cancel
Showing results for 
Search instead for 
Did you mean: 
1

Ask

2

Reply

3

Solution

Gigafast routing issue

iahmed
2: Seeker
2: Seeker

Hello

 

During the pilot program latency from Peterborough was great to every where in the UK, France and Netherlands but over the last year or two, latency has gone significantly worse. 

 

Here's a test from 2018:

 

s03 (192.168.0.11)                                                                                                                                                                    2018-12-19T14:53:09+0000
Keys:  Help   Display mode   Restart statistics   Order of fields   quit
                                                                                                                                                                      Packets               Pings
 Host                                                                                                                                                               Loss%   Snt   Last   Avg  Best  Wrst StDev
 1. 192.168.0.1                                                                                                                                                      0.0%    17    0.6   0.5   0.4   0.6   0.1
 2. 212.158.250.34                                                                                                                                                  11.7%    17   32.4   8.6   5.4  32.4   7.9
 3. 63.130.105.130                                                                                                                                                  25.0%    16    5.9   6.1   5.8   6.4   0.2
 4. 195.89.96.1                                                                                                                                                      0.0%    16    6.4   6.8   5.7  13.6   1.9
 5. 195.2.28.166                                                                                                                                                     0.0%    16   12.1  12.7  11.8  17.2   1.5
 6. 195.2.8.193                                                                                                                                                      0.0%    16   15.8  19.9  14.9  79.5  16.0
 7. 195.2.15.122                                                                                                                                                     0.0%    16   21.9 113.4  16.4 304.5  96.2
 8. 23.216.202.232                                                                                                                                                   0.0%    16   15.3  15.2  14.9  15.4   0.1

Here's a trace now:

 

 

# mtr --no-dns --report 23.216.202.232
Start: 2021-04-27T05:16:22+0100
HOST: unifi Loss% Snt Last Avg Best Wrst StDev
1.|-- 172.17.0.1 0.0% 10 0.5 1.0 0.5 5.2 1.5
2.|-- 90.247.64.1 0.0% 10 10.6 10.6 10.3 11.2 0.2
3.|-- 63.130.105.130 0.0% 10 11.7 11.6 11.4 11.9 0.2
4.|-- 90.255.251.71 0.0% 10 12.2 15.2 11.3 32.5 7.0
5.|-- 23.210.50.46 0.0% 10 88.2 90.3 85.7 101.1 5.6
6.|-- 95.100.192.244 0.0% 10 84.2 85.0 83.4 90.1 2.5
7.|-- 95.100.192.131 0.0% 10 99.7 90.1 87.5 99.7 3.8
8.|-- 95.100.192.205 0.0% 10 91.7 92.3 91.3 96.6 1.6
9.|-- 23.210.61.37 0.0% 10 20.3 22.8 19.7 38.8 6.1
10.|-- 23.210.61.141 0.0% 10 384.4 64.6 20.3 384.4 113.1
11.|-- 23.216.202.232 0.0% 10 20.0 19.8 19.7 20.0 0.1

 

 

It starts with the first hop, latency is double what it used to be and double the latency to Amsterdam and France than what it used to be. I have ran numerous test at varying times in the day, night and overnight to rule out congestion. It has been consistent for at least 1 year.

 

I'd appreciate if someone from the Vodafone team could look into this.

13 REPLIES 13

Cynric
16: Advanced member
16: Advanced member

If you are using the Vodafone DNS then change to another DNS provider and test again. The Vodafone DNS is known to have issues.

Jayach
16: Advanced member
16: Advanced member

The OP is using I.P. addresses, so no DNS involved.

Cynric
16: Advanced member
16: Advanced member

I didn't know whether the OP was using IPs outside of the examples, so I thought that it may be an idea.

Anonymous
Not applicable

Okay so here are a few pointers:

 

Vodafone is only responsible for their leg of your data getting to you - there can be hold-ups elsewhere on the internet.  In the example, you give us it would seem that if there are problems they appear to exist with servers belonging to Akamai.

 

If you use tools that look at the latency for individual hops on a route, adding up these latencies will give a figure massively larger than simply pinging the endpoint (see https://www.cisco.com/c/en/us/support/docs/ios-nx-os-software/ios-software-releases-121-mainline/127... for a full explanation).  It's simply the endpoint latency that you need to consider.  

 

In your example, the routing has gotten more complex, but that complexity is outside of VFs network and with Akamai's network - also I only see an increase from 15ms to 20ms, which is hardly earth-shattering (and is sufficiently small that even traffic variations depending on the time of day could account for differences)!

