Inexpensive Web Hosting - J2EE Hosting Sun Java, Mysql5, Java, Jsp, Servlet Developing Blog

342 CHAPTER 13 ASSEMBLING A PERFORMANCE MANAGEMENT PLAN Performance Management Process Document A PMP is built around the various phases of the software development and deployment life cycles and at its core is the performance management process document. Specifically, the performance management process and artifacts are grouped into the following sections: Architecture Development QA Production staging Production deployment Production support Capacity planning The architecture section of the performance management process document defines the criteria from which SLAs must be written. For example, it may state that all SLAs must provide an average response time, and that average response time must be true for 95 percent of invocations of the use case. Furthermore, at no time during a use case execution is the response time allowed to deviate greater than 50 percent over the SLA value. As in this example, the architecture section defines what SLAs in various environments must look like without defining each SLA; later, when the SLAs are written by the application technical owner and the application business owner, they know the standards to which they must adhere, so they can set specific values for their use cases and nail down deviation limits. The architecture section of the PMP process document ensures clarity when integrating performance criteria into use cases. The development section of the performance management process document defines the performance testing requirements for each project under development. For example, it may require functional unit tests for each piece of code submitted to the source control system that exercises more than 90 percent of the components. Furthermore, it may require developers to capture performance baseline snapshots and submit a performance snapshot difference for newly submitted code. The performance snapshot difference reveals the differences in response times for the tested code (down to the line-of-code level) as well as object creation differences. The exercise of reviewing this data should be sufficient to help developers ensure the performance of their components. Ideally, the developers performance data should be stored in a performance repository that can perform historical analysis on the developers behalf and trace changes in performance to developer notes about changes in functionality. The QA section of the performance management process document defines the performance testing process that QA must follow. QA procedures exist for functional testing, but they are seldom implemented for performance testing. Therefore, this section defines performance integration tests and performance integration load tests. It also defines how to measure performance and interpret and validate use case performance criteria. It emphasizes the importance of including coverage profiling against load scripts (while not under significant load) to validate that load scripts are truly testing the majority of the application code. The goal of the QA section is to provide QA team members with all of the performance management process information that they need to effectively perform their jobs.
Note: If you are looking for cheap and reliable webhost to host and run your mysql application check mysql web server services.

CHAPTER 13 ASSEMBLING A PERFORMANCE MANAGEMENT PLAN comments and code-level changes. By encapsulating all stages of the application performance life cycle into a single managed infrastructure, performance issues can be quickly identified and resolved, hopefully before your users ever experience them. Performance Management Infrastructure Therefore, we can define a performance management plan (PMP) as a collection of artifacts that define processes to manage the performance of an enterprise application from application inception to production management. The concept of a PMP has been lacking in the Java industry for years, which is probably why so many companies miss their production SLAs, so one of my goals is to promote the adoption of a PMP by my clients. And thus far performance teams in companies that have embraced the concept, either fully or in the part of their organization over which they have control, have found the PMP fruitful. As you read through this chapter, keep in mind that while a PMP ideally addresses your entire organization, in reality its complete adoption is difficult to achieve in a large company. Therefore, extract the parts that are relevant to your group within your company, and you will still achieve significant benefits. The term performance management plan is an abstraction for what might be better termed a performance management infrastructure, because as previously mentioned, rather than being a document, it is really a collection of artifacts. Let s now review the functions of each artifact. The performance process infrastructure includes a performance process document that defines the organization s performance standards across the software development and deployment life cycles. For each application within the organization, it maintains application performance management documents that serve to integrate individual applications into the performance process. While the performance process document focuses on the processes that are performed at each stage of the performance management life cycle, the application performance management documents focus on application-specific implementations of that process, including schedules, artifact locations, and component owners. The performance testing infrastructure specifies the schedules, repositories, and roles for performance testing for each application across the organization. This infrastructure includes performance unit tests, performance integration tests, performance integration load tests, production staging tests, performance staging load tests, and capacity assessments. In addition, it defines the initial loads as well as the load test loads to be executed at each stage for each application. The performance deployment infrastructure specifies the hardware and software topologies for each deployment environment, as well as providing deployment analysis. The purpose is to track the performance of applications historically against deployment configurations. The production support infrastructure defines the individuals in each role in the production support methodology, so during issue triaging, the owners of each component across support levels two, three, and four are unambiguous. Furthermore, the infrastructure allows you to track performance issues to their resolutions for historical analysis leading to the improvement of applications and deployment options. The capacity planning infrastructure defines the schedules and durations for capacity planning activities, including trend analysis, forecasting, and capacity planning. It identifies the individuals or teams responsible for performing capacity planning activities for each application. Finally, it provides a repository for tracking capacity planning artifacts historically, which leads to improved quality of these artifacts across the lifetime of an environment.
Please visit Domain Name Hosting services for high quality webhost to host and run your jsp applications.

