Interstate branches connect to headoffice via VPN

By lee ·
Our client has their head office located here in Melbourne and a branche office located in every other state. All the state branches are connected to head office via a permanent VPN tunnel. Head office and all states are using WatchGuard VPN device's and ADSL2+ Telstra Broadband Business Class connections.Head Office has 2 servers in use.

1.Small Business Server 2003 R2 which is the domain controller with Exchange 2003.

2. Server 2003 R2 Terminal Server which provides access to Freight Logistics software sourced from a company Jaix Logistic's.

Everything works well both at head office and state branches except for one which is the Sydney office. When they connect to the Terminal Server, login and then start the Freight software all is good until they go to print and change within the program from Connote to Manifest. The issue is time delay/lag or very slow response times in these particular area's of the software.
It can take anywhere bewteen 2 to 4 minutes for the print job or the change between ConNote and Manifest. Once in Manifest everything is fine once your are in but the printing is very slow.The strange thing is though is that before the ISP change over from iinet to telstra the slow reponse time was reversed, meaning the ConNote segment experienced the lag and not the Manifest component. None of the other branches are having any problems at all when using Jaix, all areas are very responsive as well as the printing which takes about 10 sec to print. All printers, workstations, operating systems, office versions, anti-virus solutions and domain configurations are the same. Nothing is different at any branche office. Head Office uses a WatchGuard XTM 23 while the branches use a WatchGuard 22. I have triple checked everything for consistency in all areas, made sure that both servers have double the minimum requirements in regards to specifications etc. So I am mystified as to why it is only one branch that is having the issue. If anyone can provide any possible advise/reasoning or has experienced the same issue, I look forward to hearing from you

This conversation is currently closed to new comments.

Thread display: Collapse - | Expand +

All Answers

Collapse -

Judicious packet tracing.

by seanferd In reply to Interstate branches conne ...

Use Wireshark or your preferred capture and analyzer apps. Or, use a system or application profiler like Process Monitor. See what has error or wait conditions.

I'm sure someone with some narrower ideas on the problem will be around. Check back.

Collapse -

Thanks for the tip

by lee In reply to Judicious packet tracing.

Thanks for the tip, in the process now doing just that. Hopefully it can help me pinpoint where the issue may be. I like Wireshark especialy with the other component CACE Pilot, but damn they know how to charge fo it! Shall let you know how it goes.

Collapse -

Jitter, lag, packet sizes, buffer size...

by TobiF In reply to Interstate branches conne ...

Since you say you've checked that all locations are equal (and you even indicate that different ISP seems to give different results), I take it we need to deal with the networking.

1. Check that you have the correct MTU set on both ends, sot that you're not suffering from split packets.
2. Measure delay, jitter and lost packets.
3. Try expanding tcp buffer size, so that transfer of larger files can become more efficient. (But, if your ISP serves you many lost packets, then large buffer is contraproductive!)

But, with such delays in the network, I might start thinking if the responsibilities between client and server should be drawn in a different way. (But that's more about internal software design of your application, which you can't control that much, except shopping around and trying solutions from different vendors or by talking nicely to your currently selected software vendor.)

Collapse -

Close but no cigar!

by lee In reply to Jitter, lag, packet sizes ...

Hi and thanks for taking the time to read my post and reply. As you have sugested I have checked the MTU settings a both ends and all is well. All of the state branches are using the same ISP with exactly the same plan. All sites are achieving consistant upload/download speeds of 1mb up and 12mb down. Only one system/user accesses the terminal server in Melbourne from each satelite office. At most 5 are connected at anyone time. All states have no issues but the Sydney office. I don't think the client would change the software as they paid about $50000.00 for it. No one else experiences the lag issue when using the Jaix program. I am about to get them to login using a completely new TEST user account to see if the issue still occurs. If the lag in the Jaix software does not occur then it is a user account issue which is easy to fix, but if it does happen, then we are back to the drawing board. I am running packet capture on both the servers and at the client end on the workstation in question for the entire work day, so hopefully I will have some further insight when I analyze the capture logs tonight. Shall post what I find here later.

Related Discussions

Related Forums