 

One final point is that Akamai's network can behave in ways that can seem odd.  As a company, they provide services to third parties such as co-locationing and load-balancing.  Once your target is within their network two external devices pinging the same IP address may not even be hitting the same server at the same location (or even servers in the same country) and hence have different latencies.  *It would be interesting to perform the same traceroute using IPv6 addresses!

iahmed
2: Seeker
2: Seeker

Thank you for your responses but unless you are a network engineer like myself you won’t understand the issue but I appreciate the effort. 

Cynric
16: Advanced member
16: Advanced member

The figures are not good and I get similar results too. The thing is that these systems are not Vodafone ones (note to editor: What am I doing? I am defending VF, call a doctor.) and so VF a likely to say that there's isn't much that they can do about it.

 

The Akamai forum has people asking the same question

https://community.akamai.com/customers/s/question/0D50f00006n9eTjCAI/is-any-one-experiencing-any-lat...

 

You’re misunderstanding the question and the test result. The first hop is the issue which is going through Vodafone’s network. 

Cynric
16: Advanced member
16: Advanced member

I'm not sure it is inside VFs domain. (Update. 63.130.xxx.xxx is VF Americas according to ARIN. That's odd routing).

 

C:\>pathping 23.216.202.232

Tracing route to a23-216-202-232.deploy.static.akamaitechnologies.com [23.216.202.232]
over a maximum of 30 hops:
  0  FX6100.lan [192.168.0.42]
  1  READYSHARE [192.168.0.1]
  2  192.168.1.1
  3  host-212-158-250-37.dslgb.com [212.158.250.37]
  4  63.130.105.130
  5  90.255.251.71
  6  ae2.r01.lon03.icn.netarch.akamai.com [23.210.50.36]
  7  ae10.r02.ams01.icn.netarch.akamai.com [95.100.192.255]
  8  ae0.r01.ams01.icn.netarch.akamai.com [95.100.192.190]
  9  ae4.r01.dus01.icn.netarch.akamai.com [95.100.192.205]
 10  ae7.r01.dus01.ien.netarch.akamai.com [23.210.61.35]
 11  ae2.ecix-dus.netarch.akamai.com [23.210.61.129]
 12  a23-216-202-232.deploy.static.akamaitechnologies.com [23.216.202.232]

Computing statistics for 300 seconds...
            Source to Here   This Node/Link
Hop  RTT    Lost/Sent = Pct  Lost/Sent = Pct  Address
  0                                           FX6100.lan [192.168.0.42]
                                0/ 100 =  0%   |
  1    4ms     0/ 100 =  0%     0/ 100 =  0%  READYSHARE [192.168.0.1]
                                0/ 100 =  0%   |
  2    3ms     0/ 100 =  0%     0/ 100 =  0%  192.168.1.1
                                0/ 100 =  0%   |
  3  155ms     0/ 100 =  0%     0/ 100 =  0%  host-212-158-250-37.dslgb.com [212.158.250.37]
                                0/ 100 =  0%   |
  4  ---     100/ 100 =100%   100/ 100 =100%  63.130.105.130
                                0/ 100 =  0%   |
  5  ---     100/ 100 =100%   100/ 100 =100%  90.255.251.71
                                0/ 100 =  0%   |
  6   94ms     0/ 100 =  0%     0/ 100 =  0%  ae2.r01.lon03.icn.netarch.akamai.com [23.210.50.36]
                                0/ 100 =  0%   |
  7   99ms     0/ 100 =  0%     0/ 100 =  0%  ae10.r02.ams01.icn.netarch.akamai.com [95.100.192.255]
                                0/ 100 =  0%   |
  8   98ms     0/ 100 =  0%     0/ 100 =  0%  ae0.r01.ams01.icn.netarch.akamai.com [95.100.192.190]
                                0/ 100 =  0%   |
  9   99ms     0/ 100 =  0%     0/ 100 =  0%  ae4.r01.dus01.icn.netarch.akamai.com [95.100.192.205]
                                0/ 100 =  0%   |
 10   30ms     0/ 100 =  0%     0/ 100 =  0%  ae7.r01.dus01.ien.netarch.akamai.com [23.210.61.35]
                                0/ 100 =  0%   |
 11   48ms     0/ 100 =  0%     0/ 100 =  0%  ae2.ecix-dus.netarch.akamai.com [23.210.61.129]
                                0/ 100 =  0%   |
 12   27ms     0/ 100 =  0%     0/ 100 =  0%  a23-216-202-232.deploy.static.akamaitechnologies.com [23.216.202.232]

Trace complete.

C:\>

 

[Removed]

 

It sounds like you performed a whois on an IP address. This does not give you the location of the IP address, it is merely the owner/contact information of the ASN.

 

Please, since this question is not in your skill set I kindly request that you don't provide any more responses.

[MOD EDIT: This post has been edited to remove off topic/personal information/inappropriate content please see Community Guidelines]