Revue des articles sur les méthodologies de pilotage de projets et de la gouvernance des systèmes d'information.
techniques de gestion de projets IT et informatiques,
stratégie et tactiques pour les SI.
The IT Watcher.
Magic Quadrant for Application Performance Monitoring
A lire sur: http://www.gartner.com/technology/reprints.do?id=1-1ODYDEA&ct=131219&st=sb
18 December 2013 ID:G00251250
Analyst(s): Jonah Kowall, Will Cappelli
Application performance monitoring is a key focus of IT operations groups, with growing influence from development, quality assurance and line-of-business application owners. Delivery of APM via SaaS and increased analytical capabilities are changing the pricing and competitive market dynamics.
This document was revised on 24 January 2014. The document you are viewing is the corrected version. For more information, see the Corrections page on gartner.com.
The analysis contained in this report is accurate as of 15 July 2013. Due to the dynamic nature of the market, please use standard inquiry for the most updated information on vendors and products.
Gartner defines an application as a set of steps providing specific results with the following characteristics:
The end user or program initiates a request, which triggers the execution of the software and hardware components used to respond to the request.
Some of the steps in the execution are defined and sequenced by business logic, as opposed to computer system logic.
The software algorithms work with one another as they execute, which results in a set of data.
The resultant data is delivered by using hardware and software to the user in a computer interface.
If the algorithms complete their execution successfully, then they achieve well-defined goals that meet the established requirements of some end users or end-user communities.
Application performance monitoring (APM), in turn, is defined as a process with five objectives:
Tracking, in real time, the execution of the software algorithms that constitute an application.
Measuring and reporting on the finite hardware and software resources that are allocated to be consumed as the algorithms execute.
Determining whether the application executes successfully according to the application owner.
Recording the latencies associated with the execution step sequences.
Determining why an application fails to execute successfully, or why resource consumption and latency levels depart from expectations.
To monitor these objectives, five functional dimensions are required:
End-user experience monitoring — The capture of data about how end-to-end application availability, latency, execution correctness and quality appear to the end user.
Application topology discovery and visualization — The discovery of the software and hardware infrastructure components involved in application execution, and the array of possible paths across which these components communicate to deliver the application.
User-defined transaction profiling — The tracing of user-grouped events, which comprise a transaction as they occur within the application as they interact with components discovered in the second dimension (application topology discovery and visualization); this is generated in response to a user's request to the application.
Application component deep dive — The fine-grained monitoring of resources consumed by, and events occurring within, the components discovered in the application topology discovery and visualization dimension. This may include server-side components and client-side devices and interfaces.
IT operations analytics — The combination or usage of techniques, including complex operations event processing, statistical pattern discovery and recognition, unstructured text indexing, search and inference, topological analysis, and multidimensional database search and analysis to discover meaningful and actionable patterns in the typically large datasets generated by the first four dimensions of APM.
The APM market is the market for all the technologies and services that deliver these five dimensions of functionality.
It is important to note that the audience for this Magic Quadrant is IT operations management. It is not developers or end users. Thus, products that are targeted at these other groups were not considered.
AppDynamics acquired DBTuna in 2013 and launched AppDynamics for databases along with run book automation capabilities to take actions based on application monitoring data. AppDynamics has entered the mobile APM market, an emerging focus area for a growing list of APM vendors. AppDynamics Pro Edition 3.7 is the primary product offering, with a freemium (see Note 1) version AppDynamics Lite Edition 2.0 providing value without cost to application development and application support engineers. In addition to offering these products on-premises, AppDynamics offers the identical product via SaaS delivery, and sees a greater shift toward the acceptance of cloud deployment models. With a hybrid hosting approach, AppDynamics introduced APM-in-a-box for managed service providers who want to run their own APM-as-a-service offering. We expect to see continued innovation from AppDynamics in end-user experience monitoring, the introduction of mobile monitoring, PHP deep-dive support, and continued extension of performance monitoring across the application environment.
Leading with a deep-dive and distributed transaction tracing capability, AppDynamics provides support organizations with an analytics-driven root cause analysis capability further augmented with run book automation for corrective actions to ensure high availability.
Expansion into deeper end-user experience monitoring and database performance monitoring has enhanced the portfolio.
The UI has been simplified and targeted, correcting visual complexity within prior offerings.
