• SELECT SITE CURRENCY
Select a currency for use throughout the site
IT Performance Optimization
Butler Group, November 2004, Pages: 288
IT Performance Optimization (ITPO) is concerned with extracting optimum performance from IT in the broadest sense, from software applications, to the hardware they run on, through to basic IT infrastructure such as the network and servers, and services provided in support of IT. However, optimization is only one aspect of management, which also includes administering and maintaining control over IT in order to deliver services. In this report we will use the term IT Performance Management (ITPM) where we wish to describe IT performance in the broadest sense, and use the term ITPO when we wish to emphasize the extraction of maximum performance and value from IT.
As a Composite Report, IT Performance Optimization presents material from the whole range of research, conveniently presented in a single volume, which covers investigating different applications, analysis of performance issues and desirable points in obtaining IT Performance Optimization.
Section 1: Management Summary
1.1 Management Summary
Section 2: Business Issues
2.1 Report Structure
2.2 The Problem
2.3 Optimising Existing Resources and Lowering Total Cost of Ownership
2.4 Introducing IT Performance Optimisation
2.5 Aligning IT to Business Goals: ITPO and Service Level Agreements 2.6 Conclusion
Section 3: Technology Features
3.1 Application Performance Optimisation
3.2 Java Management Extensions
3.3 Microsoft Operations Manager 2005
3.4 Application Traffic Management
3.5 Network Services Management
3.6 Database Monitoring, Tuning, and Optimisation
3.7 IT Service Management
3.8 Application Portfolio Management
Section 4: Architectures and Models
4.1 An IT Performance Optimisation Architecture
4.2 A Closed Loop: The Butler Group ITPO Model
4.3 ITPO Methodologies
4.4 Addressing the Needs of Different Size Organisations
4.5 Long-Term Strategic Initiatives by Vendors
Section 5: Market Analysis
5.1 Market Overview
5.2 Market Drivers
5.3 Market Direction
5.4 Case Study - A High Street Retailer
Section 6: Tables
6.1 Butler Group IT Performance Optimisation Features Matrix
6.2 Butler Group IT Performance Optimisation Product Capability Diagrams
6.3 Butler Group IT Performance Optimisation Lifecycle Positions
Section 7: Comparisons
7.1 Product Comparison Methodology
7.2 Vendor Comparisons
Section 8: Technology Audits
BMC – BMC PATROL, CONTROL-M, and SmartDBA
Computer Associates – Unicenter Infrastructure Management Solutions
Compuware – Vantage 9.1, STROBE 3.1, iSTROBE 2.0
HP – Solutions Suite
IBM – Tivoli Monitoring for Transaction Performance v5.3
Mercury – Application Management v5.0
OPNET – IT Guru Product Suite, Version 11.0
Quest – Application Management Solutions
Segue Software – SilkCentral Performance Manager v2.6
VERITAS – i 3 Application Performance Management
Wily Technology – The Wily 5 ™ Solution
Section 9: Vendor Profiles
Section 10: Glossary
ITPM represents a broad field comprising three main areas of interlinked activity: Application Performance Management (APM), Network Services Management (NSM), and IT Service Management (ITSM). APM is concerned with applications in the post-deployment phase, ensuring that they are performing at peak capability, and covering custom applications and enterprise applications, such as databases, Web services, Enterprise Resource Planning (ERP) systems, and integration with legacy systems. The focus on the live environment does not preclude APM from interacting with development and Quality Assurance (QA) - rather there is an important process of closing the loop with the pre-deployment stages, ensuring that knowledge gained in production is fed back to developers to avoid repeating mistakes. APM practice involves any of the following activities: Monitoring and Logging; Application Availability; Transaction Management; Problem Diagnosis using event correlation, root cause analysis, a knowledge base and team collaboration; Modeling and Simulation, including deployment testing and scenario modeling; and Application Portfolio Management.
NSM is concerned with ensuring delivery of the network service, covering network hardware and software infrastructure, including Local Area Network (LAN) and Wide Area Network (WAN) networks, servers, desktops, storage, backup, and associated software. NSM practice involves similar activities to APM except that the focus is on the network infrastructure rather than applications.
ITSM covers Service Support, Service Delivery, Planning to Implement Service Management, and Business Perspective, and is influenced by the UK Government’s IT Infrastructure Library (ITIL) best-practice initiative, which is recognized worldwide.
The emphasis of this report is on the business issue of making applications perform to their maximum capability, and therefore the starting point is APM, with additional coverage of NSM and ITSM. There is a critical need for APM tools in industry today. Many IT departments effectively rely on their users to act as performance monitoring agents, the first time help desk is aware of an issue is if a user calls in to complain. Needless to say this approach has a number of disadvantages: The IT Department ends up operating in a post-incident mode, is never quite on top of problems, and the users also get a poor service as they face frequent downtimes or slow running applications. The lack of historical performance records prevents any kind of objective analysis as to whether performance issues are real or perceived, or for analyzing events leading up to a failure or incident.
