| Author |
 Topic  |
|
|
Chico
Seasoned But Casual Onlooker
Canada
68 Posts
Status: offline |
 Posted - 10/11/2012 : 1:51:08 PM
|
Hi Everyone
Hope this is the right section, if not please move it to right location.
Looking for some feed back here.
I'm currently working at a location in Ontario where this office connects to a server in another location (Quebec) to use a ERP system. The biggest complaint I get is the delay time from when you click on an icon and how long it takes for it to pop up so you can enter data in the field, or may get the hour glass and look at the top left section of the screen and get the message (not responding).
I have checked the internet speed at this location (with Rogers DSL) and it has a 10meg d/l and a 2 meg u/l. I'm convinced that the u/l speed is causing the problem too much buffering here.
The other location where the server that we connect to is running a Mxdsl with speeds of 10 Megabits download / 3 Megabits upload.
They have switched us over to a dictated server (only us are using it) and we still haven't seen any improvement.
So my question I have and looking for feedback and/or suggestions here:
• Do I look at bringing in a second (DSL) line and bridge it or. If I bridge what equipment is needed for bridging to work effectively.
• Look at bringing in a Fibre Optic line either getting a 5/5 or 10/10 dl & ul speed.
• If I improve at this location will the other location need to improve their internet connection speed too.
• Is there something that can be done between the two servers, (both are running Windows 2008 RC2
If I do either one, would this improve things (i.e.: improve the time it takes for the window to pop up,
or will this be a waste of time and money.
Let me know, please.
|
Eric J. O’Brien
|
|
|
wkasdo
Administrator
Netherlands
7424 Posts
Status: offline |
Posted - 10/11/2012 : 2:49:39 PM
|
The problem is probably not in the line speed, but in the latency. That's just a guess of course, but traditional client/server apps are sensitive to this because they do many roundtrips. A network trace would quickly verify this.
If this is really the problem, you need to consider a different solution. RDP/Citrix comes to mind. Solving it from the network, you'd be looking at a WAN accellerator. |
Make it as simple as you can, but not simpler -- Albert Einstein |
 |
|
|
Chico
Seasoned But Casual Onlooker
Canada
68 Posts
Status: offline |
Posted - 10/11/2012 : 3:01:03 PM
|
Can you define (network trace). Is this software or hardware that is used to do this trace.
What does a WAN accellerator do.
Thanks, |
Eric J. O’Brien
|
|
|
|
wobble_wobble
Honorable But Hopeless Addict
Ireland
4523 Posts
Status: offline |
|
|
Chico
Seasoned But Casual Onlooker
Canada
68 Posts
Status: offline |
Posted - 10/12/2012 : 08:48:11 AM
|
| Can wireshark used for a LAN or can it be used in a WAN setup. |
Eric J. O’Brien
|
|
|
|
wobble_wobble
Honorable But Hopeless Addict
Ireland
4523 Posts
Status: offline |
Posted - 10/12/2012 : 09:02:49 AM
|
Both, its a network protocol analyzer and will work on both.
Netmon also works very well, but for Latency issues, I find Wireshark easier.
|
Joe
After everything that has happened during the month of Jan 07, I do believe that pigs fly backwards!
http://whatismyv6.com/ |
|
|
|
Chico
Seasoned But Casual Onlooker
Canada
68 Posts
Status: offline |
Posted - 10/12/2012 : 09:09:37 AM
|
OK I've done a ping test, can someone tell me if these times are good or bad.
Pinging srv-rds.machinex.ca [10.1.2.40] with 32 bytes of data:
Reply from 10.1.2.40: bytes=32 time=34ms TTL=128
Reply from 10.1.2.40: bytes=32 time=34ms TTL=128
Reply from 10.1.2.40: bytes=32 time=43ms TTL=128
Reply from 10.1.2.40: bytes=32 time=34ms TTL=128
Ping statistics for 10.1.2.40:
Packets: Sent = 4, Received = 4, Lost = 0 (0% loss),
Approximate round trip times in milli-seconds:
Minimum = 34ms, Maximum = 43ms, Average = 36ms |
Eric J. O’Brien
|
|
|
|
wobble_wobble
Honorable But Hopeless Addict
Ireland
4523 Posts
Status: offline |
Posted - 10/12/2012 : 09:16:38 AM
|
Try this
ping srv-rds.machinex.ca -l 1000 -n 100
And give us the report bit.
|
Joe
After everything that has happened during the month of Jan 07, I do believe that pigs fly backwards!
http://whatismyv6.com/ |
|
|
|
Chico
Seasoned But Casual Onlooker
Canada
68 Posts
Status: offline |
Posted - 10/12/2012 : 1:23:27 PM
|
Here it is.
Pinging srv-rds.machinex.ca [10.1.2.40] with 1000 bytes of data:
Reply from 10.1.2.40: bytes=1000 time=48ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=53ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=50ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=50ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=49ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=58ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=59ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=54ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=52ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=48ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=53ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=54ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=49ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=47ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=51ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=59ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=54ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=49ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=54ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=49ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=50ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=48ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=50ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=49ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=49ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=48ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=52ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=49ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=48ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=49ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=49ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=48ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=54ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=49ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=48ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=47ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=49ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=48ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=55ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=51ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=48ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=48ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=48ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=48ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=49ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=48ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=47ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=48ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=47ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=52ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=49ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=50ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=53ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=49ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=49ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=49ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=50ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=48ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=49ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=49ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=48ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=48ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=49ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=49ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=49ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=50ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=48ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=52ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=49ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=48ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=49ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=49ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=49ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=48ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=52ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=49ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=47ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=49ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=47ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=49ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=49ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=48ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=48ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=49ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=49ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=48ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=48ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=51ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=48ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=49ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=49ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=47ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=71ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=50ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=48ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=50ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=48ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=51ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=47ms TTL=128
Reply from 10.1.2.40: bytes=1000 time=52ms TTL=128
Ping statistics for 10.1.2.40:
Packets: Sent = 100, Received = 100, Lost = 0 (0% loss),
Approximate round trip times in milli-seconds:
Minimum = 47ms, Maximum = 71ms, Average = 49ms |
Eric J. O’Brien
|
|
|
|
wobble_wobble
Honorable But Hopeless Addict
Ireland
4523 Posts
Status: offline |
Posted - 10/12/2012 : 1:33:33 PM
|
Thats a latent connection.
