Professional Documents
Culture Documents
Executive Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
Executive Summary
Middleware has become an integral component of today’s IT infrastructures and is essential to service delivery. Consequently,
to ensure high quality service delivery, you must manage middleware with the same rigor and discipline as the other IT
infrastructure components. That requires a unified approach to IT service management that includes middleware.
BMC Software’s newly announced Middleware Management portfolio offers comprehensive middleware coverage with the
breadth and depth required for enterprise management of this key infrastructure component. BMC delivers this technology
through two primary solutions — BMC Middleware Management – Performance and Availability and BMC Middleware
Management – Transaction Monitoring.
You can easily integrate BMC Middleware Management solutions with BMC ProactiveNet Performance Management. Through
this integration, you can apply the power of BMC ProactiveNet Performance Management’s correlation technology to analyze
BMC Middleware Management data. As a result, middleware administrators can take advantage of BMC ProactiveNet
Performance Management’s advanced capabilities to facilitate middleware management. Conversely, the IT operations staff
can take advantage of the BMC Middleware Management solutions to gain greater insight into middleware for improved
performance management and root cause analysis.
The integration of BMC Middleware Management solutions with BMC ProactiveNet Performance Management brings
middleware management into the BMC Business Service Management (BSM) environment. It enables the IT staff to manage the
entire enterprise IT environment, including middleware, from the unified BSM platform.
This paper discusses the business criticality of middleware and the need to manage it in the context of the overall IT
environment. It examines how the integration of BMC Middleware Management and BMC ProactiveNet Performance
Management solutions addresses this need and it describes the resulting business benefits. The paper also describes how to
configure the solutions to accomplish the integration.
Each physical or virtual server may host one or more applications. What’s more, composite applications are segmented into
components that interoperate to provide application services. The components may run on different servers and each server
may host multiple components. A component may both call other components and be called by other components. Some
components may be supplied by external providers. The result is a highly complex web of interacting software components.
The components typically communicate through middleware of two main types. The first is messaging middleware that passes
messages between components that use the same messaging protocol and data representation. The two dominant messaging
middleware environments are IBM® and TIBCO middleware.
The second type is Enterprise Service Bus (ESB), or broker software, that passes messages between components that use
different messaging protocols and data representations. ESB software can also route messages. The most common ESB/
broker middleware environments are from IBM, TIBCO, and Oracle (BEA).
Because of its pivotal role in applications, middleware is a critical part of the IT infrastructure and is involved in a large number
of IT services, including critical business services.
Currently, many IT organizations manage middleware separately from other components of the IT infrastructure. This
separation increases the difficulty of overall service management. For example, when a performance problem occurs in a
service, it’s difficult to quickly determine the source of the problem. Is it the physical or virtual device hosting the application? Is it
the application? Or is it the middleware? As a result, problems often trigger highly inefficient “all hands on deck” exercises.
IT needs to take a unified approach to IT service management that encompasses all components of the IT infrastructure,
including middleware. That requires viewing middleware in the context of the big picture of business processes and service
delivery. For example, a business process may involve multiple services provided by multiple enterprise applications. Figure 1
illustrates the big picture of an order entry business process.
1
Begin Order Order Order Order End Business Process
Entry Approval Shipping Billing
Transport Layer
-JMS -MQ
-HTTP(S) -XMS
Mediation Layer
-Message Broker
-WESB
Transport Layer
-JMS -MQ
-HTTP(S) -XMS
Services
A unified approach eliminates the “all hands on deck” scenario in addressing problems in services that involve middleware.
Here’s an example:
A retail chain has a mainframe pricing application that communicates through middleware with retail outlets that use Windows.
The application keeps the retail stores updated on the latest promotional pricing. It’s critical that the pricing information reach all
retail stores in a timely fashion to prevent mischarging customers.
For some reason, the pricing information is no longer getting to the retail stores in a reasonable amount of time. Is the problem
in the mainframe application, the Windows application, or the middleware? Through a big picture view of the price updating
application, the IT staff sees that the messages are leaving the mainframe, passing though the middleware in a normal
fashion, but waiting for the store-side Windows application to process them. This indicates that the problem is in the Windows
application. Armed with this knowledge, the middleware administrator can work directly with the owner of the Windows
application to address the problem rather than wasting time digging into the mainframe application or middleware components.
