ProLAN: The fifth Level

Performance of applications eyes of users

• The fifth Level is a management system business for transactions and quality of work of business applications for "eyes of users".
• The fifth Level is an expansion of the systems of network management ProLAN and the systems of network management of other producers (Microsoft System Center, HP Open View, IBM Tivoli, etc.) an opportunity to control performance, availability and errors of business applications as they are "seen" by users.
• The fifth Level is the first (and still only) the product of category Quality of Experience Monitoring, Business Transaction Management supporting business applications of the Russian production.

Implementations of the Fifth Level allows:

#Повысить labor productivity of IT Service due to automation of process of diagnostics of failures in work of business applications. Diagnostics of failures can automatically be executed. MTTR (Mean Time To Repair) decreases. #Повысить loyalty of users of business applications due to quality improvement of operation of applications and simplification of registration procedure of incidents. Failures in work of IT Infrastructure and business applications can be registered and be eliminated to the appeal of users to Service Desk. MTBF (Mean Time Between Failures) increases. Incidents can automatically be registered. #Сократить costs of IT Service for the account:

• automatic control of observance of SLA by IT service providers. Reports on observance SLA (SLA Compliance Reports) are created automatically;
• automatic qualification of incidents. The IT Service automatically obtains information on that, CHTO occurred, GDE occurred and KOGDA occurred that allows to qualify automatically incidents and, thus, to reduce load of the first support line;
• existence of objective criterion of efficiency of the selected tactics of achievement of strategic objectives. Key criterion of compliance of IT to requirements of Business is APDEX (Application Performance Index). The reports supporting APDEX are created automatically.

Tools of the solution of these tasks are:

1. Проактивный monitoring of Quality of Experience (qualities of work of business applications on the party of users).
2. Автоматическое the notification about failures in work of business applications with detailing - THAT, WHERE, WHEN and also automatic obtaining information on the state of health of IT Infrastructure (servers, switches, routers, etc.) at the time of emergence of such failures.
3. Автоматическая registration in Service Desk of the incidents connected with the poor performance and errors in work of business applications. "From a box" products are supported: HP Service Manager, BMC Remedy.

Solution architecture

Figure 1. Solution architecture the Fifth Level

Quality of Experience (QoE) is a set of the metrics characterizing quality of work of business applications on user side ("the user's eyes") namely:

1. Время reactions of business applications, measured on user side (E2E RT, End-to-End Response Time). This time from the moment when the user, working in the business application, requests accomplishment of a certain function till the moment when this function is executed (an example of function, - search of goods in the reference book).
2. Количество the transactions executed by the user with separation into successfully executed transactions, compulsorily complete transactions on some event (for example, to an error, a timeout, etc.), the transactions completed successfully, but executed with violation of the standard sequence of actions.
3. Количество errors at accomplishment of transactions with separation on system and user. System mistakes are caused by failures in work of IT Infrastructure or errors in the code of applications. User – result of the wrong actions of users.
4. Значения APDEX (Application Performance Index). APDEX is the result of statistical processing of the measured E2E RT values characterizing satisfaction of users with performance of business applications.

Now the Fifth Level supports the GUI-applications working at the MS Windows platform. WEB, Java and console applications will be supported in the following releases.

Management of Quality of Experience includes:

• Measurement and monitoring of Quality of Experience (QoE).
• Correlation of QoE with health of IT Infrastructure.
• The notification about crucial events.

Measurement and monitoring of Quality of Experience (QoE)

For measurement of QoE the program (Windows service) is used: EPM Agent. The program is installed on computers of users (or a terminal server if users work in the terminal mode). That the EPM Agent could distinguish the transactions executed by users, they should be described previously in the corresponding Template. The template of transaction is a XML file in which key signs of transaction and its attributes are described (type, the business application, business process, etc.). Therefore, for measurement of QoE on computers of users EPM Agents and the Templates corresponding to crucial transactions of the used business applications should be set. Templates are developed by ProLAN company (or partners) at a stage of implementation of the Fifth Level.

