Figure 142. Differences in Disk Subsystems
Improving performance of the disk subsystem will usually prolong the maximum
transactions per second rate shown in Figure 142. The effect of a faster disk can
be to improve the disk cache hit rate. This is because write-data can be flushed
from the cache quicker, so that the amount of memory available for caching of
new data is increased.
5.2 Performance Analysis Tools
Using performance analysis tools is an effective method of understanding server
performance.
5.2.1 DatagLANce
IBM DatagLANce is one such tool that can give valuable insight into server
performance. DatagLANce operates at the physical layer and monitors the
actual frames being transmitted on the network.
You can use the DatagLANce analyzer to get an accurate picture of the current
activity on your network or a historical record of network activity over a specified
period of time. You can design your own screens and save them from the menu.
If you wish, you can create 32 different bar charts of real-time statistics, launch
an analysis session and call up frame summary views, protocol interpreted
views, and hexadecimal views (all color coded, highlighted, and tracked
simultaneously). You can rearrange the statistics, add to them, and save them
under a new name. You can use the DatagLANce analyzer′s alarms to let you
know when certain statistical thresholds, such as network utilization, are
reached. Since DatagLANce runs on top of OS/2, you can even use your
machine for other applications while the analyzer is monitoring your network.
The following is a summary of DatagLANce features:
•88 source and destination address pairs filtered in real time
•Real-time frame capture while monitoring
•Eight fast, super-powerful, programmable event detectors
172 NetWare Integration Guide