The performance and cost- effectiveness of server consolidation are driving IT personnel to consolidate their databases, Web application servers, and mission-critical applications to the Linux platform. Businesses need reliable hardware and software to ensure the availability and performance of their applications.

Successful implementation of Linux applications in server consolidation efforts requires a partner that takes a holistic life-cycle perspective on migrating to Linux and manages Linux applications within a heterogeneous enterprise environment.

Linux server consolidation impacts three distinct areas:

• Server consolidation planning: IT personnel must be able to extract and evaluate the performance, throughput, and responsiveness of existing workloads to accurately determine which hardware configuration requirements will be needed on Linux.
• Coordination of application software deployment: Long before applications are deployed on Linux, installation of software can be a major issue. Businesses must be able to plan seamless software installations, upgrades, and maintenance over hundreds or even thousands of Linux servers.
• Service management: Once applications are ported to Linux, IT personnel must be able to measure and assure ongoing return on investment and end-to-end response times.

Server Consolidation Planning
Moving non-Linux applications to Linux is not as simple as moving databases and applications. To ensure success, customers must understand the underlying resource requirements of workloads on their existing platforms, determine the effects of the business cycle, analyze the current service levels, and determine the appropriate workloads to migrate and consolidate. Of key importance are the performance and capacity issues that result after migration. Under- or overconfiguring the target platform can have equally undesirable results. Overconfiguring results in overspending on unnecessary hardware, while under-configuring causes systems performance issues impacting availability and service delivery.

Not only is it critical to understand which hardware configuration is necessary on the Linux target, it is equally critical to ensure that the consolidated environment can continue to meet business needs to avoid recurrence of misprovisioning. Some of the critical factors to ensure the success of Linux server consolidation plans are listed here:

• Complete understanding of the resource requirements of application workloads considered for migration and consolidation (candidate workloads)
• Complete understanding of the business cycle impact on underlying candidate workload resource utilization
• Ability to evaluate multiple alternatives to lower IT costs per transaction with associated impact (positive or negative) on service levels
• Assurance that ongoing performance of migrated and consolidated workloads will continue to meet service levels
• Visibility for all stakeholders who need to see the performance of their workloads throughout the server consolidation process and ongoing thereafter

Determining if, when, and what additional resources are needed to assure ongoing responsiveness and throughput is paramount. Applications and hardware that scale inadequately will affect business availability. IT personnel must always be aware of the hardware resources and applications in use on their systems. It's imperative that these personnel have the ability to closely forecast the future in terms of the responsiveness and throughput of business applications. It's not enough to simply trend underlying resource utilization because response time and throughput do not bear a direct relationship to underlying resource usage (such as CPU utilization). In fact, use of such trending typically leads to significant undersizing with concomitant negative impact on business availability post-consolidation and migration. A typical response to the inaccuracies of trending usually leads to significant hardware resource overprovisioning, thereby lowering the ROI of any migration and/or consolidation effort.

In many cases, the existing IT environment may have sufficient underlying resource capacity, but be inefficiently used or suffer from response time and throughput bottlenecks. Modeling capabilities are critical in these environments, offering rapid identification of transaction responsiveness and throughput bottlenecks in CPU, I/O, network, and cross-system interdependencies. These modeling capabilities can further build on existing analysis and reporting components to provide information and rapid insight into all the critical planning and analysis activities listed here:

• Eliminating overspending on unnecessary hardware resources
• Allowing deferral of hardware resource purchase by accurately projecting when upgrades are required in terms of impact on responsiveness and throughput
• Allowing easy, quick evaluation of lower cost alternatives (hardware, load balancing, tuning, etc.)
• Justifying and scheduling system upgrades by creating actual versus planned graphs and reports demonstrating need
• Ensuring ongoing cost-effective delivery of the right level of service to users and customers - with visible proof
• Providing "what if?" scenarios to better determine how workloads and systems will perform under various resource, tuning, and optimization scenarios
• Ensuring adequate capacity and required levels of service on an ongoing basis

Businesses are not only consolidating from variations of Unix, they are also consolidating Windows environments and migrating to Linux. These businesses require the same types of performance viewing, management, analysis, and planning solutions to help facilitate the ongoing performance optimization and consolidation and/or migration activities.

Coordination of Application Server Deployment
Many businesses are exploiting the cost-effectiveness, stability, and scalability of running applications on Linux. While offering obvious technical and economic benefits, deploying applications to Linux also introduces some unique deployment, system management, security, and reliability challenges. The logistical challenge associated with installing, upgrading, and managing an environment with hundreds (or even thousands) of Linux systems running diverse applications must be addressed efficiently and effectively. Make sure that your solution addresses the following business needs:

• Installation and software deployment to Linux
• Deployment of heterogeneous applications over hundreds or thousands of Linux nodes
• Reduction of scheduled maintenance downtime
• Reduction of operational costs of a decentralized Linux environment
• Controlled interoperability and application software prerequisites when software is upgraded
• Improved productivity of the IT staff that is responsible for application installations and upgrades

Open Configurations
Disparate combinations of the Linux operating system and applications running on Intel or zSeries will cause application implementation issues. Different deployment configurations and permissible methods can provide a flexible and customized environment, but they also make a single, straightforward, standardized, general procedure impractical. The concept behind flexibility is customization - making the system behave the way you want it to, not the way the engineers imagined you'd want it. You need to turn this flexibility into controlled power, not chaos. You need a way to perform specific and simultaneous installations and maintenance tasks on multiple machines, each configured in a way that best fits your business needs.

Security
Once you have solved the issues of variety, flexibility, and power, the issue of security remains, and if this issue is compromised, the rest of the solution is worthless. Security for the components within the Linux operating system is dependent upon the expertise of the personnel or the abilities of the management tool you use. An effective tool should find components that are security hazards and not only warn you of them, but provide an option to automatically replace them with secure components.

Service Management
Service management is the last, but by no means least important, area of Linux server consolidation. Customers must be able to manage their Linux, Unix, and Windows environments from a service-level perspective. If you are consolidating your servers to Intel or zSeries hardware, you need the tools to manage the process from an application perspective. You need to know how your applications are performing from an end-to-end perspective. In other words, you need the answers to the following questions:

• How did my application perform before I moved to this hardware?
• Is it performing better?
• What is my end-to end response time?

Overall Enterprise Management for Linux
Enterprise Management for Linux (EML) ensures the business availability of Intel and zSeries servers and their applications by solving day-to-day performance problems, tracking long-term performance, and providing an evaluation of existing workloads to assist in server consolidation. EML provides application deployment, end-to-end monitoring, and management for application infrastructure and tools to help measure service levels.