Slow downs on an evening all of a sudden

Forums › Spark New Zealand (including Skinny and BigPipe) › Slow downs on an evening all of a sudden

tnaseem

82 posts

Master Geek

#173427 22-May-2015 19:33

I already have a support request running with Bigpipe (who in turn have contacted Chorus), but it's been a couple of days now so I thought I'd throw this out here and see if anyone can shed some light on this...

I've been on Bigpipe since their closed beta period (VDSL), and have always had rock solid connection (apart from the odd hiccup). At this new property since December last year, we've been getting 50Mb/s down 9.7Mb/s up (4ms ping), as we're right across the street for the cabinet. That is until the last few days where the speeds have been dropping to single figures on an evening, roughly from 7pm to around 11pm. Gets to a point where even streaming on Spotify is a problem.

It does sound like congestion, but I can't understand why (a) It's started happening all of a sudden in the last 2 to 3 days, (b) why such a huge drop in speeds?

Before anyone asks, I've checked internally, and there's nothing amiss there. The problem is external.

To illustrate the drop in speeds, see below. These are readings from the last couple of days:

1 | 2 | 3

13459 posts

Uber Geek

ID Verified

Lifetime subscriber

#1310085 22-May-2015 19:41

where are you testing to?

tnaseem

82 posts

Master Geek

#1310108 22-May-2015 19:50

Spark / Auckland

Some extra tests I did last night at 9.30pm:

Speed test: https://www.speedtest.net/my-result/4375418874

Ping has started…

PING google.com (216.58.220.142): 56 data bytes
64 bytes from 216.58.220.142: icmp_seq=0 ttl=54 time=141.586 ms
64 bytes from 216.58.220.142: icmp_seq=1 ttl=54 time=118.844 ms
64 bytes from 216.58.220.142: icmp_seq=2 ttl=54 time=136.128 ms
64 bytes from 216.58.220.142: icmp_seq=3 ttl=54 time=159.864 ms
64 bytes from 216.58.220.142: icmp_seq=4 ttl=54 time=151.067 ms
64 bytes from 216.58.220.142: icmp_seq=5 ttl=54 time=171.072 ms
64 bytes from 216.58.220.142: icmp_seq=6 ttl=54 time=190.964 ms
64 bytes from 216.58.220.142: icmp_seq=7 ttl=54 time=196.307 ms
64 bytes from 216.58.220.142: icmp_seq=8 ttl=54 time=168.809 ms
64 bytes from 216.58.220.142: icmp_seq=9 ttl=54 time=211.612 ms

--- google.com ping statistics ---
10 packets transmitted, 10 packets received, 0.0% packet loss
round-trip min/avg/max/stddev = 118.844/164.625/211.612/27.586 ms

Traceroute has started…

traceroute to google.com (216.58.220.142), 64 hops max, 72 byte packets
1 192.168.1.1 (192.168.1.1) 0.515 ms 0.383 ms 0.311 ms
2 210.54.38.129 (210.54.38.129) 92.815 ms 80.809 ms 81.352 ms
3 122.56.60.217 (122.56.60.217) 140.450 ms 166.224 ms 114.254 ms
4 122.56.60.216 (122.56.60.216) 107.865 ms 173.941 ms 155.274 ms
5 * * *
6 g7-2-8-201.akbr6.global-gateway.net.nz (122.56.118.65) 131.072 ms 109.854 ms 128.256 ms
7 ae5-2.akbr7.global-gateway.net.nz (210.55.202.213) 144.168 ms 126.475 ms 90.120 ms
8 xe0-0-4.sgbr3.global-gateway.net.nz (122.56.127.94) 119.038 ms 106.402 ms 152.938 ms
9 ae2-10.sgbr4.global-gateway.net.nz (202.50.232.246) 110.046 ms 96.592 ms 108.090 ms
10 google-gsw.sgbr4.global-gateway.net.nz (202.50.237.198) 158.464 ms 96.752 ms 91.715 ms
11 216.239.51.218 (216.239.51.218) 96.466 ms 171.734 ms 129.243 ms
12 216.239.46.151 (216.239.46.151) 118.182 ms 127.305 ms 68.710 ms
13 syd09s01-in-f14.1e100.net (216.58.220.142) 70.900 ms 118.058 ms 118.587 ms

sbiddle

30853 posts

Uber Geek

Retired Mod

Trusted

Biddle Corp

Lifetime subscriber

#1310113 22-May-2015 20:07

Have you checked to see you're not saturating your upstream at all?

tnaseem

82 posts

Master Geek

#1310114 22-May-2015 20:13

Yes. Of course. :)

Like I mentioned above, it's not internal. It's the first thing I check when experiencing issues like this. My tests above have been done with a direct wired gigabit connection to the router to remove any variables internally.

One thing Bigpipe did mention in their initial email was that they were nowhere near capacity, so looks like it's local to the exchange. Chorus has been given a buttload of pathping tests that I was collating during the period yesterday, but I've yet to hear back. Being a weekend, I assume this won't be until Monday now...

hashbrown

463 posts

Ultimate Geek

#1310116 22-May-2015 20:18

I've been getting slowdowns recently too. DSL sync is 16.5 down but speedtest to Spark Auckland is 5.3Mbit/s this evening.

tnaseem

82 posts

Master Geek

#1310118 22-May-2015 20:28

Where in Auckland are you? Just to see if you're on the same exchange :) I'm in Bucklands Beach.

Might be worth contacting Bigpipe too, as they explained that I was the only one experiencing issues.

VinLew

210 posts

Master Geek

#1310127 22-May-2015 20:41

Like this?

