We are implementing Call Pilot and have noticed that the server sends out KeepAliveTimer, or KAT, packets at about 5 packets per second, (including the clients response this means 10 PPS). From what I read on other groups the this is needed for the sql server to clean up stale sessions. This is a huge burden on the server, with 1000+ concurrent sessions, the network, and for remote clients running the IP softphone from home offices. I modified the registry setting for this parameter from
HKEY_LOCAL_MACHINE\\SYSTEM\\CurrentControlSet\\Services\\Tcpip\\Parameters\\KeepAliveTime=00000000 (hex) (0 seconds before transmitting the next KAT) to
HKEY_LOCAL_MACHINE\\SYSTEM\\CurrentControlSet\\Services\\Tcpip\\Parameters\\KeepAliveTime=000DBBA0 (hex) (15 Minutes before sending the next KAT )
and this seemed to clear it up.
Here's a sniffer trace from before-
30 1.406936 client host TCP [TCP Keep-Alive ACK] 1847 >imap [ACK] Seq=165 31 1.608089 host client IMAP [TCP Keep-Alive] Response: \\000 32 1.608105 client host TCP [TCP Keep-Alive ACK] 1847 >
imap [ACK] 33 1.809304 host client IMAP [TCP Keep-Alive] Response: \\000 34 1.809342 client host TCP [TCP Keep-Alive ACK] 1847 >
imap [ACK] 35 2.010529 host client IMAP [TCP Keep-Alive] Response: \\000 36 2.010554 client host TCP [TCP Keep-Alive ACK] 1847 >
imap [ACK] 37 2.211663 host client IMAP [TCP Keep-Alive] Response: \\000 38 2.211695 client host TCP [TCP Keep-Alive ACK] 1847 >
imap [ACK] 39 2.412831 host client IMAP [TCP Keep-Alive] Response: \\000 40 2.412868 client host TCP [TCP Keep-Alive ACK] 1847 >
imap [ACK]
Here is how it is now-
94 907.489311 host client IMAP [TCP Keep-Alive] Response: \\000 95 907.489341 client host TCP [TCP Keep-Alive ACK] 1417My questions are-
Is the best way to fix this issue?
Are there any problems doing this?
Thanks for your help!
mj