AUTONOMIC AND METRICS The transition to full autonomic computing does not happen in a short time frame. The implementation must be carefully planned and monitored and cannot be solely accomplished by acquiring IBM products. Autonomic computing implementation still requires sound IT project management practices. Other skills in the organization must adapt and change, as new processes need to be introduced to create long-term success. As corporations progress through the five levels, the tools, processes, and skill requirements become increasingly sophisticated. Also the systems will begin to become more closely aligned with the business units. The Basic level—level 1—represents the starting point for most IT departments and their corporations. The IT department needs to introduce formal measurements and metrics to determine a reference baseline for the introduction of this new technology. If a cost center approach is not being applied, one needs to be introduced. This ensures that the IT resources used are reviewed and measured as an investment. The Managed level—level 2—is measured by the availability of resources for tasks such as the time it take to solve and close trouble tickets in their internal problem management systems. To improve on these measurements, IT departments must continually seek formal process improvements through existing channels and document their efforts. These channels will mainly be manual processes. This has been the formal approach for almost all IT departments for the last 20 years or more. At the Predictive level—level 3—IT departments are measured on the availability and performance of the business systems and their return on the investment. To improve at this level, IT organizations must measure and analyze transaction throughput and performance. It is at this level that IT systems are established as a critical role in business success. Predictive tools are introduced for the first time to forecast future expected IT performance. Many of these tools will make recommendations for future performance improvement and enhancements. At the Adaptive level—level 4—IT resources are automatically provisioned and tuned to provide optimal performance. Business policies are introduced, and can be defined and amended, online. Service level agreements and business priorities will guide autonomic infrastructure behavior. IT departments are measured on their performance such as response time improvements over previous levels and the degree of efficiency as they respond to new problems, and differing workloads and priorities. At the final Autonomic level—level 5—true autonomic computing is established. Measurements begin to determine how the IT department will make the business more profitable. Extensive financial metrics are available. Advanced modeling techniques are used as well, to optimize the business performance and assist in rapidly deploying new business applications at the same level, which is fully autonomic. Table 10.2 summarizes the support process needed to make the transition to fully autonomic computing. The road map to success with autonomic computing requires significant changes in processes, skill evolution, more simplified management practices, new technologies and architectures, and implementation of open industry standards. Probably the most important change to be made will be in the IT culture, as there will be fundamental changes in the way systems are managed. Table 10.2. The Processes, Tools, Skills, and Benchmarks Needed Through the Incremental Delivery Process of Autonomic Computing | | Basic Level 1 | Managed Level 2 | Predictive Level 3 | Adaptive Level 4 | Autnonomic Level 5 | | Process | Informal, reactive, manual | Documented, improved over time, leverage of industry best practices, manual process to review IT performance | Proactive, shorter approval cycle | Automation of many resource management best practices and transaction management best practices, driven by service-level agreements | IT service management and IT resource management best practices are automated | | Tools | Local, platform and product-specific | Consolidated resource management consoles with correlation of events, problem management system, automated software install, intrusion detection, load balancing | Role-based consoles with analysis and recommend-ations; product configuration advisors; real-time view of current and future IT performance; automation of some repetitive tasks, common knowledge base of inventory and dependency management | Policy-management tools drive dynamic charge based on resource-specific policies | Costing/financial analysis tools, business and IT modeling tools, tradeoff analysis; automation of some e-business mangement roles | | Skills | Platform-specific geographically dispersed with technology | Mulitple platform skills, multiple management tool skills | Cross-platform business system knowledge, IT workload management skills, some business-process knowledge | Service objectives and delivery per resource, analysis of impact on business objectives | E-business cost and benefit analysis, perfomance modeling, advanced use of financial tools for IT context | | Benchmarks | Time to fix problems and finish tasks | System availability, time to close trouble tickets and work requests | Business system availability, service-level agreement attainment, customer satisfaction | Business system response time, service-level agreement attainment, customer satisfaction, IT contribution to business success | Business success, competitiveness of service-level agreement metrics, business responsiveness |
|
