Communication between WBS & iBurst Customers

Do you want this thread replaced with a post-launch thread?

  • Start a new thread (i.e. wipe the slate clean) & close/lock this thread

    Votes: 19 86.4%

  • Leave it open for constructive criticism of WBS iBurst Customer Service & Support

    Votes: 3 13.6%
  • Total voters
    22
  • Poll closed .
Status
Not open for further replies.
The _reason_ for terrible bandwidth is high latency. This is what keeled Sentech, and this is what's keeling iBurst. When I arranged the Sentech wardrives I plotted yards and yards of data using tcp session analysis tools and the verdict was all the same:

If you'd like to know more, read "It's the Latency, Stupid" to get a bit of a grasp of how ACK latency chokes bandwidth and the reasons for it.

http://rescomp.stanford.edu/~cheshire/rants/Latency.html

Less latency = better bandwidth. Of course, bandwidth still matters, but a big fat 1meg pipe will never perform to it's capacity if it's tied to high latency media, or a high latency round-trip.

Packet loss is bad too, but I think this should be enough reading material for one day.

I think what Gatecrasher was refering to wasn't TCP ACK's as such but the PHY (hardware layer) ACK's that the tower has to transmit to a modem to acknowledge the receipt of the previous frame (and vice versa). On a physical layer, if these physical ACK's dont happen, wireless devices attempt to retransmit the lost frame a few times, before reporting an actual lost packet to the network layer. This makes wireless links less packet-loss prone, but sometimes more latent, depending on the signal quality. His assumption that a less noisy path, or a "better signal" is better for the physical transmission of the iBurst medium is definately correct.

Regardless of the signalling used TDMA, or CDMA - most wireless physical layers are built on the above principles when it comes to acknowledging frame receipt and attempts to prevent packet loss on the physical layer.
 
ic said:
Ok Slim jy is mos baie slim en heeltemal korrek :).
Click to expand...
OK i'm not too good with afrikaans but that looked like you said i'm right all the time, king dingaling and all that jazz :)
 
TheRoDent said:
The _reason_ for terrible bandwidth is high latency.
Click to expand...
so true, and wireless has bad latency, its a fact all over the world. BTW interesting page you linked there.
 
Earlier in this thread I got this (MTU=1352, 1st hop 172.16.0.101)

Code: 
C:\Documents and Settings\...>ping www.iburst.co.za

Pinging www.iburst.co.za [196.30.31.120] with 32 bytes of data:

Reply from 196.30.31.120: bytes=32 time=63ms TTL=126
Reply from 196.30.31.120: bytes=32 time=62ms TTL=126
Reply from 196.30.31.120: bytes=32 time=62ms TTL=126
Reply from 196.30.31.120: bytes=32 time=62ms TTL=126

Ping statistics for 196.30.31.120:
    Packets: Sent = 4, Received = 4, Lost = 0 (0% loss),
Approximate round trip times in milli-seconds:
    Minimum = 62ms, Maximum = 63ms, Average = 62ms

C:\Documents and Settings\...>ping -f -l 1404 www.iburst.co.za


Pinging www.iburst.co.za [196.30.31.120] with 1404 bytes of data:

Packet needs to be fragmented but DF set.
Packet needs to be fragmented but DF set.
Packet needs to be fragmented but DF set.
Packet needs to be fragmented but DF set.

Ping statistics for 196.30.31.120:
    Packets: Sent = 4, Received = 0, Lost = 4 (100% loss),

C:\Documents and Settings\...>tracert www.iburst.co.za

Tracing route to www.iburst.co.za [196.30.31.120]
over a maximum of 30 hops:

  1    72 ms    93 ms    77 ms  172.16.0.101
  2    77 ms    77 ms    62 ms  196.30.31.102
  3    77 ms    62 ms    93 ms  www.iburst.co.za [196.30.31.120]

Trace complete.

