You might want to look into BWMng for real-time monitoring, but I can live with SNMP-based graphing and reporting through munin or cricket.
On a side note, I understand that ifstat itself can be compiled for Solaris.
Oh, for throughout, detailed, protocol-level statistic dissections, ntop can't be beaten, but it requires a lot of resources and is overkill for most cases.
One of the nicest tools that I've found is nicstat. It helped me to quickly rule-out the network when we were trying to diagnose what we believed to be an NFS problem related to the network.
You might want to look into BWMng for real-time monitoring, but I can live with SNMP-based graphing and reporting through munin or cricket.
On a side note, I understand that ifstat itself can be compiled for Solaris.
Oh, for throughout, detailed, protocol-level statistic dissections, ntop can't be beaten, but it requires a lot of resources and is overkill for most cases.
You can look at the "raw" kernel stats using kstat:
kstat -p -m ce (or whatever you interface is e.g. bg)
One of the nicest tools that I've found is nicstat. It helped me to quickly rule-out the network when we were trying to diagnose what we believed to be an NFS problem related to the network.