Network errors in replication

Hi All,

We are receiving replication alerts with the following messages.

TCP Provider: An existing connection was forcibly closed by the remote host.

The process could not connect to Subscriber ''

TCP Provider: The semaphore timeout period has expired.

These set of errors comes daily at least once. I was doubting if it is network issue. To track the network we used Solar Wind tool and configured alert on ping request timing out and on packet loss.But the network alert and replication alert comes at different times.

I wanted to know is there any setting in replication due to which these errors arise.

In the Mege agent parameter listing, there is logintimeout and -querytimeout. Can some one point out the difference between two?

Thanks!