Telkom & Mweb peering again to some extent?

P

Pada

Executive Member
Joined
Feb 18, 2009
Messages
8,192
Reaction score
6
Location
Stellenbosch
I'm just starting a new thread regarding this great news, since most of you might've missed: http://mybroadband.co.za/vb/showthr...ers-Feedback?p=6789958&viewfull=1#post6789958

Here's a trace route from my Mweb ADSL connection to www.Telkom.co.za :D
Code: 
Tracing route to www.telkom.co.za [196.43.22.222]
over a maximum of 30 hops:

  1     1 ms    <1 ms    <1 ms  mikrotik [192.168.11.1]
  2    10 ms    11 ms    11 ms  41-133-108-1.dsl.mweb.co.za [41.133.108.1]
  3    12 ms    12 ms    12 ms  te1-1-0.502.cpt-ipc-1.mweb.co.za [196.28.178.166]
  4    12 ms    12 ms    11 ms  vlan92.cpt-hscore-1.mweb.co.za [196.28.179.11]
  5    14 ms    14 ms    30 ms  tengige0-2-0-0.11.cpt-up-1.mweb.co.za [196.28.178.97]
  6    85 ms    54 ms    12 ms  197-84-4-21.cpt.mweb.co.za [197.84.4.21]
  7    14 ms    41 ms    27 ms  196.43.25.205
  8    35 ms    34 ms    58 ms  196.43.39.162
  9    34 ms    32 ms    33 ms  rrba-ip-bssr-1-ge-2-48.telkom-ipnet.co.za [196.43.23.6]
 10    36 ms    36 ms    36 ms  nbsc-ip-bssr-1-atm-5-0-0-1.telkom-ipnet.co.za [196.43.23.30]
 11    38 ms    38 ms     *     nbsc-ip-rcache-1-vif2.telkom-ipnet.co.za [196.43.22.222]
 12    38 ms    38 ms    48 ms  nbsc-ip-rcache-1-vif2.telkom-ipnet.co.za [196.43.22.222]

Trace complete.

^ I don't see any CINX hops though.

I presume that Telkom clients should now be able to access Mweb hosted sites with local only cap, like news24.co.za & supersport.co.za !?
 
I have a telkom account and could play on mweb servers with 70 ping last night :0
 
Last edited:
Wonder which company gave in, both had something to lose imo.
 
the faaaakers co-located it seems. Somebody slapped a cisco in there. YEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEE!

root@N30:/home/k1dbl4ck# nmap -O 196.43.25.205

Starting Nmap 5.00 ( http://nmap.org ) at 2011-09-05 00:18 SAST
Interesting ports on 196.43.25.205:
Not shown: 999 closed ports
PORT STATE SERVICE
1720/tcp filtered H.323/Q.931
Warning: OSScan results may be unreliable because we could not find at least 1 open and 1 closed port
Device type: router|specialized|switch|broadband router|WAP
Running (JUST GUESSING) : Cisco IOS 12.X|11.X (96%), Cisco CatOS (96%), Huawei embedded (95%), ZyXEL ZyNOS 3.X (95%)
Aggressive OS guesses: Cisco 1812, 3640, or 3700 router (IOS 12.4) (96%), Cisco 836, 1751, 1841, or 2800 router (IOS 12.4) (96%), Cisco 870 router (IOS 12.4) (96%), Cisco DOCSIS cable modem termination server (IOS 12.1) (96%), Cisco Catalyst 6500-series switch (IOS 12.1) (96%), Cisco Catalyst WS-C2950G switch (96%), Cisco 2900XL switch (IOS 11.2) (96%), Cisco 1700 router (IOS 12.0) (96%), Cisco 2509 router (IOS 12.3) (96%), Cisco 2514 router (IOS 12.1) (96%)
No exact OS matches for host (test conditions non-ideal).