There are many different kinds of problems that beset applications, from bugs, to database problems, and infrastructure issues. We examine these causes and show how ITPM tools and best practice can overcome these problems, and in solving them efficiently there are cost savings to be made.
Communication problems between team members increase rapidly as the size of teams grow: For two people there is just one communication path, for five the paths grow to ten, for seven the paths number 21. Surveys have revealed that when application problems arise the average IT department will call upon five people to resolve the issue. Thus with ten communication paths between the five staff members there is a potential for misunderstandings and communication breakdown to occur. This report shows how to improve collaboration and also how to better manage problems so that the right people are involved from the start.
IT Performance Management tools address the above problems, providing visibility into the IT infrastructure for the first time, so that IT support can pre-empt problems through automated monitoring, quickly locate the origin of problems and delegate the right teams to deal with them, as well as improve collaboration between support teams through sharing knowledge in reports and knowledge bases.
There is a direct link between ITPO on the one hand and TCO and Return on Investment (ROI) on the other. ITPO maximizes the use of existing IT investments, ensuring that organizations gain maximum return from their expenditure, and minimizes the expense of downtime and problem resolution. Application users become more productive, and the greater use of automation helps reduce the cost of IT support.
The nature of IT is also changing at a rapid pace, and the direction is towards ever more complex systems. Applications are increasingly Web-enabled, throwing up new challenges in supporting distributed components and Web services, and at the same time the IT department has to maintain legacy systems and enterprise applications. In order to meet these challenges and future-proof organizations against these trends, ITPO can ensure that IT departments can match increased complexity in IT with the support tools to maintain control over it.
The three main areas in ITPO are interlinked but have historically been approached separately. As a result these historical divisions throw up a challenge for any organization that wishes to implement a unified ITPO architecture. The existing solutions will dictate to some extent the next step forward: few organizations start from scratch. Commissioning an audit is a recommended first step; it will reveal exactly what is in place and whether some rationalization can be performed before considering purchasing new tools.
This report outlines the integrated ITPO architecture, linking the various elements of APM, NSM, and ITSM, indicating how as a complete system the full advantages of optimization can be realized.
The business issues addressed in this report include:
Adopting a Performance Culture. ITPM tools will not themselves lead to optimal use of resources, rather there needs to be a policy of adoption by the IT support staff to use the tools and this may require a culture change to current practice within the organization. There also needs to be a feedback process in place so that lessons learnt in production can be passed back to development, ensuring that mistakes are not repeated. The Closed-loop Model shows how the feedback process can be implemented.
ITPO and Service Level Agreements (SLAs). The decision to implement an ITPO solution should be made at a strategic level to realize the full benefit of optimization and control over service performance levels, and is more likely to achieve a good ROI compared with any tactical deployment. ITPO can also help align IT with business, and is an essential complement to the use of SLAs. Surveys on SLAs reveal the lack of service level monitoring – this deficiency to a large extent negates the reason for and benefit of SLAs. Without feedback to the service provider, whether it is an external or internal service, there can be no enforcement or inducement to keep to an agreement. At the very least an organization needs to know whether the standard of service stipulated is being met.
Security and Compliance. IT security is an ever increasing burden on organizations, and IT Performance Management tools can help to identify the signatures of a virus breakout on the network, for example through monitoring traffic levels. Meeting compliance demands is another role these tools can fulfill. Audit reports showing the history of application and network configurations help satisfy regulatory requirements. Compliance demands pro-active measures to ensure the soundness of the IT infrastructure and ITPO helps meet that goal.
IT Department. Performance Management tools bring automation to the IT department and help reduce the pressure on IT staff. With the increase in J2EE deployment in industry there has been no corresponding increase in the availability of J2EE skilled engineers. Tools that maintain J2EE systems are therefore to be welcomed, and also play a part in skills transfer to lesser-experienced J2EE support staff.
The future will see continued growth in Internet use and corresponding reliance on it for critical business activity. To support these growth organizations need to put systems in place that will enable their IT departments to cope with the demand for high performance. Also, with the increasing emphasis today on good governance, businesses will find that ITPO will help to ensure that IT departments have control over IT; APM and NSM provide the transparency into the environment necessary for maintaining control.
There are APM tools that extract maximum performance from a given resource. These tools fall into a number of brackets depending on whether instrumentation is built-in to the application and on how the monitoring agents gather information. The tools chosen should support the technologies in use within the organization, and any likely future technologies. With the need to support Web services on the increase, organizations must check that tools are compliant with the various Web service standards.
A checklist of issues that need to be addressed before selecting a solution needs to also include: Security – given that some Application Traffic Management tools can inspect application messages containing sensitive business information, access to such tools need to be controlled, or encryption used; APM data repository - open schemas allow direct reporting against the database tables; and integration with NSM and ITSM, as many organizations will have existing solutions in place.
