As the system administrator for a VOS system you typically could care less about what other hosts are on the local subnet.
A multihomed system is a system with multiple IP interfaces. These interfaces can be on the same or on different subnets. Today I want to consider protocols like FTP and NDMP.
First a little background, TCP has a concept called maximum segment size (MSS). This is the largest segment (of data) that the TCP stack will accept.
Having two interfaces in a module on the same subnet does not give you two interfaces into and out of the module.
Paul covered CPU utilization and code paths but there is another very important aspect of many applications – network utilization.
I recently was asked to do some analysis of a trace file, it was only 161 seconds in duration and contained a little more than 2.6 million frames. I was amazed to discover that approximately 75% of those frames where either ping requests or ping replies.
Many times when working an issue I ask people to send me a network trace. A network trace often points to the root cause of the problem. At a minimum, it reduces the problem space to something that is manageable.
I recently ran into a site with several modules, each of which was forwarding packets. The worst case was at a rate of 1 every 2 seconds or so.
It is coming up on the end of daylight savings time and that means questions; specifically how to change the time zone, does ntpd need to be stopped and restarted when the time zone changes and if ntpd is running will it change the time zone automagically.
The impact of communication layer latency is typically under estimated when trying to fix application performance problems but correct understanding is critical if you are to direct your efforts toward practical solutions.