Intelligent Workload Measurement
Novell Operations Center monitors and measures service level performance
Written by Bill Tobey
If there’s any management function that becomes indispensable as we start repackaging key business services as portable workloads, it’s the ability to monitor state and measure performance no matter where those workloads are running. Measuring service performance was hard enough when everything ran on physical resources safely within our own firewalls. Today we’re virtualizing many of those physical systems, seeding clouds in our own data centers, consuming some applications as SaaS while running others in public infrastructure clouds like Amazon EC2.
These hybrid environments add tremendous complexity to service management, and the fact that we no longer own or operate all the spaces where our services run doesn’t begin to get us off the hook for our SLA commitments. We started down the road to virtualization and cloud computing in search of standardization, efficiency, service quality and economy; if we can’t measure service performance accurately there’s no way to know whether we’re accomplishing anything.
IDC describes the challenge like this: “The nearly constant stream of workload provisioning and system configuration changes that occur in highly virtualized cloud computing environments results in the creation of a large volume of events, logs, and notifications across the system. IT organizations need tools that can effectively correlate these data streams and provide relevant metrics to track service level, root cause, compliance and fulfillment status, and drive automated remediation activities as needed.”
In fact, monitoring and measurement require two distinct types of functionality:
- First, the ability to connect all the available silos of IT information, correlate events in the IT infrastructure with the services they affect, and monitor those event streams for significant service impacts.
- Second, the ability to integrate IT and business performance metrics in ways that capture the contribution and significance of each service to the business, are meaningful to business users, and are continuously available to all stakeholders in near real time.
Novell Operations Center: The WorkloadIQ Solution for Service Measurement
In the Novell portfolio of WorkloadIQ solutions, Novell Operations Center provides management, monitoring and measurement of key business services across physical, virtual and cloud environments.
Novell Operations Center monitors service performance and availability, models the relationships between services and the underlying infrastructure, and provides automated real-time measurement. It integrates the outputs of existing management tools with business context, creating a business-centric view of the most complex IT environments. It is available with connectors to most leading third-party systems and infrastructure management solutions, and is, of course, tightly integrated with all Novell identity, security and management solutions, including:
- Novell Cloud Manager
- Novell PlateSpin Recon
- Novell Sentinel
- Novell Identity Manager
- Novell ZENworks (includes Novell ZENworks Asset Management, Novell ZENworks Configuration Management, and Novell ZENworks Linux Management)
Novell Operations Center is available as a complete business service management solution, or in three focused solutions that let customers build their service management capabilities flexibly and incrementally.
Novell Operations Center for Service Measuring monitors and manages performance in real time, alerting you before service levels drop. It integrates business context and prioritization rules into measurement activities, accounting for calendars, transaction value and volume, and other business factors to reveal bottom line impacts. Historical trend analysis provides guidance for future resource use, management and improvement. All these measurements are collected and consolidated in real-time dashboards that transform complicated technology into actionable, easy-to-understand business communications.
Novell Operations Center for Service Monitoring consolidates all available data sources, providing end-to-end service health monitoring of mixed-IT environments in a true end-to-end service management view. It helps business users understand service performance and IT to recognize business impact. Important new features include:
- Event Manager collects, filters, de-duplicates and normalizes line-oriented event data from any source. Working in sync with the Novell Operations Center intelligent service model, Event Manager turns raw data into clear, pertinent information that IT and business teams can understand and respond to.
- Experience Manager uses passive, synthetic transactions to measure end-user response times for Web-based applications.
Novell Operations Center for Service Mapping provides automated mapping and policy-based management of the complex relationships between key business services and the underlying IT infrastructure. Key components of this solution include a configuration management system, federated configuration management data base (CMDB) and asset management and discovery integration.
We started down the road to virtualization and cloud computing in search of standardization, efficiency, service quality and economy; if we can’t measure service performance accurately there’s no way to know whether we’re accomplishing anything.
Service Monitoring and Measurement in Complex Mixed Environments
Novell Operations Center gives IT and line-of-business personnel a real-time view of business service health and performance, regardless of whether the workloads that deliver those services are running internally or externally; in physical, virtual or cloud environments. (See Figure 1.) All that’s required is that workloads are configured with on-board intelligence and are reporting to some management solution with a standards-based API.
For workloads running in public infrastructure-as-a-service (IaaS) cloud environments such as Amazon EC2, Novell Operations Center connects with the indigenous management solution and retrieves all relevant information and events as they are made available. In the case of Amazon EC2, a new adapter connects to Amazon CloudWatch, allowing Novell Operations Center to retrieve and consolidate operational and performance metrics for all live AMI instances. Experience Manager provides an independent measure of health and end-user responsiveness through synthetic transaction testing.
For SaaS services, Novell Operations Center monitors health and measures performance in the same manner, through integration with the service providers’ management systems and independent transaction testing. For internal cloud environments, Novell Operations Center connects directly with Novell Cloud Manager, as well as most other virtual systems management tools.
Finally, for workloads running on traditional physical infrastructure, Novell Operations Center connects with the existing systems and infrastructure management solutions to collect and integrate health and performance metrics and events.
Business-Centric Service Monitoring: The View from the Dashboard
So what does a business-centric view of an IT service actually look like? In this view of a typical Novell Operations Center dashboard, the portlet display at the upper left shows an overview of a manufacturing company’s order handling process, with a separate health indicator for each major process segment from order capture through materials procurement, manufacturing, finished product delivery, billing and payment processing. (See Figure 2.) The process detail display immediately below informs us that the data inputs for our process health indicators include feeds from an ERP monitoring system, an event management system, and an incident management system.
The yellow indicators in the process overview are alerting us to major business impacts to the order capture and manufacturing processes, while the process detail display indicators show us that the alarms have been triggered by data from the ERP and event management systems.
By selecting the Root Cause display from the Actions menu in the Overview portlet, we find that two system events are responsible for the alarms—a network card and a system node CPU have failed. (See Figure 3.) Novell Operations Center has correlated these two otherwise unrelated events into its model for the order processing service, allowing us to see exactly where these infrastructure events will impact business services and operations. With this information in hand we can prioritize our response activities accurately and appropriately to get these systems and the services they support back on line ASAP.
The process overview dashboard also provides real time information about service performance against a critical SLA. In this example the company has set a goal that 90 percent of incoming orders will be processed within four hours of receipt. The tracking portlet at the bottom of this view displays processing data from the ERP system. (See Figure 4.) It shows that current performance against that goal is safely within target range.
If we wish to see performance trends over a longer interval, historical analytics allow us to interrogate detailed historical data through any type of view we choose to define. (See Figure 5.)
Business Service Monitoring and Measurement Made Simple
As the IT infrastructure that delivers critical business services becomes an ever-more dynamic and complex mixture of physical, virtual and cloud-based resources, IT organizations and their customers throughout the business need easy-to-use tools for monitoring services and measuring their performance regardless of where the underlying workloads are running. Novell Operations Center brings together all the available information about infrastructure state and service performance in an intuitive, business-centric presentation that gives every stakeholder exactly the information they need to optimize their contribution to the business.
 IDC, White Paper Sponsored by Novell, "Intelligent Workload Management: Opportunities and Challenges," Doc.#223661, June 2010.