Advertisment: Broadband via satellite
Advertisment: Worldwide satellite services from Ground Control Company

www.satsig.net

Satellite Internet Forum.

Welcome, Guest.        Forum rules.
      Home            Login            Register          
Pages: 1

Data Volume Discrepencies

(Read 5969 times)
bixelps
Member
★★
Offline



Posts: 6
Feb 13th, 2010 at 2:20pm  
I operate a VSAT station using SkyEdge modem, 1.2M dish and 1W transmitter.  This system became operation in 2009.  Only IP traffic goes over this system.  The RX Eb/NO ranges for 8.5-11 depending on the weather.

My provider charges me based on the amount of data that is exchanged (TX+RX).  I have the modem connected to a Linux based router which provides services to several users at this site.

There is a discrepency between what the router measures as the data transmitted (RX+TX) and what the VSAT service provider is measuring.  The provider is telling me that I am sending 30% more data than the router is saying.  The provider seems to have very little technical knowledge on the subject and just claims that this is my error.  I am pretty confident in my Linux router which is directly connected to the modem.

So now I am feeling cheated.  Can someone here help me understand what my VSAT provider is measuring and how I might be seeing this type of difference?  The Gilat Modem also keeps statistics in the form of "Packets".  Is there anyway to translate packets into bytes so I would have another means to measure?

This station serves a not for profit hospital in Africa and has a pretty low budget for Sat. services.  Any advice would be greatly appreciated.

Thanks
Back to top
 
 
IP Logged
 
Oasis Networks
Senior Member
★★★
Offline



Posts: 232
Reply #1 - Feb 16th, 2010 at 7:39am  
Hello,

I would like to comment regarding couple of points.

The first one, without knowing who is the provider, I believe the way the provide measure the usage is more accurate than your Lunix router. If it were a Cisco router for example, I would have trusted the counters of the Cisco more than the measurements done by the provider.

Second, I understand that you are using the link mostly for IP traffic. In this case, it is recommended to reconsider using the Skyedge platform, which is excellent for voice traffic for small and remote locations..but if you just need a connection to the internet, I think other platforms could be more suitable.

Regards,
Nimrod
Back to top
 

www.oasisnetworks.net - Oasis Networks - Online with you!
IP Logged
 
bixelps
Member
★★
Offline



Posts: 6
Reply #2 - Feb 17th, 2010 at 2:20am  
The provider here is GlobalTT.com.

We use the link for IP traffic including VoIP.  We have some CIR dedicated for this purpose.

When it comes to the discrepency I was hoping to get some theory's about what is happening.  For instance, I do not know what overhead might be tacked onto my data to transmit on the sat link.  I also do not know what overhead the "HTTP Accelerator" techonolgy (RPA) might impose.  Also, how about retransmissions because of errors.  Is it possible that such a thing could cause this amount of discrepency?

As far as your Linux vs. Cisco comment I suppose everyone has what they prefer.  I have tested the Linux counters in the past and found them accurate.  Also I have seen that my SkyEdge modem has an HTTP interface in which it keeps track of the "Packets".  These counts agree with what the router is saying.

Certainly GlobalTT thinks their count is more accurate than mine.  But at the same time they do not offer me any data to back up their claim and cannot offer any explaination as to this discrepency.  This makes me doubt their claims.
Back to top
 
 
IP Logged
 
Oasis Networks
Senior Member
★★★
Offline



Posts: 232
Reply #3 - Feb 18th, 2010 at 2:13pm  
Hi,

30% really sounds too much, especially if you receive same data rate on your router and on your remote.

Just another thing worth to be checked:

What are the SNR values at the hub of your transmission? I have some good experience with the skyedge but I do not know exactly how the data rate is measured - what about retransmissions for example, caused because of low SNR or because of unavailable time slots? What kind of answers do you get from Globaltt? Any explanations or just a laconic answer?
Back to top
 

www.oasisnetworks.net - Oasis Networks - Online with you!
IP Logged
 
bixelps
Member
★★
Offline



Posts: 6
Reply #4 - Feb 18th, 2010 at 8:30pm  
I am not very experienced with Sat. networks so I have no idea what I can reasonably request of them in terms of data.   Maybe you can help out with that.