OS detection performed. Please report any incorrect results at http://nmap.org/submit/ .
Nmap done: 1 IP address (1 host up) scanned in 7.23 seconds
 
Pity to see the peering is only taking place in Cape Town (pings still aren't great) - but at least local-only peering works. Telkom probably succumbed because of Consumer Protection Act pressure - that they advertise local accessibility yet guys couldn't get that local service when capped
 
The peering link look very congested at the moment
Code: 
                                                               Packets               Pings
 Host                                                        Loss%   Snt   Last   Avg  Best  Wrst StDev
 1. vz4-jnb.vps.za-dns.com                                    0.0%    12    0.0   0.0   0.0   0.1   0.0
 2. 65.75.28.196.netactive.net                                0.0%    12    0.6   0.7   0.5   1.2   0.2
 3. 129.75.28.196.netactive.net                               0.0%    12    1.5   1.0   0.7   1.6   0.3
 4. tengig-1-6.vic-hscore-1.mweb.co.za                        0.0%    12    1.6   1.3   0.8   2.6   0.6
 5. TenGigE0-0-0-0.11.vic-up-1.mweb.co.za                     0.0%    12    7.0   4.7   1.3   8.9   2.1
 6. gig-1-0-2.cpt-hscore-3.mweb.co.za                         0.0%    12   23.2  24.9  22.9  29.6   2.0
 7. 197-84-4-21.cpt.mweb.co.za                                0.0%    11  977.0 971.0 916.9 1029.  38.6
 8. 196.43.25.205                                             0.0%    11  959.1 968.3 898.1 1015.  41.2
 9. wblv-ip-essr-1-atm-2-0-0-2.telkom-ipnet.co.za             0.0%    11  972.9 970.6 899.7 1021.  41.8
10. wblv-ip-www-1.saix.net                                    9.1%    11  993.9 973.0 921.9 1022.  39.8
 
Last edited:
schumi said:
Wonder which company gave in, both had something to lose imo.
Click to expand...

Considering telkom hosts nothing of importance, what reason would mweb have to give in?

Sent from my GT-I9000 using MyBroadband Android App
 
Shayd said:
Considering telkom hosts nothing of importance, what reason would mweb have to give in?

Sent from my GT-I9000 using MyBroadband Android App
Click to expand...
That my friend as a archaic way of thinking. I believe network models and business models have changed significantly is the past decade to such an extent that reasoning isn't as simple anymore. For example we push more traffic than we pull to certain networks, and still they still don't want to peer.

I can see a lot of M-Web client's want better RDP access to their Office Networks that are on SAIX, just as one example that is.

Then there is also, why would M-Web want to hand off traffic in London if they could hand it off in South Africa instead. Technically speaking, there are many reasons why both networks would want to peer, politically speaking....many reasons for both to not want to peer...

Anyways, looking at the latencies on that Local link, which seems to be M-Web backhaul that is congested, I'd say there are A LOT more traffic between the 2 networks than both expected, at least in one of the directions.
 
Last edited:
just did a trace on SAIX forum and didn't come in very low.

in the 1000ms+

can any1 do a double check for me? games.saix.net
 
Frankly I would say pinky set this I'm motion. There was no benefit to either company to not peer and this internet berlin wall served no purpose other than to make the telkom execs feeling important.

Peer for free and help both companies seems the rational option. Particularly seeing that I am willing to bet that telkom is a net traffic consumer.

Sent from my GT-I9000 using MyBroadband Android App
 
Tinuva said:
Looks like the peering link in JHB is congested. I'd wager that the CT guys to SAIX CT ips are fine however.
Click to expand...
Yup, although even the CT link seems to be slightly congested.

Trace to wbs-ip-dns.saix.net:
Code: 
                                       Packets               Pings
 Host                                Loss%   Snt   Last   Avg  Best  Wrst StDev
 1. mikrotik                          0.0%    13    0.9   0.9   0.8   0.9   0.0
 2. 41-133-108-1.dsl.mweb.co.za       0.0%    13   11.5  10.9  10.2  11.5   0.4
 3. te1-1-0.511.cpt-ipc-1.mweb.co.za  0.0%    13   12.7  12.1  11.6  12.7   0.3
 4. vlan92.cpt-hscore-1.mweb.co.za    0.0%    13   11.5  12.1  11.5  12.8   0.4
 5. tengige0-2-0-0.11.cpt-up-1.mweb.  0.0%    13   14.3  25.4  12.6 104.2  26.4
 6. 197-84-4-21.cpt.mweb.co.za        0.0%    13   69.6  24.7  11.9  69.6  18.8
 7. 196.43.25.205                     0.0%    13   68.0  27.0  12.2  68.0  18.2
 8. wbs-ip-esr-2-gig-2-0-0.telkom-ip  0.0%    13   73.6  29.3  13.2  73.6  19.8
 9. 196.43.23.62                      0.0%    13   39.3  23.1  12.4  45.8  12.9
10. ???