The data gathered by BMC Middleware Management – Performance and Availability can include:
2
BMC Middleware Management – Performance and Availability aggregates historical data in a BMC Middleware Management
database, which can be a DB2®, Oracle®, or Microsoft SQL Server. You configure the BMC Middleware Management –
Performance and Availability solution to determine what data is to be collected and how often. The solution can present the data
in user defined views, reports, and charts.
The second solution is BMC Middleware Management – Transaction Monitoring, which tracks how specific transactions are
moving through the process. BMC Middleware Management – Transaction Monitoring looks at the content of messages to allow
tracking of business transactions and their content and timing from hop to hop as they move through middleware environments.
The solution provides an advanced data correlation capability. Traditional correlation solutions employ two types of correlation.
The first is time-based in which only those events that occurred at approximately the same time are correlated. The second
type is topology-based, which takes into account the physical and logical topology of IT resources. Physical topology correlation
considers events that are generated by physically interdependent IT resources (such as servers and network devices) to be
correlated. Logical topology correlation considers those events generated by logically interdependent IT resources (such as
applications and databases) to be correlated.
In addition to time-based and topology-based correlation, BMC ProactiveNet Performance Management employs a third and
more advanced type of correlation that distinguishes between normal and abnormal behavior. It also employs an automatic
learning process that analyzes workload profiles and determines the normal operating behavior of the environment. It uses
this information to automatically set baselines for normal/abnormal event determination. It then generates events only when it
detects abnormal conditions, greatly decreasing the number of events to which the IT staff must respond. It also automatically
and continually readjusts the baselines to accommodate normal workload variations over time.
What’s more, BMC ProactiveNet Performance Management employs advanced techniques to correlate performance data and
forensic data to generate intelligent events that indicate probable causes of abnormal events. It creates a short list of probable
causes, ranked by probability, facilitating problem triage and resolution.
First, BMC ProactiveNet Performance Management provides a comprehensive IT infrastructure-oriented view. This view
includes all the hardware and software assets in the infrastructure, their physical and logical topologies, the services
provided by the assets, the business priorities of the services, and the business processes supported by the services. Second,
the middleware management solutions provide an application-oriented view of the IT infrastructure. This view shows the
transaction flow through the components of composite applications, including such components as web services, middleware,
and backend components. These two complementary views are analogous to having both a map of city streets and a view of the
traffic flow through those streets.
Middleware Administrators
Middleware administrators can take advantage of BMC ProactiveNet Performance Management’s dynamic baselining capability
to automatically set and continually adjust baselines for BMC Middleware Management objects, such as queue depths or
message rates. The administrators can configure BMC ProactiveNet Performance Management to automatically generate
intelligent events when abnormal conditions are detected, that is, when conditions exceed the baselines. For example, a large
daily batch application may always cause a queue to back up. Setting a hard queue depth threshold may result in an alert being
generated each time the batch application is run, even though the increase in queue depth is within the normal range for the
batch application. Instead of setting a hard threshold, BMC ProactiveNet Performance Management the current queue depth
3
for the batch application to that of other daily batch runs and generates an alert only if the depth exceeds the normal range.
This predictive analysis enables middleware administrators to move proactively to head off issues before they result in service
degradation while avoiding false positive alerts.
In addition, through BMC ProactiveNet Performance Management, middleware administrators can get a “big picture” view of the
IT environment to help them gain greater insight into issues. For example, a queue manager outage that normally requires an
alert to the middleware administrator may be due to an outage of the Linux system on which the queue manager runs. Because
BMC ProactiveNet Performance Management has awareness of the Linux server status as well as the queue manager status,
alerts can be better directed.
IT Operations
Operations personnel benefit in that they can view BMC Middleware Management reports and graphs to “look inside” the
middleware and gain additional insight into performance and availability issues. Here are two examples:
The graph in Figure 2 shows application processing across two load-balanced servers. The load balancing was effective for
most of the visible time period, however, from 13:00 – 13:15, all application processing was being routed to Server2 (shown
in blue) while the Server1 processing (shown in red) dropped to zero. Seeing this, the IT operations staff switched to the BMC
ProactiveNet Performance Management view and saw that Server1 was operating normally. This narrowed the failure to the
application instance running on Server1 rather than in Server1 itself. The operations staff referred the problem to the application
developer for resolution.
