Hello, we use the Java Price Engine Client API to receive the FXCM real-time price updates and have noticed our Java client applications, one instance running at our primary site and one running at are secondary site are not receiving the exact same number of update messages. I can see with netstat the client application is reconnecting with the price server every few minutes. Is the behavior expected or could this be causing me to lose real time updates?

Here is the output from netstat
Fri Jul 10 13:43:29 CDT 2009
tkr@ticker-commodity3-live:~> netstat -t | grep nyfiobridge
tcp 0 0 ticker-commodity3:39723 nyfiobridge.fx:www-http CLOSE_WAIT
tcp 0 0 ticker-commodity3:39724 nyfiobridge.fx:www-http CLOSE_WAIT
tcp 0 0 ticker-commodity3:39725 nyfiobridge.fx:www-http CLOSE_WAIT
tcp 0 0 ticker-commodity3:55062 nyfiobridge.fx:www-http CLOSE_WAIT
tcp 0 0 ticker-commodity3:55063 nyfiobridge.fx:www-http CLOSE_WAIT
tcp 0 0 ticker-commodity3:55061 nyfiobridge.fx:www-http CLOSE_WAIT
tcp 0 0 ticker-commodity3:55058 nyfiobridge.fx:www-http CLOSE_WAIT
tcp 0 0 ticker-commodity3:55059 nyfiobridge.fx:www-http CLOSE_WAIT
tcp 0 0 ticker-commodity3:55064 nyfiobridge.fx:www-http TIME_WAIT
tcp 0 0 ticker-commodity3:55065 nyfiobridge.fx:www-http ESTABLISHED
tkr@ticker-commodity3-live:~> date
Fri Jul 10 13:44:18 CDT 2009
tkr@ticker-commodity3-live:~> netstat -t | grep nyfiobridge
tcp 0 0 ticker-commodity3:39724 nyfiobridge.fx:www-http CLOSE_WAIT
tcp 0 0 ticker-commodity3:39725 nyfiobridge.fx:www-http CLOSE_WAIT
tcp 0 0 ticker-commodity3:55062 nyfiobridge.fx:www-http CLOSE_WAIT
tcp 0 0 ticker-commodity3:55063 nyfiobridge.fx:www-http CLOSE_WAIT
tcp 0 0 ticker-commodity3:55061 nyfiobridge.fx:www-http CLOSE_WAIT
tcp 0 0 ticker-commodity3:55058 nyfiobridge.fx:www-http CLOSE_WAIT
tcp 0 0 ticker-commodity3:55059 nyfiobridge.fx:www-http CLOSE_WAIT
tcp 0 0 ticker-commodity3:55066 nyfiobridge.fx:www-http CLOSE_WAIT
tcp 0 0 ticker-commodity3:55067 nyfiobridge.fx:www-http ESTABLISHED
tcp 0 0 ticker-commodity3:55065 nyfiobridge.fx:www-http CLOSE_WAIT
tkr@ticker-commodity3-live:~> date
Fri Jul 10 13:44:44 CDT 2009
tkr@ticker-commodity3-live:~> netstat -t | grep nyfiobridge
tcp 0 0 ticker-commodity3:39725 nyfiobridge.fx:www-http CLOSE_WAIT
tcp 0 0 ticker-commodity3:55062 nyfiobridge.fx:www-http CLOSE_WAIT
tcp 0 0 ticker-commodity3:55063 nyfiobridge.fx:www-http CLOSE_WAIT
tcp 0 0 ticker-commodity3:55061 nyfiobridge.fx:www-http CLOSE_WAIT
tcp 0 0 ticker-commodity3:55058 nyfiobridge.fx:www-http CLOSE_WAIT
tcp 0 0 ticker-commodity3:55059 nyfiobridge.fx:www-http CLOSE_WAIT
tcp 0 0 ticker-commodity3:55068 nyfiobridge.fx:www-http ESTABLISHED
tcp 0 0 ticker-commodity3:55066 nyfiobridge.fx:www-http CLOSE_WAIT
tcp 0 0 ticker-commodity3:55067 nyfiobridge.fx:www-http CLOSE_WAIT
tcp 0 0 ticker-commodity3:55065 nyfiobridge.fx:www-http CLOSE_WAIT

Its possible they are out of sync by some seconds but all things being equal they should receive the same information over time at the end.

For each symbol we track the number of updates and we can see that there is a difference between the primary site instance and our backup site instance. We put a sequence number on our messages so we know we are not dropping data on our side. So what it looks like is within the Java API when we make a request for data we noticed the port it connects to changes - indicating it reconnected to FXCM. Our question is - is this normal and expected behavior? Why would it reconnect and can we prevent it. We think while it is reconnecting we are losing data and that is why the counts are off.