Quote:
Originally Posted by greenot  Is there a reason that push works fine for exchange, but not for Zimbra? I really have come to rely on my push functionality and this will probably have to go into the evaluation equation if I want to fully switch things over. |
I'd reexamine the possibility that when you switched from exchange to zimbra, you chose different push settings.
Here's Microsoft's page on Direct Push:
Understanding Direct Push: Exchange 2007 Help. Basically, Direct Push opens a new connection with Exchange every 15 minutes
and that connection must stay open in order for Exchange to be able to tell the device about anything that changes in the interim. Opening a new connection is a called a "ping" (not to be confused with icmp ping). Should anything new arrive in the 15 minute interval, Exchange prompts the client to download it, and then the client then issues a new ping.
What this means is that if a device isn't receiving any data, it should be issuing a new ping every 15 minutes. If it does receive data, it will issue new pings more frequently.
In fact you can observe this activity. Just do
grep \=Ping$ /opt/zimbra/log/sync.log > ping.log
to capture all recent pings into ping.log. Then if you want to see the activity for a particular user, do
grep <username> ping.log
When I did this I saw, not too surprisingly, long ping intervals in the middle of the night, and shorter ones during the day.
On top of all this, since ping is a "long-lived https request", I would surmise that it requires the device to stay on and maintain its IP connection (and therefore its wireless connection) continuously in order for it to be able to accept the response from the server.
In short I can't really see anything from the server side that would make battery life shorter with Zimbra than Exchange.
Bugzilla
Wiki
Source
iPhone Battery Life 
Linear Mode