4
Figure 3. BMC Middleware Management report
The BMC Middleware Management report in Figure 3 shows response times of a set of WebSphere Application Server EJB
methods. (Each EJB declares an interface consisting of one or more EJB Methods. Each EJB Method is a specific function of the
bean.) Through the chart, the operations staff can see which methods are the limiting factors in the bean code. In this example,
the table is sorted by response time. As can be seen, the updateWarehouseInventory method is taking too long (over 24 seconds
which is the limit of the performance of the bean). Having gained this insight, the operations staff can task the application
developer to target efforts to improve the code for that particular method.
BSM Implications
The integration of BMC Middleware Management solutions with BMC ProactiveNet Performance Management brings
middleware management into the BMC BSM environment, so that all IT groups, including middleware administrators, operate
from the same unified BSM platform.
Bringing middleware into the BSM environment enables middleware administrators to work far more efficiently with other IT
groups in a wide variety of areas, including database management, application management, event management, incident and
problem management, performance management, and capacity management. The result is higher service quality and lower costs.
1. First, you expose the BMC Middleware Management database to BMC ProactiveNet Performance Management in the same
manner as you would expose this database to other management tools. This enables BMC ProactiveNet Performance
Management to execute queries against the BMC Middleware Management database.
2. Then, you configure ProactiveNet to execute queries periodically against the BMC Middleware Management database in
the same manner as it executes queries on other performance databases. The database appears to BMC ProactiveNet
Performance Management as just another performance history database.
5
Connecting BMC ProactiveNet Performance Management to the
BMC Middleware Management Database Server
ProactiveNet
Server
BMM Database
ProactiveNet Server
Agent BMM
Database
ProactiveNet BPM
Proxy Agent ODBC
BMM Knowledge
Module Perl Script
As Illustrated in Figure 4, the connection between BMC ProactiveNet Performance Management and the BMC Middleware
Management database server involves of the following components:
»» The Proactive Operations Platform Server. The repository for all data collected by BMC ProactiveNet Performance
Management. The server analyzes data, creating profiles, for each metric, that defines normal variation in parameter value
over time. It detects deviations from the profiles (abnormalities) and uses this information to create alerts and to perform
probable cause analysis.
»» BMC ProactiveNet Performance Management Agent. Collects data from monitored systems and other monitoring tools
»» BMC ProactiveNet Performance Management Proxy. Component within the BMC ProactiveNet Performance Management
Agent that allows the BMC ProactiveNet Performance Management Agent to connect to and collect data from multiple BMC
Performance Manager Agents.
»» BMC Performance Manager Agent. Software agent that resides on each monitored system.
»» BMC Middleware Management Knowledge Module. Provides the intelligence and encapsulated knowledge for the BMC
Performance Manager Agent. It defines what parameters the BMC Performance Manager Agent monitors, and how to collect
and present the data.
The server accesses the BMC Middleware Management database through the BMC Performance Manager Agent and the
Knowledge Module, all installed on the BMC Middleware Management database server. The Agent collects data through the
Knowledge Module using a Perl script that operates through an ODBC connection. The Knowledge Module executes the Perl
script according to the schedule that you specify.
6
Configuring the Solutions
This section describes the configuration of the BMC Middleware Management and BMC ProactiveNet Performance Management
solutions to permit BMC ProactiveNet Performance Management to access the BMC Middleware Management database. It is
assumed that you have already installed BMC ProactiveNet Performance Management and BMC Middleware Management –
Performance and Availability or BMC Middleware Management – Transaction Monitoring.
To set the history rules, select the History tab of the BMC Middleware Management – Management Console and perform the
following steps:
1. Select, from a list of available object types, the type of object (such as an MQ Queue or a Service Response Time) on which
you want to keep history.
2. Select, from a list, which parameters of that object type (such as queue depth) that you want to collect.
3. Select the data compression interval. For example, you may want to keep the response time for every transaction for 1 day,
but after that keep only the average response times for all transactions in a given hour. In this case, you would choose 1 hour
as the data compression interval.
