au from Vodafone VDSL

It's like pulling teeth... What is the packet loss like on the affected Spark connections?

Cheers - N

Talkiet:

 

It's like pulling teeth... What is the packet loss like on the affected Spark connections?

 

Cheers - N

 

Not getting any packet loss (if there's a better way to test this let me know)

Edit: tested with 20+ pings and still had 0% loss

Thanks - that's really interesting... My (Spark) connection seems unaffected so I couldn't check myself - but I'll pass along the observation that it's only latency, and no associated loss.

While increased latency isn't good for games, it's not as destructive to throughput as loss. It seems like there's some sort of mitigation going on somewhere - there's no shapers (in a non-negative sense!) in our network that would be contributing to increased latency without loss.

Cheers - N

Talkiet:

 

Thanks - that's really interesting... My (Spark) connection seems unaffected so I couldn't check myself - but I'll pass along the observation that it's only latency, and no associated loss.

 

While increased latency isn't good for games, it's not as destructive to throughput as loss. It seems like there's some sort of mitigation going on somewhere - there's no shapers (in a non-negative sense!) in our network that would be contributing to increased latency without loss.

 

 

 

Cheers - N

 

 

 

Yeah I mean it seems to be an issue on Telstra's end (people have pointed out that there's a cable fault atm?)
But is there no temporary work around spark can do for those affected by this?
Also, is your personal connection ADSL/VDSL or Fibre?

Without understanding exactly what's wrong in Telstra's network, no, there's nothing we can do to provide a temporary workaround that's not just randomly changing things, and that's not something we'd do lightly. Given there's (apparently) no packet loss, the impact should be minimal with regard to most use - although I acknowledge that for some gaming servers it won't be ideal.

My personal line is fibre, but there's no difference in how the fibre and ADSL/VDSL services are presented upstream - except that there are different IP subnets dedicated to each access type. Before you jump on that as an option - there are dozens of separate ranges EACH for ADSL/VDSL and Fibre, it's not like they are contiguous.

There's a chance that the differential impact is related to the fact we load balance across multiple links - users on one side might be hitting the bad part of Telstra and users on the other (that's HUGELY simplified) might be having no issue. It will not be related to a particular access type. And no, we can't flick a user from one to the other because the balancing isn't done per user, it's done at a larger level of aggregation - so we'd have to switch across a large number of users which may imbalance other parts of the network.

While it's still appearing to be improving (I know we have done some things at our end to mitigate the issues as best we can from our side) I don't think we'd be prepared to do significant grooming (moving users from one part of the network to another) just to fix some increased latency for some users, to a small part of the Internet.

It should be better, it will be better, but for a service with no SLA related to latency to particular offnet destinations, we're going to have to give them some more time on this.

Regards

N

yitz: Talkiet can you clarify if a customer signs up for Static IP the assigned static IP will remain within the range allocated for the particular access type for that BNG? So for example if your dynamic VDSL range on a particular BNG is 125.238.0.0/21-vdsl.bb.spark.co.nz would your assigned static IP be within that range or do you still have separate ranges from which static IP's are allocated from?

Good Q. Separate IP ranges for statics. If someone is on a static they can temporarily go to a dynamic by using "NoStatic" as their PPP username.

Cheers - N

Talkiet:

 

yitz: Talkiet can you clarify if a customer signs up for Static IP the assigned static IP will remain within the range allocated for the particular access type for that BNG? So for example if your dynamic VDSL range on a particular BNG is 125.238.0.0/21-vdsl.bb.spark.co.nz would your assigned static IP be within that range or do you still have separate ranges from which static IP's are allocated from?

 

Good Q. Separate IP ranges for statics. If someone is on a static they can temporarily go to a dynamic by using "NoStatic" as their PPP username.

 

Cheers - N

 

 

 

Do you think assigning a static IP in an unaffected range would temporarily fix the issue for those experiencing this?
If so how could we find an unaffected range to request a static IP for?

We don't authoritatively know what the unaffected ranges are. It's also possible this isn't the only predictive factor - it could be isolated to some BNGs and not others. I know it's not what you want to hear and and I WILL see if I can get an update tomorrow, but attempting to 'fix' this by putting people on temporary IPs would create more work to undo as well - plus might not fix the issue.

If you haven't already, it would be worth trying a VPN - that will effectively change your IP and might provide a temporary workaround.

Cheers - N