Managing application and infrastructure performance is consuming an increasing amount of time and attention from enterprise IT. This is a direct result of the on-going increase in the complexity and interdependency among infrastructure, applications and functions required to deliver business services. Compounding the problem, enterprise demands on and expectations from IT and data center services continues to grow despite flat budgets and static staffing.
Stronger operational links to and impact on business operations means enterprise IT faces increasing pressures to deliver business services that are more reliable, adaptable, faster and innovative in function and delivery to keep existing customers, while attracting new ones.
On the down side is the business risk of catastrophic failure of a vital service when an additional service request is made on the wrong server at the wrong time. To deal with these pressures and demands, IT must proactively manage the performance and availability of infrastructure, applications and services. Potential problems need to be identified, and, if possible, be resolved before a service failure occurs.
To maintain reliable services and avoid problems, IT must be able to visibly see the impact of change, receive early warning of potential problems and get additional insight into cross-functional operations. Additionally, a growing scarcity of experienced IT resources increases reliance on less skilled staff while highly qualified individuals are spread thin to handle multiple assignments.
The cumulative result is significant pain in the operations center. Fortunately, this is now being addressed with a new generation of intelligent, automated system performance management tools.
Integrated performance management
The changing role of IT and its relationship to the enterprise dramatically changed performance management. As staff responsibilities have moved beyond simply keeping the infrastructure up and running, contemporary performance metrics have become more complex than response times, percent processor utilization or uptime. It isn't that these things are not critical; they are and will remain significant to data center operations.
But it's an integrated and complex world. All too often forgotten is the fact that actions have consequences. In today's interdependent environments, unilateral action taken without considering impact in other parts can have disastrous consequences. Infrastructure, applications and resources can no longer be managed as separate entities or viewed as silos. To solve this requires an increase and focus on automation, data sharing, cooperation and communications across all parts of the IT and business functions, as well as tools that support such activities. The solution requires integrated, intelligent and automated management that functions across the enterprise.
The marketing name for this concept and the solutions that support it is Business Service Management (BSM). Solutions are available that will provide early warning of potential service problems through intelligent, automated analysis of the full range of enterprise and IT management data.
For our purposes, we'll focus on performance and capacity management. The ability to proactively anticipate problems and recognize actions that will result in performance degradation remains a key factor in maintaining infrastructure and service performance. Traditional operational performance management monitors and analyzes both current and past data about the performance of services and the infrastructure in order to maintain on-going and future performance at acceptable, committed service levels. To meet evolving needs, IT needs to anticipate potential problems, whatever the source, early enough to take action to avoid or minimize negative service impact.
Intelligent automation allows non-expert IT staff access and ability to effectively use analytics, reporting and visualization capabilities in operational settings without requiring extensive training. For example, intelligent analytics using workload forecast data can predict infrastructure performance and identify potential bottlenecks. End-to-end views of service delivery tied to infrastructure operations can provide advance warning of capacity bottlenecks and potential performance problems. Enterprise and IT staff working together can determine the cause of the problems such as application design, processor overload, or network saturation, and initiate proper corrective action.
Case study: BMC's Performance Assurange Management for Mainframes
All of the major players, as well as many mid-tier players offer solutions to implement BSM. There are too many, with new ones emerging all the time, to do any detailed comparison here. A recent project gave us some insight into the effectiveness of one vendor's product, BMC's Performance Assurance Management for Mainframes. We looked at two of the major proof-of-capabilities.
The first capability is connecting IT operations with business metrics. The ability to identify and track metrics that provide meaningful information about these two areas represents a major problem. BMC's Performance Assurance includes and allows input of external business metrics to permit analysis and inclusion in reports with IT metrics.
We spoke with a customer successfully combining IT performance metrics with business planner capacity forecasts. This customer was able to effectively plan for future investment and operational changes in its IT resources based on business needs.
Second, it is critical to have effective automated analysis that can correctly identify what is causing a problem. Equally important is taking the next step for automated recommendation and initiating needed remedial action. A solution that does not at least minimally match the performance of an experienced staff member is unlikely to meet enterprise needs.
At one company, a new employee familiar with BMC tools with limited mainframe skills took one hour to perform an evaluation of an application performance trend. An experienced mainframe specialist using traditional tools took about four hour for the same analysis. They both arrived at essentially the same conclusions.
There are plenty of other examples out there with solutions from other vendors. The message is that the challenges of effective performance management for the enterprise and IT can be addressed following a strategy of BSM and implemented with intelligent, integrated performance management tools. The necessary tools exist and are being used today with improvements and enhancements emerging nearly every day.