Afrihost New Network Feedback - Part 3

Status
Not open for further replies.
Morning Everyone

I just want to quickly respond to the multiple complaints from Cape Town, PE and other South Region users. We definitely know that latency is not what we'd like, and our team is working really hard to preserve the overall experience.

I just want to be clear that we're not shaping any products that shouldn't be shaped, however shaping aggressively is still not giving us the headroom we need to keep latency at levels we'd like to see. We've already scheduled critical upgrades to ensure that we can deliver a better experience. In the meantime, our team will continue to look at every possible way to deliver the best experience possible when demand is at its highest.

I see a few people mentioning promotions putting strain on the network. We've mentioned many times before that this is not the case. In fact, Free-Time is one example where we actually wanted to try to spread out demand on the network, to give our users a better experience (at our cost). However, it just hasn't really had the impact we had hoped for, but certainly has not negatively impacted the network except that it has increased our costs a lot more :(

I just want to reassure everyone here that we always do very careful analysis of the impact of any promotion prior to and during launch to ensure that it does not affect our overall network performance, and if this is the case, our client's experience always comes first.
 
If your getting 180 to bras your problem is between you and exchange. Might want to look into that

I know, had a line fault open for near two weeks that was closed which I keep asking to be re-opened etc.
I still have 8ms latency to bras.afrihost.com, which is strange.
 
I know, had a line fault open for near two weeks that was closed which I keep asking to be re-opened etc.
I still have 8ms latency to bras.afrihost.com, which is strange.

Yeah, you're issue is definitely not a result of exchange congestion at all. Our CC Team will definitely be able to find and resolve the issue.
 
Afriguys, could you prioritise gaming bandwidth/latency to london servers? Path of Exile and Diablo 3. (East Region)

I'm on a capped account and sometimes the latency spikes like the ECG of a heart attack patient :)
 
Yeah, you're issue is definitely not a result of exchange congestion at all. Our CC Team will definitely be able to find and resolve the issue.

Would you mind bumping me back to 10Mbps?

And:
Code:
Pinging bras.afrihost.com [155.239.255.250] with 32 bytes of data:
Reply from 155.239.255.250: bytes=32 time=7ms TTL=63
Reply from 155.239.255.250: bytes=32 time=9ms TTL=63
Reply from 155.239.255.250: bytes=32 time=8ms TTL=63
Reply from 155.239.255.250: bytes=32 time=8ms TTL=63
Reply from 155.239.255.250: bytes=32 time=8ms TTL=63
Reply from 155.239.255.250: bytes=32 time=9ms TTL=63
Reply from 155.239.255.250: bytes=32 time=9ms TTL=63
Reply from 155.239.255.250: bytes=32 time=8ms TTL=63
Reply from 155.239.255.250: bytes=32 time=8ms TTL=63
Reply from 155.239.255.250: bytes=32 time=7ms TTL=63
Reply from 155.239.255.250: bytes=32 time=8ms TTL=63
Reply from 155.239.255.250: bytes=32 time=8ms TTL=63
Reply from 155.239.255.250: bytes=32 time=9ms TTL=63
Reply from 155.239.255.250: bytes=32 time=8ms TTL=63
Reply from 155.239.255.250: bytes=32 time=8ms TTL=63

Ping statistics for 155.239.255.250:
    Packets: Sent = 15, Received = 15, Lost = 0 (0% loss),
Approximate round trip times in milli-seconds:
    Minimum = 7ms, Maximum = 9ms, Average = 8ms

Versus the same time yesterday:
packet loss.jpg

Afriguys, could you prioritise gaming bandwidth/latency to london servers? Path of Exile and Diablo 3. (East Region)

I'm on a capped account and sometimes the latency spikes like the ECG of a heart attack patient :)
lol
 
Would you mind bumping me back to 10Mbps?

