Telkom & backbone faults

S

snail112

Expert Member
Joined
Apr 18, 2007
Messages
1,671
Reaction score
2
Location
Pretoria
Question:

How does this Telkom backbone faults work exactly?

Scenario:

"Bob and Koos are on the same exchange, and they live about within 10km from each other. Bob has no problems with his OpenWeb account. But Koos on the other hand, experiences slow download speeds, high latency. Bob and Koos switch to a backup account, everything is fine - full line speed , low latency. They switch back to their OpenWeb accounts and Koos suffers again with the same problem."

The answer I got:

All ADSL traffic goes over Telkom's network. If they have a problem it will affect us as well. - the answer does not make sense, I know ADSL is Telkom network, but isn't it an ISP problem?

What I don't understand: Why does Koos suffer on the IS backbone, but Bob, who is also on the same backbone have no problem? They are on the same exchange, so they would be effected in the same way if the backbone is the same?

Shouldn't they get the same results? Is this the difference between Gold Uncapped and the other Uncapped packages?

I have no grudge towards anyone. I am just trying to understand how Telkom and backbone faults work - how 2 users on the same backbone (IS) and same exchange experience different connectivity.
 
Last edited:
There is lots of roads from one exch to another, until it gets to the isp mainframe. So if one off the roads got a incident, and is block, and no alternative, then traffic piles up. But on the other road traffic flows.
Hope you understand.
Also some customers buy a dedicated line on the road [more money] and even if this road is block they want to have there own lane on another road [even more money].
 
Being on the same exchange doesn't mean they are on the same DSLam (multiplexer) as well. There are lot of DSLams in the exchange, each of them containing multple "swords". Some DSLams might be congested, some might have networking trouble, some are older hardware, some are on older software. So there really is no guarantee Bob and Koos will ever have the same experience, even if they were next door neighbors. In fact most of our problems are caused in our local exchanges, long before our individual "roads" even hit any backbone.

That's also why I'm sceptical about new superduper underseas cables which reach SAfrican shores. All the awsumsauce they deliver is being messed up by some badly trained and worse organized techies at your local exchange. When Seacom cable breaks in the middle of the mediterranean, those providers are mostly able to fix it in a matter of hours. When your personal line to the exchange breaks down it takes Telkom weeks and months and years to fix it.
 
Read this article - it is spot on! http://www.businessday.co.za/Articles/Content.aspx?id=151146

Lack of preventative maintenance, techies with false reporting to meet targets, lack of pride, loss of skills, lack of experience, lack of training, lack of ownership at ground level, poor decision making, lack of accountability, etc. all contribute to poor network performance.

Some folks may say that Telkom is overstaffed; I argue that the total number of staff might be sufficient BUT their staff distribution is skewered. They need more technical staff to maintain the network (indoor and outdoor) and fewer executives in their Multi-Tower HO in Pretoria.

My insider friends tell me that Telkom must be one of the companies with the most internal measurements and reports in the world. To cut costs they retrenched plenty of support staff but then shifted their work to the technical supervisors. So instead of doing technical work they spend more than half their day with admin type work!! Wasted energy, I say. As per the article, Telkom need to start trusting their staff and policing them less and then more actual work will get done.
 
Last edited:
MickeyD said:
Read this article - it is spot on! http://www.businessday.co.za/Articles/Content.aspx?id=151146

Lack of preventative maintenance, techies with false reporting to meet targets, lack of pride, loss of skills, lack of experience, lack of training, lack of ownership at ground level, poor decision making, lack of accountability, etc. all contribute to poor network performance.

Some folks may say that Telkom is overstaffed; I argue that the total number of staff might be sufficient BUT their staff distribution is skewered. They need more technical staff to maintain the network (indoor and outdoor) and fewer executives in their Multi-Tower HO in Pretoria.

My insider friends tell me that Telkom must be one of the companies with the most internal measurements and reports in the world. To cut costs they retrenched plenty of support staff but then shifted their work to the technical supervisors. So instead of doing technical work they spend more than half their day with admin type work!! Wasted energy, I say. As per the article, Telkom need to start trusting their staff and policing them less and then more actual work will get done.
Click to expand...
Couldn't agree more ...
 