340 CHAPTER 13 ASSEMBLING A PERFORMANCE MANAGEMENT PLAN Figure 13-1. The performance management infrastructure. The performance process infrastructure spans the entire application life cycle, the performance testing infrastructure spans the architecture through production staging phases, and the remaining infrastructures address specific niches in the application life cycle. Artifact repositories are maintained at various stages in the application performance life cycle, denoted by the symbols. One of the primary components of the performance management infrastructure is the notion of repositories. Repositories may be simple directory structures or version-controlled software storage applications. The point is that at various stages in the performance management process artifacts will be generated, and those artifacts need to be managed and tracked. In the ideal situation, a software infrastructure would exist to manage these artifacts and allow them to be queried to visualize the performance impact of changes. For example, if a new version of an application is released that negatively affects the capacity assessment, you should be able to trace application code changes back to changes in the results of integration performance tests and, further back, to changes in the results of individual performance unit tests. And ideally, those performance unit test reports and results can lead back to source control
In case you need quality webspace to host and run your web applications, try our personal web hosting services.

CHAPTER 13 ASSEMBLING A PERFORMANCE MANAGEMENT PLAN Table 13-1. Performance Management Artifacts Stage Artifacts Architecture Performance criteria definition process, integration of performance criteria into use cases Development Performance testing process, performance unit test results QA Performance testing process, integration load test results Production staging Performance testing process, production staging load test results, capacity assessment results Production deployment Production deployment process, high-availability and failover requirements, scalability analysis results Production support Production support process, production support workflow, incident tracking Capacity planning Capacity planning process trend analysis, forecasting analysis, and capacity planning analysis As you review this list of artifacts, you will notice some overlap across stages as well as some artifacts unique to various stages. Specifically, the process and performance testing process artifacts appear in multiple stages. Once the application is in production, then you see the introduction of unique artifacts, such as production deployment scalability analyses; production support workflow; and trending, forecasting, and capacity analyses. Therefore, we overlay the following five sections of the performance management infrastructure across the seven stages of the application performance life cycle: Performance process infrastructure Performance testing infrastructure Performance deployment infrastructure Production support infrastructure Capacity planning infrastructure Figure 13-1 illustrates the overlay of the relationship between these five categories of artifacts and the seven performance management stages of an application s life cycle.
Please visit Domain Name Hosting services for high quality webhost to host and run your jsp applications.

338 CHAPTER 13 ASSEMBLING A PERFORMANCE MANAGEMENT PLAN I want to include this information here, so you and your organization can start benefiting from it today! Evolution of the Performance Management Plan Before formally defining a performance management plan, a review all of the activities that performed in the field of performance management may be fruitful in establishing some cohesion among them. Although performance management can begin at any stage of the development life cycle, depending on the current stages of your applications, let us consider the performance management steps from application inception to production management. The following list briefly outlines these steps at each stage of the application: Architecture: Establish performance criteria and integrate those into use cases. Development: Define specific processes to test application components for performance; specifically, run unit tests using a code profiler, memory profiler, and coverage profiler. QA: Require QA to evaluate an application against performance criteria in addition to functional criteria. Furthermore, require QA to perform performance integration tests as well as performance integration load tests. Production staging: Require the performance team to perform production staging performance testing as well as production staging performance load testing, which includes scalability testing. Before passing the application over to production deployment, as well as during the last few iterations of the application development, the performance team includes performance user acceptance testing (UAT) to ensure that the application satisfies performance requirements as well as the traditional functional requirements. Production deployment: Evaluate the results of the performance integration test to optimally tune your production environment. These results include an evaluation of and planning for high-availability and failover requirements. Production support: Define performance monitoring rules and intelligent alerts to detect performance problems before they impact your users. Define processes, so that when performance problems do occur, you can quickly triage them and send them to the most appropriate party to resolve them. Capacity planning: Once an application is running in production, perform trend analyses, build forecasts combining trends with your business domain knowledge, and construct a capacity plan to ensure application performance. A PMP needs to encompass all of these activities into a process that you can manage and track. If these activities remain disparate units of work with their own separate documents, then the process will become difficult to maintain. Rather, they need to be grouped into a larger category of performance documents under the umbrella of a PMP. Table 13-1 lists the important artifacts to capture at each of the seven stages of the application performance life cycle.
We would like to recommend you tested and proved virtual web hosting services, which you will surely find to be of great quality.