NOTE: Instead of creating a rule many times, you can create a template once and associate objects with the template. A
template applies the same history rule or set of history rules to similar objects, such as queues, throughout the enterprise.
You can edit templates to customize them to your specific requirements. BMC Middleware Management – Performance and
Availability and BMC Middleware Management – Transaction Monitoring include a number of pre-written templates.
You then set the sample interval in the BMC Middleware Management Configuration Monitor
You can create reports that present the history data in the form of spreadsheets. The BMC Middleware Management solutions
include templates that are prepopulated for each object. You can customize the reports to meet specific requirements.
You can also create charts and graphs of the data for any middleware object, enabling you to easily analyze and compare the
data aggregated by the BMC Middleware Management solution. Charting is controlled by a wizard. You simply step through the
wizard to create the general chart you want. You can then customize the appearance and labeling of the chart. Several chart
types are available, including two-dimension charts (line, scatter, area, and step), and three-dimension charts (bar, x bar and
range, Pareto, and linear regression).
NOTE: For a detailed description of configuring BMC Middleware Management history collection and presentation, refer to
the BMC product documentation for BMC Middleware Management – Performance and Availability.
7
To configure the UDFs:
1. Connect to your database as a highly authorized user (such as the schema owner).
2. Execute the BMC Middleware Management provided scripts matching your database and version.
3. Check the log files to ensure that the UDFs are correctly installed.
NOTE: The exact details of configuring the UDFs depend on your operating system and database platform. For a detailed
description of configuring UDFs for your environment (such as DB2 on AIX), refer to the BMC product documentation for
BMC Middleware Management – Performance and Availability.
1. Launch the BMC ProactiveNet Performance Management Administration Console and select Tools -> Configure -> BMC
PATROL Proxy
2. Enter the connection details for the BMC PATROL Proxy Agent
3. Add the BMC Performance Manager Agent, ensuring that all connection credentials are correct
4. Once the proxy has successfully connected to the agent, a list of application classes present on the agent is displayed
5. Expand QNAMI_TOPIC and select all instances of the QNAMI METRICS application class
6. Click Finish
NOTE: For more details on the configuration procedure, see the BMC ProactiveNet Performance Management
Administration Guide.
8
Configuring Launch of BMC Middleware Management from the
BMC ProactiveNet Performance Management Console
To configure BMC ProactiveNet Performance Management to launch the BMC Middleware Management Console, you create a
script from the ProactiveNet Administration console, using the BMC ProactiveNet Performance Management Diagnostic Wizard.
The procedure is straightforward:
You can now cross-launch to the BMC Middleware Management – Management Console from the BMC ProactiveNet
Performance Management User Console by navigating to an instance of a QNAMI_Metrics Monitor and selecting Tools ->
Diagnostics -> BMM Console
9
Conclusion
Composite applications play an essential role in the delivery of critical business services and middleware is a crucial link in
most composite applications. Consequently, it’s imperative that IT keep middleware components running at the performance
and availability levels required by the business. Now you can tap the power of BMC ProactiveNet Performance Management to
enhance middleware management by integrating with BMC Middleware Management solutions through simple configuration.
Through this integration, you’ll increase the efficiency and effectiveness of both middleware administrators and the IT
operations staff in managing middleware. Middleware administrators can take advantage of BMC ProactiveNet Performance
Management’s proactive management capabilities. The operations staff can gain greater insight into middleware through the
availability of BMC Middleware Management aggregated data.
In addition, you’ll integrate middleware management into the overall fabric of Business Service Management. That results
in higher quality of service, lower costs, and less risk of service disruptions and noncompliance with internal polices and
external regulations.
BMC, BMC Software, and the BMC Software logo are the exclusive properties of BMC Software, Inc., are registered with the U.S. Patent and Trademark Office, and may be registered or pending
registration in other countries. All other BMC trademarks, service marks, and logos may be registered or pending registration in the U.S. or in other countries. All other trademarks or registered
trademarks are the property of their respective owners. © 2010 BMC Software, Inc. All rights reserved. *127217*