A systematic approach to the USP Clear presentation of information and clarity of thought when drawing up a unique selling proposition is not everything. They may not give the expected effect. The USP must be formed in such a way as to influence not only the reasonable perception of the client

System unit It's no secret that among buyers of new computers there are not so many people who seriously think about why they are buying it? But in vain! Only by correctly outlining the circle of your real needs can you with sufficient accuracy

System Monitor After entering the command line Run perfmon.msc command, the Microsoft Management Console accesses the contents of the registry branches HKEY_CURRENT_USERSoftwareMicrosoftSystemMonitor and HKEY_LOCAL_MACHINESOFTWAREMicrosoftWindows NTCurrentVersionPerflib, the parameters of which will be discussed a little later. After entering

System block The system block is the information processing block. It is interesting to us because it is on it that the button, or key, for turning on the computer is located. It should be noted that there are also other quite important devices, such as: disk drive for

System Timer The system timer plays a more significant role in keeping track of the passage of time by the kernel. Regardless of the hardware platform, the idea behind the system timer is the same - it provides a mechanism for managing interrupts that

System unit This is usually a metal box, large or not very large. You may be surprised, but this is the most important component of a computer (in principle, this is a computer). Outside this box (on its front part) there is a very important thing - a power button

Monitor Externally it resembles a TV. Therefore, the dimensions of the monitor are also the same as regular TVs, - is determined by the length of the screen diagonal in inches - 15, 17, 19, 21 (by the way, one inch is equal to 2.54 cm). Just as a DVD player without a TV is just a box, so

Introduction to monitoring systems.

Let's consider monitoring tools: Console Performance and Network Monitor

Console Performance.

To launch the Productivity console, open the Administrative Tools folder in Control Panel and select the Performance.

The Productivity Console is actually two tools:

• System Monitor snap-in,
• Standalone Performance Logs and Alerts snap-in - to view monitoring log files.

Equipment System Monitor.

With the Performance Monitor snap-in, you can measure the performance of your computer or other computers on the network:

• Collect and view current system performance data on local computer or on multiple remote computers.
• View current data or data previously collected using the Performance Logs and Alerts snap-in.
• Present data in the form of a graph, histogram, or report that can be printed.
• Implement opportunities System Monitor V Microsoft Word or other applications Microsoft package Office using OLE automation.
• Create HTML pages to view performance.
• Create monitoring configurations that allow reuse, which can be installed on other computers using the MS Management Console

Objects and performance counters.

Windows XP gets performance information from computer components. System components generate performance data as they operate. These components are called performance objects.

The OS has a number of performance objects, usually corresponding to major hardware components such as memory, processors, etc. Applications can also install their own performance objects.

Each performance object provides counters that collect performance data. For example, the Pages Exchanged/sec counter of the Memory object tracks the extent to which pages are cached.

The objects that are most often used to track the operation of system components are:

• logical drive;
• physical disk;
• memory;
• flow;
• process;
• CPU;
• system;
• swap file;

In Windows XP the object Physical disk is enabled by default, and the object Logical drive disabled by default. To enable counters for monitoring logical drives or volumes you need to run the command in the console diskperf -yv and reboot the system.

To see an explanation of what data a particular counter provides, click the Explain button in the Add Counters dialog box, which you can access by clicking the plus icon on the snap-in toolbar.

Some objects (such as Memory and Server) have only one instance, although other performance objects can have multiple instances. If an object has multiple instances, you can add counters to track statistics for each instance or for all instances at the same time.

For example, if the system has several processors installed, then the object CPU will have many copies. Moreover, if an object supports multiple instances, then when the instances are combined into a group, there will be a parent instance and child instances that will belong to that parent instance.

Setting up counters.

When you select the System Monitor snap-in, a graphing window will open by default in the results pane. Steps to add counters:

If you do not have the appropriate permissions to monitor your computer, you will receive an error message. The counter will be indicated on the histogram, but no data will be displayed.

If the counter that you want to monitor is not on the screen, it is possible that the service or element that is the counter object is not installed or enabled on this computer, and you should add this service first.

You can place the System Monitor control in Microsoft document Word. To do this:

• Open Word document and place the cursor at the location where the control is supposed to be placed.
• From the View menu, select Toolbars -> Controls.
• Click the More Items button and select the System Monitor Control option from the list.
• The System Monitor graph will be inserted into specified place in the document. IN at the moment the element is in design mode so you can work with it in the editor Visual Basic.
• To change settings or add counters, click the Exit Design Mode button to exit Design Mode.
• To add counters to the chart, click on it right click mouse and select Add counters from the context menu.

Settings appearance.

The System Monitor snap-in provides three viewing tools for performance information: two graphical (Graph and Histogram) and one textual (Report). To customize the appearance of the monitoring window, right-click the graph and select Properties from the context menu. In the window that opens, you can set a series for the graph and histogram additional parameters display:

• name of the graph or histogram and coordinate axes;
• value output range;
• characteristics of a curve on a graph or bars on a histogram, such as color, thickness, style, etc.

On the tab General You can specify the required type of monitoring tool. By default the option is selected Schedule. You can also display performance data as a bar graph or report.

Working with the System Monitor snap-in.