High growth is fueled by large venture investments, which add pressure to execution. Sustaining high growth rates is resulting in increased offering diversification, which can be a detriment if investments are not aligned with buyers' needs.
A single collection point requiring multiple agent technologies adds complexity to deployments, and limits the extension of the technology to solve problems within diverse application environments.
Self-service SaaS and management was released in July 2013, later than many competitors, allowing those vendors to catch up and even surpass AppDynamics execution.
AppNeta is a newcomer to this Magic Quadrant, having acquired Tracelytics in June 2012, broadening its SaaS-based network performance offerings to include APM. Its solution is delivered only via SaaS, with a choice of physical or virtual appliances to measure and collect network performance and health information. The offering consists of TraceView v.2.1 for deep-dive application analysis and transaction tracing. AppView Web v.2.3 provides single and multistep synthetic monitoring from local or remote locations. FlowView v.2.1 is used to perform passive network performance analysis via deep packet inspection and to analyze flows provided by other network devices, such as NetFlow, sFlow and JFlow. PathView v.6.5 uses patented technology to measure path capacity on LAN and WAN links. TraceView provides script-injected real user monitoring, as well as performing deep-dive support for Java, PHP, Python and Ruby on non-Windows environments, but looks to fill out alternate platform and language support this year. Tracing in depth between many application languages is differentiated from many SaaS providers. AppNeta has brought together the synthetic transaction capabilities of AppView Web with the TraceView Real User Monitoring (RUM) technologies with a single UI. It has done similar integration with PathView and TraceView modules to overlay these data sources. For example, it has integrated data sources such as overlaying quality of connectivity from the network with application transaction path information. Although the TraceView product has not been fully integrated, initial integrations have been delivered, and additional changes are planned through 2013, which will result ultimately in a singular product UI for these features.
AppNeta's SaaS-only platform provides in-depth cross-language tracing and attractive pricing, appealing to those running modern Web application environments or the lower-cost network performance monitoring (NPM) functions needed by those running large WANs.
Unique visualizations and a modern UI provide additional insight into application usage and performance.
Combining NPM and APM technologies appeals across support groups, although integration is not yet complete.
Support of Java technologies is newer and more limited than those from the competition. Tracelytics focused on emerging languages initially; hence, the TraceView functionality has to catch up to other offerings.
Lack of support for Windows or Microsoft's popular .NET language will limit adoption.
AppNeta lacks the ability to automate root cause analysis by applying analytics for faster root cause analysis.
This year, BMC Software submitted its End User Experience Management (EUEM) product set to satisfy requirements for end-user experience monitoring; Atrium Discovery and Dependency Mapping in conjunction with its configuration management database (CMDB) capabilities to support runtime application architecture discovery and modeling; BMC Middleware Management (BMM) to deliver user-defined transaction profiling; Application Diagnostics for deep-dive, agent-based monitoring of Java and .NET; and BMC ProactiveNet Performance Management (BPPM) to support the IT operation analytics dimension of APM. The most notable changes in BMC's APM portfolio were the introduction of a new, in-house-developed deep-dive application component monitoring platform, enabling BMC to curtail its reliance on AppDynamics' deep-dive monitoring technology, and a significant expansion of the APM SaaS offering. While BMC has strength in the EUEM and centralized analytics components, it lacks depth in the deep-dive product due to the young age of the solution.
The BPPM (formerly ProactiveNet) statistical pattern discovery engine is one of the most powerful on the market, assisting with event correlation and root cause analysis. BMC has successfully exploited this engine to draw out the meaning inherent in non-end-user experience-related APM data.
BMC's mainframe application and middleware performance monitoring capability gives its APM portfolio good insight into the complexities of the z/OS world.
BMC's deep-dive monitoring capability is still comparatively immature and does not, as yet, match the functionality that BMC was able to deliver previously via its now defunct relationship with AppDynamics.
BMC's APM product portfolio is comparatively unintegrated and, as a whole, difficult to implement and maintain. Integration has occurred between the EUEM modules and the Java and .NET modules, but other components remain isolated, detracting from the single-vendor story.
The ongoing privatization of BMC raises questions about the road map for current and future BMC products, including BMC's APM offerings.
