Right-sizing Hardware

Right-sizing and configuring systems can increase efficiency and therefore profitability. Wrong-sizing has the opposite effect. Too little capacity and you suffer poor response, low productivity, and lost business. Too much capacity and you unnecessarily waste resources, increase complexity, and decrease manageability, wasting both time and money. And right-sizing hardware can help reduce software costs too. Since many licensing fees are tied to server size, downsizing systems can bring significant software savings.

Whether you are undergoing a hardware refresh, consolidating servers, or purchasing hardware for new application roll-outs, it is valuable to have accurate and objective advice regarding the right amount of capacity to purchase and the right hardware and software configuration to use to get the most benefit from your IT investment.

TeamQuest Software Addresses Right-sizing Hardware and Software Purchases

TeamQuest software provides you with objective, accurate analysis for justifying and right-sizing hardware and software purchases. Using TeamQuest Analyzer and TeamQuest Predictor, you can examine different configurations to find the most cost-effective solution over a specified period of time that best satisfies service levels and business needs.

TeamQuest Analyzer enables you to do in-depth analysis of resource utilization. Using powerful workload processing where all system activities can be analyzed based on the application, service, or business unit that initiated them, TeamQuest Analyzer provides detailed insight into the characteristics and behaviors of your existing systems from a business service point of view.

TeamQuest Predictor adds the ability to predict the outcome of potential future scenarios. Starting from a viable baseline from an existing system or from a pilot of a new system, a wide spectrum of what-if scenarios can be quickly evaluated using analytic modeling. Without having to reconfigure a single piece of hardware or software at the server, the following changes can be evaluated:

• Transaction intensity by workload (application, service or business unit)
• Number of users
• Reconfiguration of hardware
• Consolidation of applications running on separate physical servers

Click for larger view Click for larger view

These charts show components of response time before and after CPU was added. Comparing these charts shows the effects of adding processing power. Several other useful statistics are also supplied, such as resource utilization, active resources, resource utilization by workload, stretch factor and many others. Stretch factor is the single most important number for showing bottlenecks in a system. It is the ratio of response time divided by the response time with no queuing delays. The lowest possible number is 1.0, where there are no queuing delays. When the stretch factor exceeds 2.0, it means that queuing delays equal the actual time spent doing work. With linear trending, the trend will continue in a straight line. Most servers do not behave this way. In reality, as the server becomes more saturated, the response time will grow exponentially. This curve pattern is often referred to as a hockey-stick pattern and is accurately determined by TeamQuest Predictor using an analytic queuing network solver. To take the guesswork out of hardware and software configurations, it is essential to understand your current systems and use tools to evaluate the effects of changes. A formal, regimented approach to system sizing will lead to greater predictability, less firefighting, and increased productivity.

Need to right-size hardware? Contact TeamQuest