Equinox IT

Cloud spending continues to surge globally, but most organisations haven’t made the changes necessary to maximise the value and cost-efficiency benefits of their cloud investments. Download the whitepaper From Overspend to Advantage now.

wanghou168

453 posts

Ultimate Geek

#1310129 22-May-2015 20:52

I am with Spark Auckland Epsom area.

Experienced up and downs tonight since 5pm

https://www.speedtest.net/my-result/4377821389

4.66 mbps download and 0.89 mbps upload

MOBILE: Huawei P30
OTHER: Thinkpad E14 + Asus Vivobook + Intel NUC5i3 + A family of Raspberry Pi x 9 + Amazon Echo

hashbrown

463 posts

Ultimate Geek

#1310139 22-May-2015 21:11

tnaseem: Where in Auckland are you? Just to see if you're on the same exchange :) I'm in Bucklands Beach.

Might be worth contacting Bigpipe too, as they explained that I was the only one experiencing issues.

If it were an exchange issue other ISPs would be impacted.

I might do a few more tests just to isolate everything on my network, although nothing has changed on it recently.

tnaseem

82 posts

Master Geek

#1310218 22-May-2015 23:58

Sorry, I meant local exchange (as in cabinet). When I had a speed problem last year, it was at the cabinet. a slightly frayed cable, which Chorus replaced and moved me to a new port just to make sure.

BigPipeNZ

1170 posts

Uber Geek

Trusted

BigPipe

#1310314 23-May-2015 09:34

hey guys,

There is definitely nothing on our part of the network that looks wonky as far as I can tell (although I've asked our network team to have a closer look just in case). We have plenty of capacity on things like international links etc, and nothing appears to be capped. We've had very support requests come through like this.
Checked a couple of other staff members (in auckland) connections at peaks times, and they all look fine (including mine) so it's definitely not anything that is effecting everyone.

As an ISP it's very hard to see these issues come through since the symptoms are virtually identical to the very common issue that is either local wifi congestion or saturating the connection with other traffic (flatmates torrenting, software updates, cloud backup etc).

Those issues seem less likely in this case, given you guys know what you're talking about, so I'll keep digging on this and see what I can find out.
It may require escalation to Chorus and/or Telecom Wholesale to get investigation kicked off from their end.

With the speed only dropping at peak times, that suggests congestion somewhere (rather than a fault - although it doesn't rule it out) so it's really a matter of figuring out where.

If any of you haven't yet opened a ticket with us, please do so (and put a link to this thread in the form). Issues become a lot easier to diagnose when we have a bigger list of customers having the issue so we can look for commonalities e.g. all auckland based, all off one exchange, all VDSL etc etc
and make sure you let us know what you have already checked, so we don't waste time ruling out the common things.

Remember, you don't have to wait on hold for hours to let us know about the issue, just a couple of minutes filling out a form and keep an eye on your emails. :)

bigpipe.co.nz
https://www.facebook.com/BigPipeNZ
https://twitter.com/BigPipeNZ

tnaseem

82 posts

Master Geek

#1310362 23-May-2015 11:46

It does sound like congestion to me too, but in my case it's only started happening in the last several days, and it's a sudden drop from 50Mb/s to around 6Mb/s over the space of 15 minutes (varies around 6 to 7pm start), then a 'sudden' upsurge back to normal at around 10.45pm. Normally, I'm still rock steady at this time here.

My speed test graph below (December 2014 to now - more readings over the last few days which is why they take up most of the graph!):

The 4 dips in that graph are the days I've been getting this issue. Times between 7pm(ish) and 11pm(ish). The first dip, I only took one reading, hence the downward spike. Last reading was just now. Prior to that (19th May), it's been steady as a rock since last December.

Anyway, I do have a support request in (ref:_00D90oVYy._50090pDYE), and was expecting to hear back yesterday, but I guess Chorus contractors went home early on Friday :)

hashbrown

463 posts

Ultimate Geek

#1310401 23-May-2015 14:27

I've started monitoring today. Interestingly today at around 11:00AM my latency to Spark's Auckland OOKLA server jumped from a consistent 11ms up to >40ms suggesting someone is making changes. Hopefully these are to address this issue. Downstream is consistent at 14.1 - 14.2Mb/s which is what I'd expect for my sync rate.

Lurch

1061 posts

Uber Geek

#1310458 23-May-2015 16:11

Vodafone is also doing the same thing around those same what I would call peak hour times. Started happening around the same time you have mentioned. I put it down to Vodafone not having enough bandwidth available at those times, which wasn't the case a few weeks back.
Gave up logging a call, but it's interesting to see Bigpipe has the same woes.

hashbrown

463 posts

Ultimate Geek

#1310477 23-May-2015 17:30

Lurch: Vodafone is also doing the same thing around those same what I would call peak hour times. Started happening around the same time you have mentioned. I put it down to Vodafone not having enough bandwidth available at those times, which wasn't the case a few weeks back.
Gave up logging a call, but it's interesting to see Bigpipe has the same woes.

Interesting. That would tend to confirm that this is congested backhaul and possibly a Chorus issue. Speeds already slowing down for the evening. Down to 8.2Mb/s compared to the reliable 14.1Mb/s I was getting this morning.

The increased latency is interesting too. 50ms just to get to what is presumably a Bigpipe BNG. As per my earlier post, this extra latency only kicked in at about 11:00AM today.

Tracing route to 203.96.123.1 over a maximum of 30 hops
1    <1 ms    <1 ms     2 ms 192.168.1.1
2     1 ms    <1 ms    <1 ms 192.168.175.254
3    50 ms    48 ms    49 ms 100.70.28.1
4     *        *        *     Request timed out.
5     *        *        *     Request timed out.
6    49 ms    51 ms    50 ms 203.96.123.1

Edit:formatting

1 | 2 | 3