And:
Code:
Pinging bras.afrihost.com [155.239.255.250] with 32 bytes of data:
Reply from 155.239.255.250: bytes=32 time=7ms TTL=63
Reply from 155.239.255.250: bytes=32 time=9ms TTL=63
Reply from 155.239.255.250: bytes=32 time=8ms TTL=63
Reply from 155.239.255.250: bytes=32 time=8ms TTL=63
Reply from 155.239.255.250: bytes=32 time=8ms TTL=63
Reply from 155.239.255.250: bytes=32 time=9ms TTL=63
Reply from 155.239.255.250: bytes=32 time=9ms TTL=63
Reply from 155.239.255.250: bytes=32 time=8ms TTL=63
Reply from 155.239.255.250: bytes=32 time=8ms TTL=63
Reply from 155.239.255.250: bytes=32 time=7ms TTL=63
Reply from 155.239.255.250: bytes=32 time=8ms TTL=63
Reply from 155.239.255.250: bytes=32 time=8ms TTL=63
Reply from 155.239.255.250: bytes=32 time=9ms TTL=63
Reply from 155.239.255.250: bytes=32 time=8ms TTL=63
Reply from 155.239.255.250: bytes=32 time=8ms TTL=63

Ping statistics for 155.239.255.250:
    Packets: Sent = 15, Received = 15, Lost = 0 (0% loss),
Approximate round trip times in milli-seconds:
    Minimum = 7ms, Maximum = 9ms, Average = 8ms

Versus the same time yesterday:
View attachment 413442


lol

I'll ask our CC Team to do this for you ASAP. :)
 
Afriguys, could you prioritise gaming bandwidth/latency to london servers? Path of Exile and Diablo 3. (East Region)

I'm on a capped account and sometimes the latency spikes like the ECG of a heart attack patient :)

We always do our best to prioritise this. But it really depends on various things such as your DNS, gaming protocols and the server you're connecting to. If you give me a little more detail around this (maybe over PM) I'll do my best to assist you. :)
 
Morning Everyone

I just want to quickly respond to the multiple complaints from Cape Town, PE and other South Region users. We definitely know that latency is not what we'd like, and our team is working really hard to preserve the overall experience.

I just want to be clear that we're not shaping any products that shouldn't be shaped, however shaping aggressively is still not giving us the headroom we need to keep latency at levels we'd like to see. We've already scheduled critical upgrades to ensure that we can deliver a better experience. In the meantime, our team will continue to look at every possible way to deliver the best experience possible when demand is at its highest.

I see a few people mentioning promotions putting strain on the network. We've mentioned many times before that this is not the case. In fact, Free-Time is one example where we actually wanted to try to spread out demand on the network, to give our users a better experience (at our cost). However, it just hasn't really had the impact we had hoped for, but certainly has not negatively impacted the network except that it has increased our costs a lot more :(

I just want to reassure everyone here that we always do very careful analysis of the impact of any promotion prior to and during launch to ensure that it does not affect our overall network performance, and if this is the case, our client's experience always comes first.

If you don't mind me asking, could you please be a little bit more technical regarding where the latency issues are coming from?

You have said multiple times that it is not capacity issues during peak times, and in your second paragraph mentioned that aggressive shaping is not giving the headroom required to keep latency levels low (implying congestion during peak traffic?).

Just curious..
 
If you don't mind me asking, could you please be a little bit more technical regarding where the latency issues are coming from?

You have said multiple times that it is not capacity issues during peak times, and in your second paragraph mentioned that aggressive shaping is not giving the headroom required to keep latency levels low (implying congestion during peak traffic?).

Just curious..

We believe we have adequate capacity, but demand is a fluid thing and it it changes all the time. Deciding to upgrade capacity is a super expensive thing and can not be a quick decision. This is why we have to monitor, try out different options and find effective ways of managing protocols.

Having "enough" capacity really means adding up everyone's line speed and making that our IPC, which I hope you understand is completely unfeasible as that would translate into a Petabyte per second IPC. (which would be nice)
 
We believe we have adequate capacity, but demand is a fluid thing and it it changes all the time. Deciding to upgrade capacity is a super expensive thing and can not be a quick decision. This is why we have to monitor, try out different options and find effective ways of managing protocols.

Having "enough" capacity really means adding up everyone's line speed and making that our IPC, which I hope you understand is completely unfeasible as that would translate into a Petabyte per second IPC. (which would be nice)

Thanks for the explanation!

It then appears to be similar to what I had thought. First measure would be to try and maximize efficiency of your existing IPC by means of certain tools such as shaping, and if that does not work well enough an IPC upgrade would be considered.

