This will mostly be a function of the [avg latency/hop in WAN] and [# of
hops in WAN] and [max throughput of slowest link in WAN] and [efforts made
to tune packet size to avoid fragmentation - that is, adjusting packet MTU
size to match WAN limitations]. For the last factor, if you're using
Windows, or just for general information, see web pages that discuss DRTCP,
a free Windows utility that allows easy access to various Windows settings
in the Registry that control the TCP stack.

Here's the fun part: Latency of any given hop in WAN will vary depending on
load. If you have a exclusive link (like a point-to-point T1 or fiber link),
then you have a much better chance of getting that reliable stats on latency
than if you've got an IPSec VPN going over an Internet connection, because
over the Internet you have no way of predicting load, unless you have an SLA
with each entity between you and your destination.

If your link is over an IPSec (or other) VPN over the Internet, I'd suggest
multiple traceroutes over a period of time (days to weeks) to determine most
likely route, including number and latency of hops. There's a tool called
smokeping to automate and graph this, if you care to work with Open Source
tools.

Murugan, Palani said:
| Hi
|
| I need to configure a node for backup through WAN.
|
| And they want to know the throughput on the network
|
| The data on the node = 20GB
|
| Could any of you tell how to calculate the throughput on the network ?