CA Technologies submitted Application Performance Management 9.5 for distributed environment end-user experience monitoring, application runtime architecture discovery and modeling, and deep-dive application component monitoring. The same technology is available for mainframe environments delivered by Cross-Enterprise Application Performance Management 3.0. Application Delivery Analysis 9.3 as the source of user-defined transaction profiling functionality, while ITOA is delivered through functionality dispersed across products already mentioned and Executive Insight for Service Assurance. Application Performance Management-as-a-service 1.1 is the primary platform through which CA delivered APM SaaS functionality, while a synthetic transaction-based end-user experience monitoring as a service capability is delivered through the company's APM Cloud Monitor.
CA's APM portfolio effectively integrates the deep-dive monitoring of many application domains, including WebSphere MQ and other z/OS environments.
A shared service and transaction model enables CA customers that have APM and infrastructure management products to correlate infrastructure and application performance data.
CA's packet capture-based end-user experience monitoring technology is tightly coupled with its application-aware NPM, supporting effective synergies between the two disciplines.
CA's offering is complex, especially the deep-dive monitoring, consuming more resources than many products on the hosts where it's deployed. The solution is often difficult to implement and maintain.
Data gathered from CA's application topology visualization technology is not well unified with the results of its deep-dive monitoring capability.
For three years, APM played a small role in CA Technologies' positioning. This has caused many CA customers and prospects to question the company's commitment to staying on the leading edge of innovation in this market. The positioning has changed, and APM now plays a prominent role, with a greater level of investment; but there is still much lost ground to be recouped.
The dynaTrace PurePath technology has been extended across a broad array of domains, including z/OS and multiple end-user environments. At the same time, particularly in application server environments, dynaTrace PurePath constructs can be used as a pivot from which to zoom into granular monitoring of method-level events. This integrates user-defined transaction profiling and deep-dive monitoring.
The dynaTrace platform's historical appeal to the developer community makes the Compuware portfolio popular among organizations pursuing tighter collaboration between development and IT operations.
The Compuware portfolio is complex and has undergone a number of fundamental restructurings during the last several years. The company has lacked consistent delivery methods and product integration.
While Compuware is committed to APM, it lacks infrastructure monitoring, and broader IT operations management technologies. The lack of a partner ecosystem, and reluctance to partner with other infrastructure and technology providers forces buyers to seek additional relationships.
The importance of mainframe management software to Compuware's overall corporate position and the difficulties that vendors in that market have faced during the past 18 months have raised concerns about the future of Compuware.
The Foglight platform is one of the few unified infrastructure and application monitoring platforms known for its capabilities within database performance monitoring. These modules are among the most powerful available and can be easily integrated into the technology's overall APM functionality.
The Foglight platform is equipped with easy-to-use and powerful application topology visualization that helps users understand not only how the various logical elements of their applications fit together, but also how the applications fit into the virtual fabric and network infrastructure.
Dell's packet-capture-based end-user experience monitoring provides rich analytics and the ability to users to play back user interactions.
Since Dell acquired Quest, communication regarding the way in which the Foglight portfolio is meant to cohere with Dell's overall strategy has been sparse and confusing. The Dell sales channel has also been slow to learn about the product.
Dell has yet to publicly address the emergence of tablets and smartphones as key application access platforms.
There is little awareness, let alone traction, with regard to Dell APM SaaS offering, even among longtime Foglight customers.
HP released APM 360 v.9.22 as a bundle that includes Business Process Monitor v.9.22 and Real User Monitor v.9.22 to cover the functional dimension of end-user experience monitoring, Diagnostics v. 9.21, SiteScope v.11.22 for application component deep-dive monitoring requirements, Service Health Analyzer v.9.22 for IT operations analytics, and Service Level Management v.9.22. Other HP APM products are Discovery and Dependency Mapping Automation v.10.01, as well as Configuration Management System v.10.01, which provides capabilities in the runtime architecture discovery and modeling functional domain. Operations Manager Smart Plug-ins, all v.9.10, are used for additional application instance monitoring. Business Process Insight v.9.22 and TransactionVision v.9.22 were released together for the user-defined transaction profiling functional domain, while Service Health Analyzer v.9.22, ArcSight Logger v.5.30, Service Health Reporter v.9.20 and Service Health Optimizer v.9.20 provide support for IT operations analytics. HP delivered APM SaaS on the BSM SaaS v.9.13 and HP Performance Anywhere platforms. The APM portfolio has seen a number of key developments. Most notable are HP's extensive deployment of analytics functionality across its APM portfolio in a centralized and decentralized architecture, as well as the launch of its modern APM as a service capability, and the articulation of a strategy that places big data and analytics at the heart of HP's APM and other monitoring disciplines.
