International Capacity of MWEB/Vumatel FTTH - a concern. "4mb ADSL vs 20mb FTTH"

[OUPA]MrNutz

Expert Member
Joined
Jan 21, 2005
Messages
1,788
Hi Guys - I need a bit of constructive thoughts and opinions. My CEO (South Korean) requested FTTH so that he and his family can watch South Korean media streaming from SEOUL. He is however complaining that his FTTH is not 1 inch better for that specific kind of service vs ADSL.

Either the international capacity is really poor on both platforms or the links are saturated.

FTTH (20mbps) : (carries traffic around earth via eastern routes to S.Korea)

When doing a local speedtest , 19-20mbps both up and down is easily achieved on speedtest.net from using a JHB based server.
When using connectivity test to UK , 10.45mbps down and 6.93mbps up is achieved.
When using connectivity test to Baltimore(New York) , 6.97mbps down and 1.21mbps up is achieved.
When using connectivity test to SEOUL , 3.94-4.43mbps down and 0.34-0.46mbps up is achieved (which confirms my suspicion about poor backhaul capacity between asian partners)
When using torrent traffic (p2p) , 19.45mbps down is achieved.

ADSL (4mbps) : (carries traffic westerly via UK->US->to S.Korea)

When doing local speedtest , 3.83mbps down 0.42mbps up is achieved.
When using connectivity test to Baltimore(New York) , 3.76mbps down and 0.33mbps up is achieved.
When using connectivity test to UK , 3.4mbps down and 0.35mbps up is achieved
When testing to SEOUL , 3.20mbps down and 0.30mbps up.
When using torrent traffic (p2p) , 4.59mbps down is achieved.

No doubt that FTTH is faster locally , but internationally - something just seems...wrong?
How can one attempt to constructively improve this and achieve more international speed to S.Korea on FTTH?
I've logged a call with MWEB.
 

XenitXTD

Well-Known Member
Joined
Sep 26, 2014
Messages
104
I am with crystal web for fiber... and while i havent tested asian sources i can at least say i retain 60-80% of my speed to the US in speed tests... lowest being 60% highest was probably like 85% but those are momentary glitch attempts at FTL... FTL drives are still not available yet.

But yeah... I was once MWEB... will never ever go back to them... ever since they sold out it was the dawn of their decline... i watched that decline as a customer of theirs...

I mostly went with Crystal because they have usenet news servers and all my buddies with them had good experiences and I can officially say I too am in good experience so far with them.

I would say to make a request for people on different ISP's to test and see what they get so that you can just see what figures look like with different ISP's so you can see that its not the same and on which ISP's it is.
 

Mr Scratch

Expert Member
Joined
May 15, 2013
Messages
4,838
Check how your ISP peers with whoever is hosting the SK content, and preferably choose an ISP that has a peer that has presence in an SK datacentre.
 

PBCool

Cool Ideas Rep
Company Rep
Joined
Jan 11, 2016
Messages
13,304
[OUPA]MrNutz;20674454 said:
Hi Guys - I need a bit of constructive thoughts and opinions. My CEO (South Korean) requested FTTH so that he and his family can watch South Korean media streaming from SEOUL. He is however complaining that his FTTH is not 1 inch better for that specific kind of service vs ADSL.

Either the international capacity is really poor on both platforms or the links are saturated.

FTTH (20mbps) : (carries traffic around earth via eastern routes to S.Korea)

When doing a local speedtest , 19-20mbps both up and down is easily achieved on speedtest.net from using a JHB based server.
When using connectivity test to UK , 10.45mbps down and 6.93mbps up is achieved.
When using connectivity test to Baltimore(New York) , 6.97mbps down and 1.21mbps up is achieved.
When using connectivity test to SEOUL , 3.94-4.43mbps down and 0.34-0.46mbps up is achieved (which confirms my suspicion about poor backhaul capacity between asian partners)
When using torrent traffic (p2p) , 19.45mbps down is achieved.

ADSL (4mbps) : (carries traffic westerly via UK->US->to S.Korea)

When doing local speedtest , 3.83mbps down 0.42mbps up is achieved.
When using connectivity test to Baltimore(New York) , 3.76mbps down and 0.33mbps up is achieved.
When using connectivity test to UK , 3.4mbps down and 0.35mbps up is achieved
When testing to SEOUL , 3.20mbps down and 0.30mbps up.
When using torrent traffic (p2p) , 4.59mbps down is achieved.

No doubt that FTTH is faster locally , but internationally - something just seems...wrong?
How can one attempt to constructively improve this and achieve more international speed to S.Korea on FTTH?
I've logged a call with MWEB.
Do you have an IP of the service in South Korea?

These things often aren't that simple. Like for instance most of Dubai routes their traffic to London via the US due to political reasons. Even though they are far closer to EU.
 

[OUPA]MrNutz

Expert Member
Joined
Jan 21, 2005
Messages
1,788
Thx Guys - They want to use ikoreantv.com [211.110.44.163] . And it seems that indeed it is SK , which is better than KT.
 

ponder

Honorary Master
Joined
Jan 22, 2005
Messages
92,825
You should probably try asking people here on different isps to do a traceroute and download a linux dvd from a korean mirror to gauge speeds.

On openserve that ip routes via frankfurt-newyork-seattle
 
Last edited:

PBCool

Cool Ideas Rep
Company Rep
Joined
Jan 11, 2016
Messages
13,304
[OUPA]MrNutz;20676602 said:
Thx Guys - They want to use ikoreantv.com [211.110.44.163] . And it seems that indeed it is SK , which is better than KT.

On our network it traces via following, so also via US:

4 2 ms 1 ms 2 ms 154.0.1.66
5 2 ms 2 ms 2 ms 154.0.1.142
6 164 ms 164 ms 163 ms 154.0.1.106
7 164 ms 165 ms 164 ms v1000.core1.lon3.he.net [216.66.80.169]
8 194 ms 181 ms 163 ms 100ge14-1.core1.lon2.he.net [184.105.64.237]
9 229 ms 229 ms 238 ms 100ge4-1.core1.nyc4.he.net [72.52.92.166]
10 240 ms 267 ms 239 ms 100ge14-1.core1.tor1.he.net [184.105.80.10]
11 262 ms 261 ms 271 ms 100ge6-1.core1.ywg1.he.net [184.105.64.102]
12 278 ms 280 ms 274 ms 100ge10-1.core1.yyc1.he.net [184.105.222.98]
13 285 ms 287 ms 295 ms 100ge10-2.core1.yvr1.he.net [184.105.64.113]
14 287 ms 287 ms 287 ms 100ge10-2.core1.sea1.he.net [184.105.64.109]
15 * * * Request timed out.
16 517 ms 517 ms 513 ms 58.229.14.20
17 512 ms 504 ms 507 ms 39.115.132.46
18 513 ms 525 ms 487 ms 218.50.5.170
19 497 ms 499 ms 496 ms 116.125.121.38
20 * * * Request timed out.
21 * * * Request timed out.
22 * * * Request timed out.

The peering exchange in SK is KINX and HE has a port there, so it seems it's more about the host of the stream buys capacity directly from the states only, or something to that effect. I dont see you managing to get a stream from that server working easily, unless you can tunnel to a provider in SK that peers at KINX and has local access to that ISPs network.
 
Last edited:
Top