One of the most significant problems continues to be a lack of knowledgeable DBA resources, especially in the area of managing Oracle databases and applications. This skills crisis continues to worsen, necessitating many IT managers to search for tools to reduce the number of DBAs required to maintain the systems and automate more manual processes in order to do more with less staff. Not surprisingly, a recent commerce statistic estimated that there are over 400,000 IT jobs open due to the lack of qualified candidates. As the gap continues to increase, IT managers are looking for tools with built-in intelligence and inter-automation capabilities that can simplify the process of managing a highly distributed environment.

For most IT organizations today, around-the-clock management is not just a goal, it’s a requirement. And global companies are often faced with a shrinking maintenance window of minutes or hours, instead of days. If needed work cannot be completed, mission critical systems go down and can have a large financial impact to the business. As these time windows continue to shrink, IT managers are forced to prioritize database, server and network maintenance and enhancements, completing only the most critical tasks. Companies that fail to put the infrastructure in place prior to running production systems usually also fail to allocate sufficient time and resources to set up a proactive monitoring and management environment. The result is that they often experience the pain of unpleasant surprises like downtime, excessive fragmentation, memory, and poorly running SQL. There is insufficient time to fix problems and prevent them from reoccurring in the future.

As a company’s business expands, what will the growth rate of the support staff need to be? Is it possible to delay or defer hiring by deploying tools designed to automate repetitive tasks? James Martin in his book, "Cybercorp" said, "Corporations cannot design and maintain the complex integration of systems that is needed unless computers themselves are used for that task. "The automation of automation is essential. ^1". Automation is one of the most significant reasons that companies have embarked on framework initiatives to integrate disparate technologies or have tried, with limited success, to integrate tools on their own.

IT managers face constant pressures to firefight and solve problems reactively. Workloads continue to increase without adequate IT staffing to expand the infrastructure. This causes an increasing number of operational problems and taxes the help desk. And as more and more companies approach new technologies like the Intranet and Internet, existing client server models and capacity plans fall apart. Managers can no longer gauge the impact of new applications on IT resources, so they use their best judgement to over-engineer networks, hoping to guarantee fast response time and adequate I/O ratios. Many problems also occur because data about the operation of the enterprise is not shared collaboratively so that staff can work together to solve problems.

In the midst of all of this change, financial managers are enforcing new policies, which force IT to quantify and cost-justify all technology investments to clear business objectives. IT Managers scramble to develop cost justifications and estimate capacities to support the budget approval process typically making a variety of assumptions that are sometimes incorrect -- and more often understated -- based on the tasks at hand.

While the challenges continue to compound, technology continues to evolve and push toward real automation of the enterprise. Creating a managed, proactive environment, although difficult, can be accomplished through good planning, knowledge of critical business processes and implementation of software tools which allow you manage and improve the IT infrastructure.

One of the most important first steps in developing a managed environment is managing application and systems performance. Many IT managers make the mistake of approaching performance management from only a singular, checklist type of approach without really considering all the aspects of performance that relate to a specific application or business process. Chris Loosley, co-author of the book, "High Performance Client Server," ² suggests that many companies need to realize that the scope of performance needs to be expanded to consider the infrastructure, the business process, application flow, and business impact, as well as the hardware and software environment.

Diagram 1: High-Performance Client/Server

Loosely suggests that, "At the most fundamental level, almost everything we need to know about performance of any computer system can be described in terms of five simple concepts,"³ which must be understood, including:

According to Loosley, workload, response times and throughput represent external sources that describe the observed performance. He says, "Resource utilization and resource service times describe how the computer system behaves internally, providing the technical explanation for the externally observed performance characteristics."³ This concept for managing performance is interesting because it takes a more holistic approach to managing the process of performance and considers not only the typical infrastructure metrics, but other factors that have a direct impact on the ability of the company to deliver service and manage the business process. The first step in delivering service for many IT managers is in clearly defining the monitoring environment and in implementing best of class tools to avoid downtime, slow response time, and other problems. PLATINUM’s ProVision suite of tools provides a valuable approach because it actually integrates the tools to make them inter-operate, automate repetitive processes, and take corrective actions based on correlated events that occur in the environment.

Hurwitz Group defines service level management as "a process that includes establishing service level objectives and expectations, creating service level agreements (SLA’s), monitoring service levels, collecting and analyzing data, and establishing user feedback and escalation procedures". ⁴

The renewed focus on service levels has been driven by a number of issues. First, key industry analysts all agree that there is a renewed interest in service level management because of the increasingly competitive environment. Many companies are looking toward IT to establish measurable service levels with the user community not only to identify which key services are delivered to users, but also to measure and document how well IT organizations are delivering the demanded services requested by the user community. In fact, many user communities are demanding that SLM measures be written in contracts to guarantee systems availability, reliability and response time. Financial incentives or disincentives are even used occasionally to encourage compliance.

Over the last year, a dozen or so service level management tools have entered the marketplace and provided partial solutions targeted at market niches. From the applications perspective, companies like Micromuse, InfoVista, NextPoint Networks, Avesta, Quallaby, Concord, HP, and others have developed tools that typically access MIB , logfile, and other information sources to quantify and evaluate some of the common SLM measurements. The reason SLM continues to be an effective and necessary strategy for many companies is because it is built on the concept that IT Infrastructure exists to support the business users. A well orchestrated SLM implementation can yield significant financial and operational benefits because these measurements are typically put in place for mission critical business processes – areas where improvements can have the highest financial impact.

Hurwitz Group, in their research report on the Evolution of SLAs identified that "Service is fast becoming a competitive differentiator in many industries. Because of this, many companies are turning to service level agreements (SLAs) that guarantee that a service provider will provide a certain level of service to its customers."⁵ Highly distributed, complex client server environments have increased the need for some measurement standards to ensure availability, reliability, and high levels of performance. Users, demanding a yardstick from IT, have led the way in establishing contracts with their data centers to ensure they effectively and proactively manage service delivery. In addition, users are demanding and specifying how quickly problems must be resolved and standardizing a process to that end. This has led to the establishment of Service Level Agreements. SLAs are typically written contracts between IT managers and users specifying specific measurable performance in areas like application availability, response time, reliability, application and process health, transaction SQL analysis, trouble ticketing, and others.

One of the challenges in writing a meaningful SLA is ensuring that the correct performance statistics are selected and that they clearly provide meaningful and measurable information. What’s the process for creating and maintaining a service level agreement? Below is a recommended seven-step iterative process for developing, maintaining and measuring service level agreements.