CHAPTER 13 Assembling a Performance Management Plan Looks like we re about done. We ve covered everything that I can think of with respect to performance management. Is there anything else you want to share with me? John asked. He and his team had worked very hard to put into place the various methodologies we discussed, and they were back on track for success. Ah, yes, one last important thing, I replied. We ve talked about performance management from the inception of a product through its architecture, development, and testing; constructed high-performance deployment strategies; defined the optimal workflow to troubleshoot production issues; and even looked at trend analysis, forecasting, and capacity planning. But let s tie them all together, so I m not leaving you with a set of disparate activities. You re right, that s quite a list. What do we need to build a coherent story including all of these activities? John asked. Performance management is an evolving practice, and as such, I expect to see some formal standards emerge in the next couple of years, but for now you need three things: process, tracking, and analysis. You need to define what performance activities you are going to perform at every stage of your application s development and deployment life cycles, track artifacts generated at each stage, and then analyze both the performance of your applications as well as the effectiveness of your process. Or, in a single term, you need a PMP. That’s a performance management plan. Sounds reasonable, but how big of a document is this thing going to be? If only it was as easy as building a document, I began. No, rather it is more of an infrastructure. Let me walk you through it . . . Formal performance management and tracking are currently undefined arts. I envision the formal definition of performance management standards and the appearance of performance management infrastructure management software emerging over the next couple of years. In the meantime, let this chapter serve as my definition of the artifacts and processes that must exist, regardless of the names and conventions that they eventually use. I realized the performance management processes explained in this chapter, and then took them to the street, so to speak I introduced many of these performance management processes into Fortune 500 companies and refined them into what is presented in this chapter. The next steps moving forward for the industry will be to establish a standards committee for performance management and formalize its artifacts. But rather than wait years for the standards committee,
Searching for affordable and reliable webhost to host and run your web applications? Go to our java web server services and you will be pleased.

CHAPTER 12 TRENDING, FORECASTING, AND CAPACITY PLANNING Forecast Risk Detailed Analysis While the forecast risk assessment provides summarized information for each forecasted potential risk with a recommended resolution, the forecast risk detailed analysis provides detailed analysis of each risk, including the detailed steps required to resolve the problem. In this example, we would replace this is the result of changing from a file-based basic authentication scheme to an LDAP solution with the LDAP authentication process at the projected load saturates the CPU and causes LDAP threads to wait. And we change a hardware upgrade to the LDAP server to add two additional CPUs to the LDAP server. In this capacity, the recommendation describes the exact problem, so that if approved, the resolution is not ambiguous. Finally, the risk detailed analysis provides the deep analysis from the capacity assessment describing the exact nature of the problem and the degradation models of the binding resources and requests, which serves as justification for each recommendation. Summary Trending, forecasting, and capacity planning may sound similar at first, but the differences between them can be summarized as follows: Trending is the analysis of data with the intent of identifying discernable patterns. Forecasting is the projection of those identified patterns to understand the impact on business processes. Capacity planning is the response to forecasts to ensure the integrity of business processes. We begin by constructing models of the behavior of various components of our environment, including usage patterns, heap usage patterns, resource utilization patterns, and response time patterns. Then we analyze those models against historical data to detect trends. We combine those trends with business domain expertise and corporate insight to forecast the impact of those trends on the environment. Finally, we compare those forecasts against business processes with application business owners to assess the forecasts impact on end users and the costs to resolve any problems. All of this information is summarized in the capacity plan and presented to decision makers within your organization. This chapter began by stating that trending is a science; forecasting is a methodology; and capacity planning is an art. But with the process presented in this chapter, and a little experience, successful capacity planning is very attainable. In the next chapter, we complete the analysis of Java EE performance management in production by assembling a formal performance management plan. The performance management plan assembles the various performance-related activities discussed in this book into a cohesive performance management solution.
From our experience, we can recommend PHP Web Hosting services, if you need affordable webhost to host and run your web application.