The principle of work the EPM Agent consists in automatic tracking of all processes and the related events executed on the computer of the user. If process which is not described in the Template is executed, then the events corresponding to it are ignored. If process which is described in the Template is executed, then all related events are automatically registered and timed. It allows to measure in real time:

• number of the executed transactions of each type;
• runtime of each registered transaction;
• number of the system and user mistakes which occurred at accomplishment of the registered transactions;
• APDEX values for the corresponding business applications.

All results of measurements are automatically exported to the consolidated database located on the selected computer – the Probe. On the Probe the Expert System estimating results of measurements on a five-point scale is executed (it is good, admissible, requires attention, on the verge, badly). Results of measurements and expert evaluations are displayed on QoE Management console. Thus, the Probe and the Management console of QoE provide:

1. Отображение the current QoE values (results of measurements and expert evaluations) on the system of logically linked graphic cards of network.
2. Автоматическое the notification of IT Service about exceeding by QoE values of the set thresholds (E-Mail, the SMS, start of the external program and many other things).
3. Статистическую processing of QoE values, their retrospective analysis and creation of Expert Reports (SLA Compliance Reports).

Support of SNMP (integration with other management systems).

Important advantage of the EPM Agent is support of SNMP and private ProLAN MIB (Private Management Information Base) containing descriptions of the metrics measured by the EPM Agent.

Support of SNMP provides a possibility of use as QoE Management console as SLA-ON family products (ProLAN-Administrator, ProLAN-Analitik, ProLAN-Ekspert), and any other systems of network management supporting SNMP, in particular, of the product Observer Infrastructure of Network Instruments company which is also offered by ProLAN company.

Figure 2. Support of SNMP (integration with other management systems)

Correlation of QoE with health of IT Infrastructure

For management of satisfaction of users it is not enough to measure QoE only. It is not less important to be able to define what it depends on, for example, why the application works too slowly that is the reason of this or that error, etc. That to define it, QoE needs to compare with metrics of health of IT Infrastructure (to carry out component analysis).

For comparison of QoE to health of IT Infrastructure in network, along with EPM Agents, the monitor of health of IT Infrastructure on the basis of any product of the ProLAN SLA-ON family is installed (ProLAN-Administrator, ProLAN-Analitik, ProLAN-Ekspert). Information on health of IT Infrastructure measured by means of such product automatically is imported to the same database where the QoE-metrics measured by EPM Agents get. Moreover, metrics of health of IT Infrastructure and a QoE-metrics automatically "become attached" to the general time line that allows to define the root reasons of deterioration in QoE quickly. Within operational monitoring for the solution of this task the Cockpit card of the SLA-ON Operations application is used. Within the retrospective analysis - the TrendViewer.NET or Trend Analyst program.

The notification about crucial events on user side

If on the computer of the user there is a crucial event (for example, a system mistake), then it is necessary to inform service Service Desk on it at once. It can become by, at least, three methods:

1. Отправкой SNMP ladder EPM Agent.
2. Отправкой messages Probe (e-mail, sms, another)
3. Отправкой messages of HelpMe containing the comprehensive information about an event namely:

• Description of an event (for example, name of an error).
• The account of the user (including the domain).
• Computer name of the user.
• The name of division where the user works.
• The name of active process on the user's computer at the moment at the time of emergence of an event.
• Heading of an active window at the time of emergence of an event.
• Values of variables of the environment (environment string).
• Screenshot at the time of emergence of an event.
• The name of the transaction executed at the time of emergence of an event (the used IT service).
• Other information.

The message of HelpMe is sent by the application of the same name which is a part of the solution the Red Button which is transparent the Fifth Level is integrated with the solution.

The idea of integration consists in the following. If on the computer of the user there was a crucial event, for example, a system mistake, then the EPM Agent automatically identifies this error, "presses" the Red Button and reports to the HelpMe application the name of an error. Having obtained information on an error, the HelpMe application automatically "gets" all the rest (takes the screenshot, reads variable environments, defines the transaction executed at the time of an error, etc.) then automatically writes all acquired information in the consolidated database. Thus, the IT Service automatically obtains the comprehensive information about an error and will be able to qualify it to the appeal of the user to Service Desk. Also, if it is necessary, all information on an error can automatically be registered in Service Desk. "From a box" are supported: HP Service Manager and BMC Remedy.