- Evolution of IT systems: As architectural patterns moved from monolithic systems to distributed systems, not all IT systems were moved to the newest patterns. Some new systems were built with new technologies and patterns whereas existing systems that were performing well enough continued on earlier technologies.
- Best of breed approach: With multi-tiered architectures, enterprises had the choice of building each tier using best of breed technology for that tier. For example, one system could be built using a J2EE container from vendor A, but a database from vendor B.
- Avoiding single vendors and technologies: Enterprises wanted to avoid dependence on any single vendor and technology. This led to systems being built using different technologies. For example, an order-booking system built using .NET technologies on Windows servers, but an order shipment system built using J2EE platform on Linux servers.
- Acquisitions and Mergers: Through acquisitions and mergers, enterprises have inherited IT systems that were built using different technologies. Frequently, new systems were added to integrate the systems of two enterprises but the new systems were totally different from the existing systems. For example, using BPEL process manager to integrate a CRM system with a transportation management system.
We see that each factor for diversity in the data center has some business or strategic value. At the same time, such diversity makes management of the data center more complex. To manage such data centers we need a special product like Oracle’s Enterprise Manager Grid Control that can provide a unified and centralized management solution for the wide array of products.
In any given data center, there are lots of repetitive operations that need to be executed on multiple servers (like applying security patches on all Oracle Databases). As data centers move away from high-end servers to a grid of inexpensive servers, the number of IT resources increases in the data center and so does the cost of executing repetitive operations on the grid. Enterprise Manager Grid Control provides solutions to reduce the cost of any grid by automating repetitive operations that can be simultaneously executed on multiple servers. Enterprise Manager Grid Control works as a force multiplier by providing support for executing the same operations on multiple servers at the cost of one operation.
As organizations put more emphasis on business and IT alignment, that requires a view of IT resources overlaid with business processes and applications is required. Enterprise Manager Grid Control provides such a view and improves the visibility of IT and business processes in a given data center. By using Enterprise Manager Grid Control, administrators can see IT issues in the context of business processes and they can understand how business processes are affected by IT performance.
In this article, we will get to know more about Oracle’s Enterprise Manager Grid Control by covering the following aspects:
Key features of Enterprise Manager Grid Control:
- Comprehensive view of data center
- Unmanned monitoring
- Historical data analysis
- Configuration management
- Managing multiple entities as one
- Service level management
- Automating provisioning
- Information publishing
- Synthetic transaction
- Manage from anywhere
- Enterprise Manager Product family
Range of products managed by Enterprise Manager:
- Range of products
- EM extensibility
Enterprise Manager Grid Control architecture.
- Multi-tier architecture
- Major components
- High availability
- Summary of learning
Key features of Enterprise Manager Grid Control
Typical applications in today’s world are built with multi-tiered architecture; to manage such applications a system administrator has to navigate through multiple management tools and consoles that come along with each product. Some of the tools have a browser interface, some have a thick client interface, or even a command line interface. Navigating through multiple management tools often involves doing some actions from a browser or running some scripts or launching a thick client from the command line.
For example, to find bottlenecks in a J2EE application in the production environment, an administrator has to navigate through the management console for the HTTP server, the management console for the J2EE container, and the management console for the database.
Enterprise Manager Grid Control is a systems management product for the monitoring and management of all of the products in the data center. For the scenario explained above, Enterprise Manager provides a common management interface to manage an HTTP server, J2EE server and database. Enterprise Manager provides this unified solution for all products in a data center.
In addition to basic monitoring, Enterprise Manager provides a unified interface for many other administration tasks like patching, configuration compliance, backup-recovery, and so on.
Some key features of Enterprise Manager are explained here.
Comprehensive view of the data center
Enterprise Manager provides a comprehensive view of the data center, where an administrator can see all of the applications, servers, databases, network devices, storage devices, and so on, along with performance and configuration data. As the number of all such resources is very high, this Enterprise Manager highlights the resources that need immediate attention or that may need attention in near future. For example, a critical security patch is available that needs to be applied on some Fusion Middleware servers, or a server that has 90% CPU utilization.