vs. trace to www.telkom.co.za:
Code: 
                                       Packets               Pings
 Host                                Loss%   Snt   Last   Avg  Best  Wrst StDev
 1. mikrotik                          0.0%    17    0.9   0.9   0.9   1.1   0.1
 2. 41-133-108-1.dsl.mweb.co.za       0.0%    17   10.6  10.8  10.0  11.6   0.4
 3. te1-1-0.502.cpt-ipc-1.mweb.co.za  0.0%    16   12.2  12.2  11.5  12.8   0.4
 4. vlan92.cpt-hscore-1.mweb.co.za    0.0%    16   12.1  12.4  11.7  15.2   0.8
 5. tengige0-2-0-0.11.cpt-up-1.mweb.  0.0%    16   16.1  16.4  13.7  33.9   4.8
 6. 197-84-4-21.cpt.mweb.co.za        0.0%    16   41.8  64.0  11.9 147.8  44.3
 7. 196.43.25.205                     6.2%    16   28.9  62.3  12.5 141.9  46.4
 8. 196.43.39.162                    12.5%    16  1009. 1012. 969.1 1049.  20.9
 9. rrba-ip-bssr-1-ge-2-48.telkom-ip  0.0%    16  1005. 1011. 976.7 1051.  21.1
10. nbsc-ip-bssr-1-atm-5-0-0-1.telko  6.2%    16  1007. 1015. 978.2 1060.  22.1
11. nbsc-ip-rcache-1-vif2.telkom-ipn  0.0%    16  1006. 1015. 981.1 1061.  20.6
 
They're peering but probably severely underestimated the traffic between MWEB and Telkom. This is a traceroute to telkom.co.za from a server hosted at MWEB JHB

Code: 
[root@mgmt ~]# tracert telkom.co.za
traceroute to telkom.co.za (196.43.22.222), 30 hops max, 40 byte packets
 1  vz2-jnb.vps.za-dns.com (196.2.147.180)  0.050 ms  0.020 ms  0.015 ms
 2  65.75.28.196.netactive.net (196.28.75.65)  0.806 ms  0.819 ms  0.897 ms
 3  129.75.28.196.netactive.net (196.28.75.129)  1.080 ms  1.093 ms  1.353 ms
 4  tengig-1-6.vic-hscore-1.mweb.co.za (196.22.169.65)  1.336 ms * tengig-1-6.vic-hscore-1.mweb.co.za (196.22.169.65)  1.338 ms
 5  TenGigE0-0-0-0.11.vic-up-1.mweb.co.za (196.22.169.225)  4.765 ms  4.782 ms  4.783 ms
 6  197-80-4-117.jhb.mweb.co.za (197.80.4.117)  906.748 ms  906.772 ms  906.773 ms
 7  196.43.25.137 (196.43.25.137)  906.731 ms  904.762 ms  904.703 ms
 8  196.43.39.166 (196.43.39.166)  904.674 ms  905.792 ms  905.716 ms
 9  rrba-ip-bssr-1-ge-2-48.telkom-ipnet.co.za (196.43.23.6)  906.777 ms  905.208 ms  905.211 ms
10  nbsc-ip-bssr-1-atm-5-0-0-1.telkom-ipnet.co.za (196.43.23.30)  911.405 ms  911.513 ms  911.519 ms
11  nbsc-ip-rcache-1-vif2.telkom-ipnet.co.za (196.43.22.222)  908.435 ms  908.392 ms *
 
Hummercellc said:
Peering anywhere in SA it better then peering through London.....
Click to expand...
That really depends. I'd prefer 400ms (via London) over 1100ms with packet loss, where as I'd assume that people on Telkom local only cap would prefer 1100ms with packet loss...
 
Pada said:
That really depends. I'd prefer 400ms (via London) over 1100ms with packet loss, where as I'd assume that people on Telkom local only cap would prefer 1100ms with packet loss...
Click to expand...

I doubt that would be the case, they probably just turned on the peering and the high pings and packet loss are just teething issues. MWEB did do network maintenance this past weekend..
 
You must log in or register to reply here.
Back
Top
Sign up to the MyBroadband newsletter
X