When monitoring your system, keep a few things in mind: useful rules, which will allow you to use system resources most efficiently:

• Determine the configuration of monitoring tools. To track work Windows installations XP, you can view data as a graph using Performance Monitor, or collect data in performance logs for viewing and analysis in other applications. Configure the Alerts and Performance Logs snap-in to collect data from selected counters at specified intervals. The resulting data logs can be used to generate reports and analyze overall system performance, as well as plan future upgrades.
• Keep the resources required for monitoring low. Monitoring tools configured for consumption minimum quantity resources. However, in some cases it is necessary to take additional measures to reduce it. Presentation of performance data in the System Monitor window in the form of a graph, frequent data sampling, large number monitored objects and counters - all of which increase the amount of resources spent on performance monitoring.
• Analyze performance data and establish a performance baseline. It is generally useful to establish a performance baseline for a typical workload. This can be done by outputting the data to graphical form in the System Monitor window.
• Setting up alerts. Set to generate alerts when counter values ​​exceed acceptable values.
• Performance tuning. Use performance data to tune system settings for optimal handling of system load.
• Planning. Monitor trends in server load changes and the need to upgrade the system hardware.

Choosing a monitoring method.

For current monitoring of system operation in real time on a local or remote computer, it is convenient to use graphs. Snap-in logs Performance logs and alerts are useful for recording entries. The logged data can be used to generate reports and presented as graphs or histograms using System Monitor.

Selecting the recording frequency.

Care should be taken when choosing the frequency and duration of data logging because frequent updates data leads to the generation of a very large amount of data that will be difficult to work with. This may also result in increased performance overhead for running the Performance Logs and Alerts snap-in.

IN general case The frequency of data updates depends on the length of the logging interval. If the registration interval is 4 hours, then it is recommended to install the update every 15 seconds. For an 8-hour logging interval, select a refresh period of at least 300 seconds (5 minutes). In general, for continuous monitoring, the interval is set to 15 minutes.

Selection of counters.

Monitoring should begin by tracking the following four components in this order:

• Memory.
• Processors.
• Disks.
• Net.

The table below shows minimum set counters that should be used to monitor the server. As you work, you can add additional counters for the performance objects you are interested in.

Degree of use

Memory\Bytes available
Memory\Byte Cache

Bottlenecks

Memory\Pages/sec
Memory\Read Pages/sec
Memory\Transit errors/sec
Memory\Bytes in Paged Pool
Memory\Bytes in the nonpaged page pool
Counters are also useful
Swap File\% Usage
Cache\% hits when displaying data

CPU

Degree of use
Bottlenecks

Processor\% CPU load (all occurrences)
Processor\Interrupts/sec
System\CPU Queue Length (all occurrences)
System\Context switches/sec

Degree of use
Bottlenecks

Physical disk\Disk read accesses/sec
Physical disk\Disk write accesses/sec
Physical Disk\Average Disk Queue Length (all occurrences)

Network interface

Degree of use
Performance

Network segment\% network usage
Protocol transfer counters
Network interface\Total bytes/sec
Network interface\Packets/sec
Server\Total bytes/sec
or
Server\Bytes sent/sec
And
Server\Bytes received/sec)

Selecting the computer that will be used for monitoring.

When monitoring remote computers, several options for collecting data are possible. For example, you can run Performance Logs and Alerts on an administrative computer and display data from all remote computers. Or you can start a data collection service on each computer and run it at regular intervals batch program to transfer data to the administrator’s computer for subsequent analysis and archiving.

Centralized data collection (collecting data from all remote computers on the local computer) is the easiest to implement. Collecting data from multiple systems can be recorded in a single log file, but this increases network traffic and requires more memory on the administrator's computer.

Distributed data collection (that is, data collection conducted on each monitored computer) does not increase traffic and does not require additional memory on the administrator's computer. However, the transfer of data to the administrator's computer will occur with a certain delay.

Performance data analysis.

Analysis of monitoring results includes checking meter readings recorded during system execution various operations. During this process, you should identify the most active processes and also identify programs or threads that are exclusively using any resources. As a result, you must figure out how your system handles the workload.

As part of this analysis, you should determine the level of system performance when a typical workload is being processed and all necessary services which is called basic level. The base standard is determined by the administrator based on the workload. This level is determined based on a number of performance counters and represents the state of the system when it satisfactorily handles all user requests.

Determining acceptable meter readings.

In general, determining an acceptable level of performance is a rather subjective decision. However, the table below provides limit values ​​for a number of meters to help you determine if your meter readings indicate a problem. If the meter readings remain consistently at the maximum level, then there is every reason to talk about the presence of a bottleneck in the system, and measures should be taken to configure or upgrade the loaded resource.

Limit meter readings:

@ECHO OFF ECHO Adding counter "New_counter"... rem Create new group data collectors logman create counter New_counter -f bincirc -c ^ "\Processor(_Total)\%% Processor Time" ^ "\Memory(_Total)\Page/sec" ^ "\Memory(_Total)\%% Uncommited Bytes In Use" ^ "\Memory(_Total)\Available Bytes" ^ "\System(_Total)\Processor Queue Length" ^ "\PhysicalDisk(_Total)\Avg. Disk Queue Lenghth" ^ "\PhysicalDisk(*)\Avg. Disk Queue Lenghth" ^ "\Network Interface(*)\Bytes Total/sec" ^ -si 15 -v ​​mmddhhmm rem Launch the performance monitor C:\Windows\System32\Perfmon.exe ECHO done

Counter names are enclosed in quotation marks and separated by a space. The names of the counters can be written both in English and in Russian (but then the bat file will only work in Russian-language Windows versions). If the meter name contains the sign % , it must be replaced by signs %% (double percentage).

Having completed this file, we will immediately be taken to the performance monitor, where we will see a newly created data collector group named New_counter.

The counters described are usually sufficient for overall assessment system load. Other counters are used to investigate problems in detail if it is determined that there is a problem lack of productivity equipment.

Did this article help you?