1. REAL-TIME PERFORMANCE MANAGEMENT

Jennifer’s intuitive monitoring screen combines different types of information into one easily accessible display. When Jennifer identifies a performance issue, featured automatic alert system is triggered to notify the system administrator. Jennifer is also capable of tracking and tracing communication between WAS and the back-end system (CICS, TUXEDO), and interpreting DBMS SQL query history to identify exact location and cause of the issue encountered down to the exact class/method.

- REAL-TIME MONITORING

• Real-time monitoring of concurrent users
• User visits per day, per hour
• List of currently active services
• Active JDBC connection assignments and status
• Response time graph and profiling per each class/method
• Real-time throughputs and load monitoring
• JDBC/SQL monitoring with SQL BIND variables
• CTG, WTC, Jolt, J-Link, TUXEDO, CICS monitoring
• Requested count and response times of applications, CPU usages
• Un-closed JDBC resources (Connection/Statement/ResultSet)
• System CPU, memory, JVM heap usage
• HTTP session dump with detailed size of the element objects
• Opened TCP/IP socket status within applications
• Opened file status within applications for reading or writing
• Live objects counting and dump support
• Collection objects tracing for memory leakage analysis
• Stand-alone EJB server monitoring
• Loaded class/interface dependency identification
• Real-time code disassembly of loaded classes into Java byte codes
• HTTP POST key/values real-time tracing

- ALGORITHM OF CONCURRENT USERS

Due to HTTP protocol “connection-less” characteristic, quantifying actual real-time concurrent user counts was not possible until now. Jennifer boasts world’s first software solution that utilizes the “Performance Theory” to monitor, define, and measure concurrent users.

- ACTIVE SERVICES MONITORING

When performance issues arise, number of applications or active services count increases significantly. Jennifer displays these changes in active services count and service processing time as well as individual thread status in multi-colored graphs and charts, down to the originating class/method/SQL unit in real-time basis.

- RESPONSE TIME DISTRIBUTION GRAPH

Each application response time is displayed in this graph as a point, and total response time is also displayed in the summary table. Analysis of the trend in these points is crucial. If response time points are lined-up or concentrated in a specific area, this can point to new statistical perspectives that were previously unknown. Furthermore, if the points are selected and dragged into the data table, user can track the response time or CPU usage specifically for the associated method/SQL group in detail.

2. PERFORMANCE PROBLEM DETERMINATION

Possessing a vast library of system performance issues and solutions collected over many years by our system performance analysts and consultants, Jennifer can utilize this knowledge to provide relevant clues and recommendations to system administrators, so that they can accurately and timely determine the cause of the issue encountered and provide counter-measures to correct them.

Web based system performance issues can be broken down into two broad categories: Relative Performance Problem caused by increase in workload and Conditional Performance Problem caused by unusual system activities. An example of the Relative Performance Problem is an increase in the number of concurrent users which can cause application response time to gradually increase, causing delays in data processing. An example of the Conditional Performance Problem can be JDBC connection failure or memory leakage which can cause temporary decreases of normal system operation.

- RELATIVE PERFORMANCE PROBLEM

Response time graph (X-View) can be used to compare and analyze SQL query transaction times to obtain clues of system performance issues. Providing this information along with SQL BIND variables can be of great help in SQL tuning. These tools can also be used to obtain quantified performance data which can identify the application that is most affected by the performance issue.

With a few mouse clicks within the dashboard, user can analyze connection count, request load, and active services count to pinpoint the issue to the problem application or back-end transaction. Dashboard can also be used for application tuning by collecting basic information such as CPU usage, response time, and SQL occupancy ratio. By analyzing changes in system CPU usage, daily concurrent user counts, and other statistical data, the user can estimate the acceptable maximum number of concurrent users. These are some of the most valuable factors needed in capacity planning and system expansion.

- CONDITIONAL PERFORMANCE PROBLEM

Jennifer is capable of monitoring and controlling JDBC connection resources by identifying JDBC connection status (IDLE/ALLOCATE/ACTIVE) and alerting the administrator of JDBC connection misuse and force connection termination as necessary to maximize process efficiency. By monitoring the system CPU/memory usage, JVM heap size changes, live object/collection counts, HTTP session dump, and TCP/IP socket usage, Jennifer can manage application vs. JDBC connection or system file usage to ensure efficient use of total system resources.

In case of dead-locks of threads, thread dump can be initiated to analyze the cause of the critical problem. By monitoring WAS/frameworks and other third party modules, process response time and CPU usage can be managed at the class/method level. Application experiencing hang-ups or queue build-ups can be paused or stopped as necessary and processing priority can also be controlled with ease.

- AUTOMATIC ALERTS

In the event of performance problem detection, Jennifer is capable of sending out automatic alerts to users and administrators via email or SMS system so that the alerts can be expeditiously sent to appropriate personnel for speedy problem diagnosis and resolution.

3. APPLICATION / SQL QUERY TRACE & TUNING

- APPLICATION CLASS/METHOD PROFILING

By examining application CPU usage and response time per class/method, Jennifer can identify which module is experiencing performance problems and convey the information in detailed reports.

- SQL QUERY TUNING & QUERY PLAN

Extraction of runtime for all SQL queries is possible. Comparative analysis of SQL query runtime and total application runtime can be used in query tuning and identifying bottle-neck issues per application that uses the said SQL query. These data can also be used for query planning.