You said I should request the SNR at the hub.  I suppose this means that there is a separate SNR for each direction of transmission.  In my OP I listed the SNR I see at my modem which I am going to assume is for the signal I am receiving only.  So how about their end.  Is the SNR of interest measured by the satellite itself or at the ground station only?  Does the satellite interpret the signal at all or just retransmit it in analog form back to the ground station.  Any reference you could recommend?

I literally begged for support on this from GlobalTT and so far all I have gotten is that its my fault and they cannot help me.  I did get the response that I am charged for all retransmissions.  Since I do not understand how this works I can only assume this is a possible explanation.  I know how TCP/IP works and yes it does involve retransmissions but such retransmissions cannot be the source of this problem because they originate in the end terminal computers and so would be counted by my router.  The only type that could be responsible would be ones that are performed between my modem and their ground station without my knowledge.   

I read that TCP/IP connections are transmitted as UDP/IP packets over sat links because of the delays involved.  I wish I could learn more about how this works.

How much can I expect GlobalTT to know about this issue?  Are they just a reseller with no technical capability?  Do you have any idea what their relationship would be like with the satellite owner?

Sorry for bending your ear on this.  I appreciate your interest.

Paul
Back to top
 
 
IP Logged
 
Oasis Networks
Senior Member
★★★
Offline



Posts: 232
Reply #5 - Feb 19th, 2010 at 6:58am  
Dear Paul,

From your description, the situation is really puzzling and I wish someone from Globaltt would step in and provide more details from their end. In this business there are lot of virtual carriers, lousy resellers and etc...but Globaltt are certainly not one of those and one would expect more detailed explanation from their end regarding this situation.

Regarding the technical questions, here are some answers, sorry for being too general maybe, it just that I will try to explain the general idea rather than diving into details.

Regarding the SNR: Yes, for two-way communication there are different SNR values at each side of the link. The SNR values on your end reflect the quality of your reception of the hub TX signal; the same way, there are SNR values at the hub that reflect the quality of the hub's reception of YOUR TX transmission.

If you have good SNR but the SNR value at the hub is poor, it means that will be lot of errors on the path from your terminal to the hub, in other words, on your transmission path. If there are many errors, the terminal will have to retransmit the faulty packets. Now this could be one reason why they measure more usage on your transmission than what you see on your router: Because your router will measure the traffic on your LAN, between your router and the VSAT which is the gateway, but it will not be able to measure the traffic between the gateway (the terminal) and the hub. If there are lot of packets retransmissions on this path, the hub will notice the increased capacity and you wouldn't. What is required to find out is how the terminal measure the traffic volumes - is it on the interface toward your router or is it on the segment between the terminal and the hub?

Another issue that should be checked is the availability of the time slots on the inbound carrier/carriers. Without getting too deep, shared platforms like Skyedge, use shared space-segment (or shared bandwidth if you like) for the outbound traffic (traffic going from the hub to the remotes) and also shared bandwidth for the inbound traffic (from the remotes to the hub). Now if the service provider over-subscribe, your terminal will try to transmit but will not manage to hold of free time-slot, like when you try to use your mobile but because the network is congested you can not make the call. If the situation is really bad, your terminal might try again and again to get for free time slot and will not manage to. The result will be lot of overhead instead of traffic, but I am not sure the hub will measure this overhead as traffic.

I would strongly recommend you to ask Globaltt what is the SNR value of your transmission at their end, I think this could be a possible reason. But really I would be happy if someone from Globaltt would step in and give some more answers.

Regarding the other technical questions - again, without getting into too much details:
1. You can think of the satellite like a mirror in the sky, that reflects signals from one side to another. This is much more complicated than this, but to make things simpler, you can consider that the signal you transmit is received at the hub and there its SNR should be checked; not on the satellite itself.
2. Regarding IP traffic over satellite: This is maybe one of the most ivestigated issues; companies are trying to improve their solutions and technology to face the two main challenges of satellite communication, which is long delay and expensive bandwidth. There are many solutions and tricky and creative features here, basically one of the most important one is what is called Spoofing, which means that the terminal receive the packets, open them and transmit them over the satellite in a different manner, discarding the acknowledge packets for example and other overheads traffic.
Back to top
 

www.oasisnetworks.net - Oasis Networks - Online with you!
IP Logged
 
Pages: 1