The following figure shows one such view of a data center, where users can see all entities that are monitored, that are up, that are down, that have performance alerts, that have configuration violations and so on. The user can drill down to fine-grained views from this top-level view.
The data in the top-level view and the fine-grained drill-down view can be broadly summarized in the following categories:
Data that shows how an IT resource is performing, that includes the current status, overall availability over a period of time, and other performance indicators that are specific to the resource like the average response time for a J2EE server. Any violation of acceptable performance thresholds is highlighted in this view.
Configuration data is the configuration parameters or, configuration files captured from an IT resource. Besides the current configuration, changes in configuration are also tracked and available from Enterprise Manager. Any violation of configuration conformance is also available. For example, if a data center policy mandates that only port 80 should be open on all servers, Enterprise Manager captures any violation of that policy.
Status of scheduled operations
In any data center there are some scheduled operations, these operations could be something like a system administration task such as taking a backup of a database server or some batch process that moves data across systems, for example, moving orders from fulfillment to shipping. Enterprise Manager provides a consolidated view of the status of all such scheduled operations.
Enterprise Manager provides a listing of all hardware and software resources with details like version numbers. All of these resources are also categorized in different buckets – for example, Oracle Application Server, WebLogic application Server, WebSphere application are all categorized in the middleware bucket. This categorization helps the user to find resources of the same or similar type. Enterprise Manager. It also captures the finer details of software resources—like patches applied.
The following figure shows one such view where the user can see all middleware entities like Oracle WebLogic Server, IBM WebSphere Server, Oracle Application Server, and so on.
Enterprise Manager monitors IT resources around the clock and it gathers all performance indicators at every fixed interval. Whenever a performance indicator goes beyond the defined acceptable limit, Enterprise Manager records that occurrence. For example, if the acceptable limit of CPU utilization for a server is 70%, then whenever CPU utilization of the server goes above 70% then that occurrence is recorded. Enterprise Manager can also send notification of any such occurrence through common notification mechanisms like email, pager, SNMP trap, and so on.
Historical data analysis
All of the performance indicators captured by Enterprise Manager are saved in the repository. Enterprise Manager provides some useful views of the data using the system administrator that can analyze data over a period of time. Besides the fine-grained data that is collected at every fixed interval, it also provides coarse views by rolling up the data every hour and every 24 hours.
Enterprise Manager gathers configuration data for IT resources at regular intervals and checks for any configuration compliance violation. Any such violation is captured and can be sent out as a notification. Enterprise Manager comes with many out-of-the-box configuration compliance rules that represent best practices; in addition to that, system administrators can configure their own rules.
All of the configuration data is also saved in the Enterprise Manager repository. Using data, the system administrator can compare the configuration of two similar IT resources or compare the configuration of the same IT resource at two different points in time. The system administrator can also see the configuration change history.
Managing multiple entities as one
Most of the more recent applications are built with multi-tiered architecture and each tier may run on different IT resources. For example, an order booking application can have all of its presentation and business logic running on a J2EE server, all business data persisted in a database, all authentication and authorization performed through an LDAP server, and all of the traffic to the application routed through an HTTP server.
To monitor such applications, all of the underlying resources need to be monitored. Enterprise Manager provides support for grouping such related IT resources. Using this support, the system administrator can monitor all related resources as one entity and all performance indicators for all related entities can be monitored from one interface.
Service level management
Enterprise Manager provides necessary constructs and interfaces for managing service level agreements that are based on the performance of IT resources. Using these constructs, the user can define indicators to measure service levels and expected service levels. For example, a service representing a web application can have the same average JSP response time as a service indicator, the expected service level for this service is to have the service indicator below three seconds for 90% of the time during business hours.
Enterprise Manager keeps track of all such indicators and violations in the context of a service and at any time the user can see the status of such service level agreements over a defined time period.