- STATISTIC REPORTS OF APPLICATIONS

Jennifer can trace response time, CPU usage and other statistical information for all active applications per class/method and display it on screen via easily viewable tables and visuals

- BACK-END TRANSACTION TRACE

Jennifer is also capable of tracing communication between WAS and back-end systems (CICS, TUXEDO, WTC/Jolt, Mainframe CICS/CTG module) thus providing min/max response times, request loads, and other statistic data for all active transactions.

4. STATISTIC ANALYSIS & REPORTING

- STATISTIC SERVICES ANALYSIS

Jennifer can provide statistical data for service information such as the number of connections, response times, system resource usage, service call records, SQL transaction history, back-end transaction history, etc…

- CUSTOMIZABLE REPORTS

Stored data can be used for second level processing and displayed in visually effective charts, diagrams, and graphs on a separate screen. Using user defined customizable optional report templates, Jennifer can output various informative performance reports in Rich Formatted report files.

- QUANTITATIVE PERFORMANCE DATA

By collecting information such as connection counts, user counts, system resource usage, and other statistical performance data, Jennifer can provide quantified evidentiary data needed for planning system expansion, and application tuning.

5. PEAK LOAD CONTROL

The ability to adjust processing priority, number of active services, and service queue and perform load-balancing to meet system needs in the event that work load exceeds system processing capacity is called PLC (Peak Load Control).

This is an exclusive technology developed and used only by Jennifer. If used effectively PLC can produce system environments that can sustain 24X365 continuously functioning systems without system hang-up or downtime.

A typical performance issue is usually related to back-end database or host transaction process being overloaded, causing service queue to build up in the system. In this instance, even though the root-cause of the problem is in back-end, the effects can spread from back-end to front-end, ultimately causing WAS server and server processing capacity to max out, disabling any service capability and resulting in hang-ups and server crashes. In this scenario, the performance issue cannot be resolved by simply increasing the number of hardware machines, WAS servers.

The only effective solution for service queue build-up is to limit the number of active services according to true system capability thus eliminating the bottleneck and denying queue access of incoming service requests made by individual users while alerting them with a message, “Service is temporarily unable to process your request at this time, Please try again later. Thank you for your cooperation”. This method can be effectively used in preventing system hang-ups and crashes and minimize system and service downtime, enabling businesses to accommodate a reasonable level of service demand.

Jennifer can utilize PLC function without changing any of the application code levels but only some basic configurations. In addition, depending on the application priority, different PLC rules can apply for individual applications. The PLC priority level can be broken-down into 3 levels as described below.

6. FEATURES OF JENNIFER

• REAL-TIME MONITORING AND CONFIGURABLE DASHBOARD
• Effective and Intuitive centralized system monitoring
• Real-time active service status monitoring
• Real-time concurrent user count monitoring
• Real-time response time graphs
• JDBC connection health status monitoring
• Collection object / live object monitoring
• TCP/IP socket monitoring
• Application’s throughput (tps) monitoring
• User’s request pattern, Think time measure
• System resource monitoring (CPU, Memory, Disk, Network)
• Real-time peak load control support
• HTTP post key tracking and statistical analysis capability
• HTTP session dump
• Real-time memory leakage detection
• Real-time problem detection support
• Real-time trace of application name in URL and client’s IP address
• Standalone Java application (pure JAVA) monitoring
• Non-Java application monitoring support with user defined adapters based on Adaptable Framework technology
• Real-time Oracle and SQL query plan capability support
• Various adapter type support (shell script, telnet terminal, etc…) and real-time visualization
• User defined business data adapter support and real-time visualization on dashboard

• PROBLEM DETERMINATION
• Real-time performance bottleneck detection (class/method/SQL/transaction)
• Memory leakage detection (collection elements / live objects)
• Automatic service dump capability
• JDBC resource usage tracking capability
• Detection capability of double assigned JDBC connection resource
• Flexible and simple alerting feature (Alerting popup box / Email / SMS)
• System or JVM process down detection capability
• Counting capability of success or failure of applications
• SQL query information trace with detailed SQL BIND variables
• User customizable error alerting and delivery feature (ADF)
• Centralized management of all errors and warnings

• WORKLOAD CONTROL
• Peak Load Control (PLC) feature (applied to service groups)
• Real-time termination of active threads
• Changing of the thread priority

• MONITORING OF BACKEND TRANSACTION SYSTEM
• Trace of CTG (for CICS), WTC (for TUXEDO), Jolt (for TUXEDO), WebT (for TMAX)
• Tracing capability of any type of user defined application modules such as Socket, URL Connection, 3rd-parth framework
• CORBA based server monitoring
• Integration with 3rd party database monitoring solution
• Support for web server, SSO security server, and other non-WAS backend solution
• Support for Oracle ERP, SAP monitoring
• Support for Oracle database real-time performance monitoring

• STATISTICAL INFORMATION REPORTS
• Application level summarization
• Daily management of system errors by types???
• Top 10 list of bad performing applications and SQL queries
• Search capability of used and un-used classes
• 2nd level processing of information stored in the database
• User defined visual reporting with configurable user templates
• Capability for storage and visual reporting with user defined adapters
• Rich formatted visual reports with user defined report templates

• AUTHORIZED ACCESS BY USER TYPE AND GROUP
• Authorized access management per user type and group
• Authorized information data management per user type and group (visual components, SQL bind variables, thread termination feature, etc.)
• Authorized access management for critical functionalities (real-time thread termination, garbage collection)

• COMPREHENSIVE ENVIRONMENT
• Web based interface
• Integrated dashboard and display for several types of resources such as WAS, system, services, concurrent users, response time
• Configurable dashboard and drag & drop functionality