It would be quite interesting to know how much IPC specific ISPs have in relation to their customer base and how they make it stretch without issues :O
 
We believe we have adequate capacity, but demand is a fluid thing and it it changes all the time. Deciding to upgrade capacity is a super expensive thing and can not be a quick decision. This is why we have to monitor, try out different options and find effective ways of managing protocols.

Having "enough" capacity really means adding up everyone's line speed and making that our IPC, which I hope you understand is completely unfeasible as that would translate into a Petabyte per second IPC. (which would be nice)

If I can chime in here. Stating that you cannot just upgrade capacity I understand, but ultimately this has been going on for a while. Users in the South has been complaining for months. At some point that decision has to be made.
And it's getting progressively worse. I've exported my trace route results for December and Jan up to now
Latency
If you just look at YouTube. We've been having around 80-100ms latency early in Jan. Last night we were sitting at 150+ms. Early in December it was 80ms.
At some point something has to give.
 
We believe we have adequate capacity, but demand is a fluid thing and it it changes all the time. Deciding to upgrade capacity is a super expensive thing and can not be a quick decision. This is why we have to monitor, try out different options and find effective ways of managing protocols.

Having "enough" capacity really means adding up everyone's line speed and making that our IPC, which I hope you understand is completely unfeasible as that would translate into a Petabyte per second IPC. (which would be nice)
Without sounding aggressive, this has been going on for quite an extended (I do mean EXTENDED) period of time. Shouldnt "analysis and projection phase" be done by now?
Surely AH should be either at "getting funds" or "planning" or even better "implementation"stage by now.

Oops! Reread post... Upgrades have been scheduled. My bad.
Strange it took so long to say. Think users would have been more empathetic if it was made known earlier
 
Last edited:
If I can chime in here. Stating that you cannot just upgrade capacity I understand, but ultimately this has been going on for a while. Users in the South has been complaining for months. At some point that decision has to be made.
And it's getting progressively worse. I've exported my trace route results for December and Jan up to now
Latency
If you just look at YouTube. We've been having around 80-100ms latency early in Jan. Last night we were sitting at 150+ms. Early in December it was 80ms.
At some point something has to give.

We never said that we can't upgrade capacity. It can be a lengthy process and, as mentioned, is usually considered as a last resort. Our Network Team has been taking your feedback into account and are actively working on improving the experience for our users in the South. :) If the only fix is to upgrade capacity, then so it shall be.
 
We never said that we can't upgrade capacity. It can be a lengthy process and, as mentioned, is usually considered as a last resort. Our Network Team has been taking your feedback into account and are actively working on improving the experience for our users in the South. :) If the only fix is to upgrade capacity, then so it shall be.

Maybe just a couple of questions then
  • When will this be resolved is probably the question on most people's lips. We've had this problem for a while and it's getting more and more frustrating. Also ironically it's the portion of the bandwidth we're paying for that performs the worst, with FreeTime mostly OK. It's all good saying it's in process, but it has been months.
  • Should we log a formal ticket for this issue? I've been told no by the representatives here, but I also don't want to be bitten that there is no formal ticket on your system. Will more tickets logged with this help get to a solution sooner?
 
We never said that we can't upgrade capacity. It can be a lengthy process and, as mentioned, is usually considered as a last resort. Our Network Team has been taking your feedback into account and are actively working on improving the experience for our users in the South. :) If the only fix is to upgrade capacity, then so it shall be.

OK so I see my question has been answered. It will be sorted end of Jan, as soon as FreeTime stops :cry:
 
Sigh the yearly Latency issues raises its ugly head again.

I can pretty much set my clock to this. Every year same time latencies to the EU go up. Between 8 - 11:20 pm every evening playing LoL is just about impossible. 300+ ms. Deja Vu
 
OK so I see my question has been answered. It will be sorted end of Jan, as soon as FreeTime stops :cry:

I'm hoping you'll see the improvement before then. We honestly do appreciate the feedback that you guys provide us with, and it does factor heavily into our decision making process. So if you guys notice anything strange or have a less than ideal experience, let us know here.
 
Status
Not open for further replies.
Top
Sign up to the MyBroadband newsletter
X