We recently had Time Warner nLAN service installed. It hands an Ethernet connection off to us from equipment the installed at our two sites which bonds, using multilink PPP, two clear channel t1s. We have the same setup at both our locations creating our WAN. We have a server that runs the Windent (dental application based on MS SQLdatabase) application at one locaition, with clients and just clients at the other location accessing the Windent application over the WAN. Winednt uses a client app that is run on the local computer.
Since the installation we have not seen any performance imprpovement, it is as slow as it was before accross the WAN. The clients located in the same office as the server have no performance issue. Wondering what may cause this. I know that is a lot to ask given all the variables.
Testing and utilization checks have shown no signs of congestion. Standard ping packets show 11ms latency, ping packets w/ 1500 payload show 31ms latency. No drops.
We ran voice testing app using UDP traffic (so we could measure loss) that utilizes an agent on our windent server and an agent on a PC at our remote, set the test to use 2.7Mbps of the bandwidht. Experienced no drops, 11ms delay (20ms payload) jitter less than .1ms.
When I do a windows drag and drop copy from the server or a pc at the main location to a pc at the remote, I only get 2Mbps bandwidht utlization when this is the only traffic on the pipe/s. If I drag and drop copy from two PCs at each end, 2.7Mbps will get used.
Would like to know if there is a reason, perhaps SQL or other that our Windent application performance is not improved and why a Windows drag and drop copy only leverages 2 of the 3 (minus overhead) Mbps connection we have?