MickeyD said:
Read this article - it is spot on! http://www.businessday.co.za/Articles/Content.aspx?id=151146

Lack of preventative maintenance, techies with false reporting to meet targets, lack of pride, loss of skills, lack of experience, lack of training, lack of ownership at ground level, poor decision making, lack of accountability, etc. all contribute to poor network performance.

Some folks may say that Telkom is overstaffed; I argue that the total number of staff might be sufficient BUT their staff distribution is skewered. They need more technical staff to maintain the network (indoor and outdoor) and fewer executives in their Multi-Tower HO in Pretoria.

My insider friends tell me that Telkom must be one of the companies with the most internal measurements and reports in the world. To cut costs they retrenched plenty of support staff but then shifted their work to the technical supervisors. So instead of doing technical work they spend more than half their day with admin type work!! Wasted energy, I say. As per the article, Telkom need to start trusting their staff and policing them less and then more actual work will get done.
Click to expand...

+1 Billion YEN
Post of the month, if only the executives at Telkom would actually take this advice...
 
snail112 said:
Thanks guys, now it make sense.
Click to expand...
Well, its actually a lot more complicated than it may seem. If you undertand the technical mechanics of the entire access path, you'll see there many reasons for performance variance.

As has been mentioned, with the exception of the smallest exchanges, they all run multiple DSLAMs (e.g. big ones can have up to 30). Each DSLAM is in turn uplinked (virtually) to its parent broadband aggregation router (ESR) to terminate the user's IP session (1st hop). The DSLAMs in a single exchange may be uplinked different ESRs (which may not even be located in the same major aggregation exchange), so there is room for variance here.

Then on any particular ESR, each ISP has their own VRF (virtual routing instance) which has to transfer packets across Telkom's IPnet backbone to that particular ISPs IPC links (in different locations). Again here there is room for variance as the end-to-end path of one ISP from a ESR could be performing fine, while another's has some issue or is congested.

Then finally, the larger ISPs run multiple IPC links which may be fed from different exchanges (local to the ISPs location). Again more room for variance as traffic is generally balanced by destination IP address, so one user can be going through an IPC link that is performing fine (by virtue of allocated IP), while another can be going through one that has issues or is congested due to traffic imbalance.

The permuations are many, so its very possible that 2 users on the same exchange using the same ISP can get varying performance results.
 
MickeyD said:
^ technically 100% correct ^
Click to expand...
Yeah, thought of another one. Even in the larger/newer exchanges that are metro ether fed, there will still be legacy DSLAMs that are uplinked via ATM, so the transport path between DSLAM & ESR could be totally different. Again more room for performance variance.
 
Roman4604 said:
Yeah, thought of another one. Even in the larger/newer exchanges that are metro ether fed, there will still be legacy DSLAMs that are uplinked via ATM, so the transport path between DSLAM & ESR could be totally different. Again more room for performance variance.
Click to expand...
Still plenty on ATM. Perversely enough the launch of 8ta has accelerated the ME rollout and this is to the benefit of ADSL users!
 
Roman4604 said:
<brilliant post> The permuations are many, so its very possible that 2 users on the same exchange using the same ISP can get varying performance results.
Click to expand...

The overwhelmingly majority of "backbone" faults will probably be cable related, whether it be congestion, damage, breakage or obsolescence. Most of this is directly due to the lack of preventative maintenance; a S.O.P. that was stopped to cut costs a few years ago.
 
OK, basically, (in plain English): although 2 people are on the same exchange, they might not be on the same DSLAM, which also mean they are not on the same "pipe" to the exchange; and the different IP address range might have an impact how they experience the interwebs on backbone X?
 
snail112 said:
OK, basically, (in plain English): although 2 people are on the same exchange, they might not be on the same DSLAM, which also mean they are not on the same "pipe" to the exchange; and the different IP address range might have an impact how they experience the interwebs on backbone X?
Click to expand...
... can also be on different "pipes" out of the exchange as well (into the 'cloud').
 
You must log in or register to reply here.
Back
Top
Sign up to the MyBroadband newsletter
X