I am getting packet loss whilst playing games and I dont know why

[JohnK19]

As the title says, I am getting huge packet loss and I dont know why. (I just made a post on this and it was deleted and marked as spam....)

I've tried rebooting everything, Refreshing my Fibre IP. Restarting my PC. Nothing helps and it has been persistent for hours.

Can anyone help?

[jdvega]

 To me, this stinks of IX congestion.  

I emailed my traces to their IP NOC, dealing with the telephone support is a waste of time for these kind of issues:

role: Vodafone IP GSOC
address: Vodafone Group PLC
address: Melbourne Street
address: Leeds
address: LS2 7PS
address: United Kingdom
phone: +44 1344 602224
remarks: ------------------------------------------------------------
remarks: For network issues contact Network Control phone
remarks: +44 1344 602224, email ncipsupport@vodafone.com
remarks: ------------------------------------------------------------

[Ripshod]

Were you asked for the traces and given the email to send them to? If not the emails will just be "spam" to them.

Are you in the South West of England? I ask because your gateway is in Swindon.

Vodafone can't do anything about the internet spine.  It's even out of Cityfibre and Openreach's hands, and we all know how great they are. 

[jdvega]

It's IX related - via LonAP.  Nowt to do with Cityfibre.

Maybe AS5378 directly do not control peering, but AS1273 Vodafone Group (the upstream for AS5378) absolutely DO control the IX peering, transit and PNIs.

[jdvega]

However, the below would suggest that AS5378 DOES have a direct peer to LonAP, not via Voda group/upstream.

https://bgp.he.net/AS5378#_ix

[jdvega]

Using the Looking Glass for AS1273 (the upstream for AS5378) i.e Vodafone Group:

https://portal.vodafone.com/web/lookingglass

Checking the 'IP route' for an impacted Blizzard server:

185.60.112.157

The valid routes are via Arelion/Telia and Level3 (not LonAP) - so the control of traffic to peer via LonAP, is on the AS5378 side:

inet.0: 962071 destinations, 16268213 routes (961972 active, 2 holddown, 197 hidden)
+ = Active Route, - = Last Active, * = Both

185.60.112.0/23 *[BGP/170] 3d 16:49:41, MED 0, localpref 80, from 195.2.1.26
AS path: 57976 I, validation-state: unverified
[BGP/170] 1d 11:35:47, MED 0, localpref 80, from 195.2.1.14
AS path: 57976 I, validation-state: unverified
[BGP/170] 11w5d 14:48:43, MED 0, localpref 80, from 195.2.1.28
AS path: 57976 I, validation-state: unverified
[BGP/170] 20w6d 23:38:57, MED 0, localpref 80
AS path: 1299 57976 I, validation-state: valid
> to 62.115.61.190 via ae28.0
[BGP/170] 3w5d 22:02:07, MED 0, localpref 80
AS path: 3356 57976 I, validation-state: valid
> to 4.68.72.93 via ae13.0

[jdvega]

At 12:55 GMT+0 it looks like Vodafone AS5378 briefly shutdown their LonAP peering - this forced traffic via Level3 instead, for those 5 minutes everything was back to normal.

... then the AS5378 peer via LonAP came back up again, and the problems returned.

[jdvega]

I see on the Blizzard forums, various people suggested this coincided with a DDoS attack on their infrastructure.

Now, given the size of Blizzard, you'd have to assume (hope) that Vodafone Limited (AS5378) have a PNI (Private Network Interconnect) with Blizzard.

... perhaps Blizzard were getting DDoSed, decided to shut down some of their PNIs, perhaps they can only DDoS scrub on some links, not others, and if the DDoS was likely a botnet, chances are the bulk of the DDoS came via compromised devices on residential connections.

So when Blizzard perhaps shut down their PNIs to Vodafone, it is now sending traffic via LonAP ... demonstrating that Vodafones LonAP connectivity is not large enough to deal with such an issue.  But perhaps IS for every day usage.

... I'm not sure what the usual route is to Blizzard from Vodafone, as I'm not a gamer.  But it is certainly impacting everything (including not Blizzard) that is routed via LonAP IX.

[jdvega]

Seems the routes to OpenDNS also are via LonAP at the moment, so anyone using OpenDNS for resolvers is going to see patchy DNS resolution.

[jdvega]

Route to OpenDNS changed to LINX, not via LonAP, this resolved the issue accessing OpenDNS for about 3-4 minutes.

LonAP peer came back up again, packet loss is back to OpenDNS via LonAP.

[SebF]

I am also having the same issues:

1 <1 ms <1 ms <1 ms ZenWiFi_XT8-89A0 [192.168.1.1]
2 1 ms 1 ms 1 ms 192.168.0.1
3 7 ms 7 ms 7 ms 90.246.0.1
4 14 ms 14 ms 13 ms 63.130.172.39
5 15 ms * * lonap-ix.blizzardonline.net [5.57.81.160]
6 * * * Request timed out.
7 * * * Request timed out.
8 * * * Request timed out.
9 * 20 ms 20 ms 137.221.78.51
10 * 21 ms 21 ms 137.221.66.45
11 21 ms 21 ms 20 ms 185.60.112.158

Vodafone "wifi xperts" support told me the below at 13:33 today:

I am very sorry I just checked with my advisor now and there is no solution from our side

at the moments there is no issues from our side yet, but we are investigating