In the week where nearly 2 million people converged into the (relatively) tiny area of The Mall, in Washington DC, I couldn't help but think about the additional load put on the communication infrastructures in the area for that day. It brings the ideas of 'busy hour' and 'peak load' into sharp focus.

If you're a network manager who happens to have the Performance SPI recording and reporting upon your network then you just might want to go back to Tuesday the 20th of January and take a look at some of the usage patterns. You'll find that the SPI stores data in three distinct ways which are used for different needs.

When you launch the Report Menu you see a series of tabs, each tab represents a different reporting area. The first is 'Component Health', providing reports on things like CPU, Buffer and Memory from device sensors. The second is Interface Health, providing reports based primarily upon MIB-II data from the ifTable - utilization, discards and errors; the last is the Diagnostics tab which has reports on the iSPI for Performance system itself.

Each tab contains a series of thumbnails, representing the various reports you can access. The thumbnails have up to three words below them:

Live

Summary

Archive

These clickable links are variations of the same report, but each serves a different purpose. Live reports show you data for the last 6 hours, at the same granularity you collected it. Summary reports show you data which has been summarised up to the one hour grain, but they keep that data for 70 days. Archive reports are there purely for archival purposes. The data is stored at the most granular level and can be accessed for up to 70 days. However, 70 days can be a lot of data to store so the system defaults the retention period for these reports to 14 days. If you want to change it use the Configuration UI.

So when would you use each of these types of reports? Here are some scenarios:

1) I want to examine data as it is collected in real time, or investigate a problem that has just happened or continues to happen: Use a live report, you have all the detail available and it stays around for 6 hours. The live reports all auto-refresh, so consider running them on your wall board.

2) I want to look at traffic patterns on a switch, aggregated on a weekly basis, for the past two months: Use a summary report, you don't need the really fine grain data but you do need to see the patterns over a long period.

3) I want to investigate traffic patterns at 12.00 noon EST on the 20th of January: Use the archive reports, select the day you need from the list available and start digging into the data.

Until next time – for the Network Management Center Damian.

Join the other conversations

Operations - hp.com/go/ITOpsBlog

Business Service Management - hp.com/go/BSMblog