Windows is a self tuning operating system. It’s kernel is already optimized to work in a variety of scenarios. There are only limited operations that can be performed to optimize the system.

The benefits associated with such a system include:

• Server Administration becomes easy.
• There are less chances of damaging the system by fiddling with the kernel level.

Performance Analysis is more of an art than a science. This article mainly focuses on familiarizing the readers with Performance Console also known as System Monitor Tool, by explaining some technical terms that are associated with it. Towards the end of the article, four major Subsystems that affect the performance of the system are also explained.

The basic technical terms that you must be familiar with in order to use Performance Monitor are as follows:

On a Production system, bottlenecks take a common place. It is in order to identify the system bottlenecks, we use System Monitor. The System Monitor tool helps to detect the system bottlenecks by capturing and analyzing the performance data on a local machine or any computer in the network.

To start System Monitor:

Start -> Programs -> Administrative tools -> Performance OR Start -> Run -> Type perfmon /low

By default, there are certain performance monitor counters. You can add new counters by clicking on the “Add” button ( a “+” sign) on the System Monitor tool bar. When this button is clicked, the “Add Counter” dialog box appears.

The System Monitor tool provides you with three different display formats to view the performance data.

• Graph
• Histogram
• Report

You can choose the display format by using the three buttons on the tool bar of the System Monitor. The Report view is normally preferred.

There are mainly four Subsystems that affects the performance of a system.

1. Memory
2. Processor
3. Disk
4. Network

Each of them is represented by a Performance Object. Windows provides a large number of Performance Objects and associated counters that you can analyze.

Memory Subsystem

When a program references to a data, the operating system first checks for it in the physical memory. If it is not present in the physical memory, the next place that the OS checks for it is the logical memory/page file. In case the data is not found here, the operating system will have to retrieve it from the disk.

Memory – Available bytes : Indicates the amount of physical memory that is available for the running process. In short, it indicates the free memory that is available on your system. If this value is less than 4 MB, you should consider adding more memory.

Memory – Pages/sec : Indicates the number of times that the required information was not found in the memory and had to be retrieved from the disk. The value of this counter should be less than 20 on an average.

Page File – % usage : This counter is used to determine how much of the allocated page is in use. If you notice that the number is consistently over 99%, then you may need to add more memory.

Page File – Usage peak : It records the maximum usage of page file in your system.

Processor Subsystem

Though, we have to monitor this Subsystem to make sure that processor utilization is normal, processors are not an important source of system bottleneck according to many senior system administrators.

Processor time : It indicates the amount of time that the processor spends to respond to the system requirements. If the value is consistently over 85%, it may indicate a processor bottleneck.

Interrupts/sec : The value of this counter indicates the number of hardware interrupts the processor receives each second. In a Pentium class computer, if the value of this counter is more than 3500, it indicates that the server could be having a problem with the program/hardware that is generating the interrupt.

The System Monitor tool helps to detect the system bottlenecks by capturing and analyzing the performance data on a local machine or any computer in the network.

Disk Subsystem

Disk access time is the total time that is required to access your disk, to retrieve the data that is required by the OS. The two factors that determine disk access time are:

• Average disk access time on the hard drive
• Speed of the disk controller

Physical Disk – Disk time : Indicates the amount of time that the physical disk is busy in serving read/write requests. If the value is greater than 90%, then you can be sure of a bottle neck in the disk sub system.

Physical Disk – Current disk queue length : Indicates the number of pending disk requests, that are waiting to be processed. The value of this counter should be less than 2, on an average.

Physical disk – disk reads/sec, writes/sec : We can not specify an exact value for this. It varies according to the manufacturer.

Physical disk – Free space : As the name suggests, it indicates the amount of free space that is available on the hardware. The value should always be greater than 15.

Network Subsystem

System Monitor provides many counters for monitoring the network interface and network protocols installed on the computer.

Network interface – Bytes Total/sec : The value of this counter indicates the total number of bytes that are sent or received from a particular network interface. (It includes all the network protocols that are bound to the network)

TCPv4 – Segments/Sec : It indicates the total number of bytes that are sent or received from a particular network interface. (It includes only the TCP protocol)

Conclusion

This article mainly focuses on familiarizing the reader with the System Monitor tool. Only the major performance objects and a few counters associated with them have been explained here. There are still a large number of performance objects and associated counters that are not explained. I advise all readers to go through every one of them in the Help section that is provided within the System Monitor tool itself. Use this tool to monitor the utilization of your system resources and view the real-time performance data in the form of counters for your server resources.

About the Author:
Hari Vishnu, Software Engineer has been working with Bobcares for more than an year now. He has expertise in both Windows and Linux server Administration, and he is considered to be a master when it comes to Windows servers. Apart from the technical side, he has gained a reputation as a gifted stage performer too.