The technical issues addressed in this report include:
Monitoring. The crucial frontline technology used in ITPO is monitoring, as this provides the raw data for further analysis. Of particular interest is the emergence of using built-in instrumentation for APM in Java technology through Java Management Extensions (JMX), and in Microsoft environments through Microsoft Operations Manager (MOM) 2005. The alternative to built-in instrumentation is to deploy agents on host machines that track application activity, including availability and performance. In NSM the network management protocols are quite mature.
Database Optimization Databases are complex applications that have been supported with tuning and optimization tools by independent vendors as well as by the database makers themselves. In fact it is more recently that database vendors have started to focus on easier administration, including self-tuning and automated management. A number of vendors have specialized in monitoring the leading databases with agent software and plugging into the rich data stored in the database’s own performance statistics tables and logs.
Tuning. Tuning transactions is possible using end-to-end deployment of APM tools, as these tools identify latencies wherever they occur, in the application, in the network, or in the database. In fact knowing where a bottleneck is occurring is half the battle won – this visibility into how applications are actually behaving in production serves as a key benefit in APM.
Traffic Management. Load balancing and network traffic management are established activities in NSM, whereby processing and message communications are evenly distributed across a network for an optimal use of resources. To this activity can be added Application Traffic Management – allowing traffic packet messages to be inspected for particular applications, and following business rules routing can be prioritized for mission critical applications or other contingencies to be acted upon using a more intelligent traffic management policy. Similarly tools offering Bandwidth Management can allocate network bandwidth to applications and services based on priority. This ensures that bandwidth hungry activities such as Web surfing do not swamp the needs of applications critical to the business.
Legacy Systems. Given that mainframe platforms and enterprise applications on legacy systems are still prevalent, there is a need to consider how Performance Management tools for modern Web-enabled networks can integrate with these older systems. Often legacy platforms will have their own monitoring technology and tools, so the issue is to interface the two systems. A number of APM tools operate on the basis of receiving information from management tools on these platforms but are not capable of deploying agents to monitor events at a granular level.
Diagnosis and Resolution. Diagnosing and resolving problems is a major function of APM and NSM. There are tools that provide the intelligence to correlate events, mine knowledge bases, and help perform root cause analysis. Reporting forms an important part of this process, giving sufficient detail for technicians, mapping applications to infrastructure, and helping identify problems. Reporting to senior management at an aggregated and summary level is also an important function.
Modeling. Network and application modeling and simulation are specialist activities that only a small number of vendors can offer. These systems are suitable for WAN networks, or smaller but complex networks, where a range of decisions can be made without disrupting the live environment, for example concerning deployment changes, security risk assessments, and managing capacity. What-if scenarios can be played out to assist in planning and forecasting.
Application Portfolio Management. This activity aims to ensure that the right applications are being used in the first place, before APM is used. Managing application portfolios bridges the gap between development and deployment, by helping to authorize projects based on need, by analyzing source code for quality issues, and by rationalizing applications to avoid duplication and too much overlap. This also ensures that performance management resources are targeted on the right applications.
The different historical backgrounds of the three main areas covered by ITPO are reflected in the distinct products available in the market. The large IT players have consolidated the Infrastructure Management space with ITSM tools that have wide ranging functionality, but this covers mainly service support and delivery. These products are still distinct from NSM tools, which tend to be specialized and are concerned mainly with hardware and the IP stack. The NSM tools market is also mature and there are many small players who support industry standards, so there is a degree of commoditization in this market. The situation with the APM market is different again: this is a relatively new market so customers are in an early adopter stage. The APM market emerged from development and testing specialists, with the need to monitor and tune applications in increasingly complex, Web-enabled, production environments.
In reality one cannot separate applications from the infrastructure they run within, so it is natural that closer integration between APM and NSM will result. That is what is expected to happen, with products emerging that offer end-to-end performance management functionality for applications and infrastructure. The ITSM players will also continue to build their end-to-end suites, so consolidation in this market space, particularly in adding APM functionality, will take place.
A number of long-term initiatives by major IT players HP, IBM, Microsoft, and Sun, will have an impact on ITPO. The Report investigates these initiatives as they affect ITPO, including utility computing, virtualization, and autonomic, or self-managing, systems. We find that even if vendors’ full long-term ambitions are not achieved the directions indicate what kinds of products will be delivered along the way that touches on performance optimization.
Currently there is no single vendor offering a complete end-to-end portfolio of products that deliver APM, NSM, and ITSM. Despite activity in the market with acquisitions and consolidation, as vendors aim to plug gaps in their portfolios, there is an insufficient degree of integration. Therefore customers will want to be clear about their needs and implement best-of-breed point solutions to achieve an all-round ITPO capability.
F5 Networks, Inc