334 CHAPTER 12 TRENDING, FORECASTING, AND CAPACITY PLANNING The executive summary provides all key points covered in the report with minimal justification. Any additional justifications for the material in the executive summary that the reader needs can be found in the body of the report. Forecast Risk Assessment This section in the capacity plan presents the results of the forecast risk assessment that was performed earlier in the process. The main questions that you need to answer for the reader are as follows: What are the risks? What are the impacts of the risks? How many users will be affected by the risks? How will the risks affect current users? How will the risks affect future sales? What are the costs to resolve each risk? What are your recommendations? For each forecast that you address, you need to answer each of the aforementioned questions. Establishing a need before proposing a solution and measuring that solution with a cost is important. Consider the example risk assessment presented in the following sidebar. RISK 1: DEGRADED LOGIN FUNCTIONALITY The login functionality of the Acme application is forecasted to begin significantly exceeding SLAs during next month s promotion: once the user load ramps up to current usage, then all subsequent users will be affected, accounting for 20 percent or more of our users. The impacts of this risk are as follows: Support calls and e-mails by existing customers Site abandonment by promotion targets (future customers) This is the result of changing from a file-based basic authentication scheme to an LDAP solution. The cost to resolve this problem is a hardware upgrade to the LDAP server itself, estimated at $10,000, as well as 20 labor hours to implement the new solution. Because of the high-profile nature of this problem, it is recommended that the solution be implemented as soon as possible. As you can see from this example, each of the main questions of the forecast risk assess ment is addressed, but at a surface level. The forecast risk detailed resolution plan section will better clarify the nature of the problem and present the detailed steps required to attain the solution.
In case you need quality webspace to host and run your web applications, try our personal web hosting services.

CHAPTER 12 TRENDING, FORECASTING, AND CAPACITY PLANNING Figure 12-4. As the number of users increases, system resource utilization and request throughput increase. But at the point in which resource utilization becomes saturated, throughput drops and response time increases exponentially. Capacity Plan The capacity plan is the culmination of trending, forecasting, and analysis efforts. It summarizes identified trends, their forecasted impact on business processes, observed behaviors in a production staging environment, the cost to fix them, and your recommendations. The capacity plan is the formal mechanism used to communicate the steps that you propose to mitigate impending risks. It is composed of the following sections: Executive summary Forecast risk assessment Forecast risk resolution plan Executive Summary If the resolutions you suggest are expensive or time consuming, be assured that you will need to present your findings to someone who does not have time to read through all of the details. The information the person needs is fairly simple: What are the potential risks? How many users will be affected by the risks? What is the cost to resolve the risks? What are your recommendations?
Searching for affordable and proven webhost to host and run your servlet applications? Go to Linux Web Hosting services and you will find it.

332 CHAPTER 12 TRENDING, FORECASTING, AND CAPACITY PLANNING Capacity Assessment As discussed in Chapter 9, the purpose of a capacity assessment is to identify the following key data points: The response time of your requests at expected usage The usage when the first request exceeds its SLA The usage when the application reaches its saturation point The degradation model from the first missed SLA to the application s saturation point The process in creating a capacity assessment is to integrate a graduated load generation tool to reproduce end-user behavior at graduated steps until the environment breaks, constructing a degradation model for each request from the point that the first SLA is violated until the envi ronment reaches its saturation point. When performing a capacity assessment as a result of a forecast, we are anticipating a specific user load performing a specific set of functionality. For example, if at a tradeshow presentation, someone reported the availability of a white paper on your Web site to 5,000 people, then you can expect an increased load in the functionality that serves white papers. Previous experience as well as industry reports might reveal that up to 20 percent of the people in attendance download white papers sometime within the week, 5 percent the following week, and then things return to normal thereafter. If this is the case, then the load scripts in the capacity assessment can be crafted to produce the anticipated load, and a capacity assessment can be constructed around that load. This capacity assessment empowers you to execute forecasted scenarios and project their impact on your production environment. In addition the capacity assessment identifies SLA violation points, saturation points, and request and resource degradation models. Therefore in the capacity plan, the impact of the forecast can be quantified and used to justify change requests. For example, if you expect to have an additional 500 people accessing your Web site for a week after a tradeshow, and specifically searching your whitepaper archive, then you can present the scenario, including the observed behavior and degradation model of the environment under the projected load. Supporting your recommendations with hard numbers is much more powerful than with projections. As shown in Figure 12-4, applications and environments do not degrade linearly. Note Because the degradation model is not linear, you have to be very careful when making projections. I once worked for a company that outsourced load testing in order to assess the capacity of its system. The company came back and told us that at 500 users, our response time was well below our SLA requirements, and because the company tested only one of our four machines, we could comfortably satisfy over 2,000 users. As soon as the user load increased in conjunction with a major promotion, application servers started crashing well below 800 users. I seriously question the company s testing strategy, but whatever its strategy, the organization s projections were completely off. We might have been able to support 500 users within our SLAs, but if our CPU was at 95 percent, then we had no hope of supporting an additional 50 users, let alone 1,500 more. Understanding your resource saturation point is paramount to making accurate projections.
If you are in need for chaep and reliable webhost to host your website, our recommendation is http web server services.