TS Server/ Citrix solution (or get a browser addon, if available) for the application.
Your probably getting a lot of time outs. |
Joe
After everything that has happened during the month of Jan 07, I do believe that pigs fly backwards!
http://whatismyv6.com/ |
|
|
|
Chico
Seasoned But Casual Onlooker
Canada
68 Posts
Status: offline |
Posted - 10/12/2012 : 2:26:22 PM
|
| You lost me here. Can you explain more here. |
Eric J. O’Brien
|
|
|
|
wobble_wobble
Honorable But Hopeless Addict
Ireland
4523 Posts
Status: offline |
Posted - 10/12/2012 : 4:02:43 PM
|
The reason for the delay your users are seeing is that the messages back and forward to the ERP system. Depending on how well written the ERP system is will dictate how much traffic goes back and forward. Not adjust that refresh/ response by adding 50 ms to each trip. So if the software is efficient not too much traffic....
If badly written, every screen change ( you add a symbol or character it number to a field) causes a screen redraw.
To alleviate this you really need either web browser based ERP system, think Google Docs Spreadsheet, Microsoft Live Docs.
Or you need to get a server based computing solution, like Microsoft Remote Desktop Services or if you have more than say 50 remote users a Citrix solution.
There are other options, but they are much more complex and in some cases more expensive, like a VDI solution or the new VMware Streaming solutions.
edit - grammer and missing words. |
Joe
After everything that has happened during the month of Jan 07, I do believe that pigs fly backwards!
http://whatismyv6.com/ |
Edited by - wobble_wobble on 10/12/2012 4:10:39 PM |
|
|
|
Chico
Seasoned But Casual Onlooker
Canada
68 Posts
Status: offline |
Posted - 10/15/2012 : 11:35:52 AM
|
| So in your option going to a fiber optic line with 10/10 speed we would not see any imporvementthen, what about a WAN accellerator as mentioned above. |
Eric J. O’Brien
|
|
|
|
Curt
Moderator
USA
6659 Posts
Status: offline |
Posted - 10/15/2012 : 11:58:19 AM
|
Most ERP systems that allow a full CAL for an end user use a "fat" client. As in a client server model and hopefully the ERP program is not on the database server that supports it.
That being said there are most likely RPC calls accross the line and it sounds like things are not set efficiently.
This is kind of the reason why application publishing still exist given that even a browser based app can be a "Fat" client today when it comes to a business application.
For a long time I've used RDS and Citrix to get the functionality out to areas of low bandwidth. We are still doing that today with a client I have using Great Plains. It would be a night mare of cost to provide a fibre connection to a "Harvesting area and cooler" that's mobile so publishing the application to the desktop allows for less data being moved to the remote client.
|
Curt Spanburgh
Microsoft Certified Business Solution Specialist.
Dynamics CRM MVP
Contributing Editor, Windows IT Pro
He that is walking with wise persons will become wise, but he that is having dealings with the stupid ones will fare badly.
Proverbs 13:20
|
|
|
|
Chico
Seasoned But Casual Onlooker
Canada
68 Posts
Status: offline |
Posted - 10/17/2012 : 08:14:27 AM
|
Well this is what is being used here is an RDS set up but the lag time is painfull sometimes.
So I'm kind of getting the impression that either adding a second line and/or going to a fibre optic 10/10MB line really won't do much then with regards to response time. |
Eric J. O’Brien
|
|
|
|
wobble_wobble
Honorable But Hopeless Addict
Ireland
4523 Posts
Status: offline |
Posted - 10/17/2012 : 08:51:59 AM
|
Eric,
If you are already using RDS, then a fibre connection will improve the latency significantly.
These are estimates from an Irish point of view, but the latency should drop from 40mS to anywhere less than 10 mS, but I would expect 4mS.
Your mileage will vary, depending on what you buy, your provider connection type etc.
|
Joe
After everything that has happened during the month of Jan 07, I do believe that pigs fly backwards!
http://whatismyv6.com/ |
|
|
|
Curt
Moderator
USA
6659 Posts
Status: offline |
Posted - 10/17/2012 : 10:05:15 AM
|
Yes, it your using RDS and there is that much latency then there truly is a problem.
Although, it would be interesting to look into a IP trace and see what is happening.
|
Curt Spanburgh
Microsoft Certified Business Solution Specialist.
Dynamics CRM MVP
Contributing Editor, Windows IT Pro
He that is walking with wise persons will become wise, but he that is having dealings with the stupid ones will fare badly.
Proverbs 13:20
|
|
|
|
Chico
Seasoned But Casual Onlooker
Canada
68 Posts
Status: offline |
Posted - 11/16/2012 : 3:02:18 PM
|
Hi everyone, thanks for everyone's feedback on this issue. It has forced our corp location to do testing using Citrix solution using vritual (Win 7) desktop setup. I have seen improvements in speed, but won't see the real improvement until they install a WAN accellerator. This will happen next year. As for the fibre optic line we are holding off until we complete the testing and go live with this, but I think we are on the right track.
Once again thanks. |
|
|
| |
Topic |
|
VPN between 2 locations