Now I get this... again (MTU=1432, 1st hop 196.30.31.100):

Code: 
C:\Documents and Settings\...>ping www.iburst.co.za

Pinging www.iburst.co.za [196.30.31.120] with 32 bytes of data:

Reply from 196.30.31.120: bytes=32 time=91ms TTL=126
Reply from 196.30.31.120: bytes=32 time=109ms TTL=126
Reply from 196.30.31.120: bytes=32 time=46ms TTL=126
Reply from 196.30.31.120: bytes=32 time=62ms TTL=126

Ping statistics for 196.30.31.120:
    Packets: Sent = 4, Received = 4, Lost = 0 (0% loss),
Approximate round trip times in milli-seconds:
    Minimum = 46ms, Maximum = 109ms, Average = 77ms

C:\Documents and Settings\...>ping -f -l 1404 www.iburst.co.za

Pinging www.iburst.co.za [196.30.31.120] with 1404 bytes of data:

Reply from 196.30.31.120: bytes=1404 time=277ms TTL=126
Reply from 196.30.31.120: bytes=1404 time=389ms TTL=126
Reply from 196.30.31.120: bytes=1404 time=327ms TTL=126
Reply from 196.30.31.120: bytes=1404 time=311ms TTL=126

Ping statistics for 196.30.31.120:
    Packets: Sent = 4, Received = 4, Lost = 0 (0% loss),
Approximate round trip times in milli-seconds:
    Minimum = 277ms, Maximum = 389ms, Average = 326ms

C:\Documents and Settings\...>tracert www.iburst.co.za

Tracing route to www.iburst.co.za [196.30.31.120]
over a maximum of 30 hops:

  1    68 ms    93 ms   249 ms  196.30.31.100
  2    74 ms    93 ms    93 ms  196.30.31.102
  3    77 ms   218 ms   108 ms  www.iburst.co.za [196.30.31.120]

Trace complete.

So Shaun, why are the MTU's different? Assuming these are two different basestations, can the network (or my connection) be working optimally when different basestations have different configurations? Apart from which, I am now fed up of constantly changing my TCP settings...
 
Change to an MTU of 576 and you'll be sorted, without having to MTU hop all the time. It will however feel like being back on 56k, but that's all iBurst is promising with their "package" isn't it? 64K... ;)
 
haha and thats exactly what you get with them. adsl is awesome and if anyone other than telkom offered it.. I would take it in a second. But for now i'd rather be with iBurst which could or could not be crap in the future as apposed to telkom which is guaranteed to be crap forever.
 
Shaun Green said:
It is totally up to you as a client of the S.A. broadband offering as to what product you would like to use/support.

I would never knock any other B/B supplier or degrade any competition that we have so long required in our country.

If you have discovered that Iburst is not for you....Contact me directly.

State your reasons and lets us assist your conversion to whichever solution you find (in your opinion) better than Iburst.

I have found constant criticism from certain members.

I would not like to have a heart attack in trying to please the unpleasable.

Too many people hide behind their pseudo's ... Someday i mightmeet IC and not even realise it... I will not shoot first.... All opinions are valued. I would however like to know who i am speaking to.

Shaun Green
Click to expand...

Exact same tactic Sentech used to de-ball their clients
 
speakign of fownloading at crazy speeds GET ALL YOUR ISOS NOW PEOPLE, WINTERS COMING AND SO IS THE CAP
 
sihen said:
A question for you guys, have you seen a improvement in the call centre lately?
Click to expand...

No improvement as far as I can see. Haven't spoke to them in over 2 months. :D
 
call centre

sihen said:
A question for you guys, have you seen a improvement in the call centre lately?
Click to expand...
I have called 3 times in the past week (during normal office hours). Once you get a person on the other end, the service/knowledge/helpfulness is very good.

However, the time on 'hold' with music and adverts is far too long. One has the feeling after 8 to 10 minutes that one has been 'forgotten'. Could an operator not break in once in a while and reassure the caller?
 
90% of our calls in the last 2 weeks are being answered within 20 seconds. We have hired more staff and are training lots.
 
Call goes dead

It has happened three times to me in the last month that the call goes dead while waiting a long time. I must say that it has also happened about twice that my calls was answered under 5 sec.
 
Status
Not open for further replies.
Back
Top
Sign up to the MyBroadband newsletter
X