HP has made strides to tie together its APM technology with its application life cycle management (ALM) portfolio, going further than other providers. This has been apparent in the tie between HP's SaaS products Performance Anywhere and Agile Manager.
Already equipped with the powerful statistical pattern discovery capabilities of Service Health Analyzer, HP has a compelling vision of APM in which analytics and a focus on cloud-enabled environments play a central role. HP has embedded many of these technologies within the Performance Anywhere offering.
HP is able to integrate infrastructure monitoring with APM, particularly via its SiteScope technology, giving users a synoptic picture that unites application and infrastructure behavior. Further movement toward enhancement of the popular SiteScope product is planned.
HP sales and account management have failed to effectively communicate its APM road map and vision to customers and prospects.
While simplification has been a goal of the APM technology, the APM portfolio remains complex. In particular, results derived from end-user experience monitoring technologies, user-defined transaction profiling technologies and deep-dive monitoring technologies are difficult to correlate with one another. The new SaaS offering, Performance Anywhere, remains separate and inconsistent with data contained in the on-premises BSM offering.
Sales execution is often inconsistent globally. Some HP sales representatives do not present the entire portfolio to prospects clearly, and responses to RFPs are insufficiently targeted to specific customer requirements.
IBM developed SmartCloud Application Performance Management v.7.6 to cover the functional dimensions of end-user experience monitoring and user-defined transaction profiling in distributed environments. User-defined transaction profiling is extended into z/OS environments via Omegamon XE v.5.1. The runtime application architecture discovery and modeling dimension was covered by Tivoli Application Dependency and Discovery Management v.7.2.1. IBM Tivoli Composite Application Manager (ITCAM) for Application Diagnostics v.7.1 was released as the application component deep-dive monitoring solution, while the IT operations analytics dimension was represented by SmartCloud Analytics. APM as a service is primarily delivered through IBM Tivoli Live. During the last 12 months, the most notable changes to the technology were a broad infusion of statistical pattern discovery capabilities across the product line, and the introduction of simpler, more modern UIs, while not replacing UIs that were needed regularly. IBM expanded the APM as a service presence, but still lacks a competitive offering. From a positioning perspective, IBM profoundly altered the status of the Tivoli brand, retaining the Tivoli name for a small number of products, but labeling most of the portfolio with the SmartCloud tag.
Particularly for Omegamon XE and ITCAM for Application Diagnostics, IBM effectively exploits its knowledge of z/OS and WebSphere Application Server to deliver superior insights into the behavior of applications running in those complex, but widely deployed, environments.
The algorithms at the core of Tivoli Analytics for Service Performance are among the most effective in the market at discovering causal, as opposed to correlational, relationships among system behaviors and, hence, are particularly successful at short-term event prediction.
IBM's services depth and global breadth can be applied to its application portfolio, and underwrite the company's ability to execute APM on a worldwide scale.
Despite attempts to rationalize and simplify the APM portfolio, new technology deployments and proof-of-concept executions are often unsuccessful.
While IBM APM fully supports deep dive for Java, .NET support is limited to basic Windows Management Instrumentation (WMI) metric collection.
ManageEngine provides a broad offering of management tools, and is a division of Zoho, a growing SaaS provider of business applications. The APM offering consists of the SaaS platform Site24x7 for synthetic monitoring and server monitoring. Applications Manager v.11 is delivered as traditional on-premises software providing support for all five dimensions of APM. Applications Manager v.11 introduced features allowing deep-dive analysis of applications written in .NET, which was a prior limitation of the product that supports Ruby and Java. Applications Manager not only focuses on APM, but also provides server, virtualization and application instance monitoring. The diversity of this single product makes it a good fit for organizations that value implementation simplicity over absolute functional depth.
Products are priced attractively, giving buyers with limited budgets and time to implement APM a viable and alternative option. Applications Manager is part of the IT360 bundle, which makes it easy to purchase a suite of integrated tools.
Functionality is diverse, going beyond the core APM use cases into server and application instance monitoring for availability and health metric collection.
Although ManageEngine targets small or midsize businesses (SMBs) with a light touch (i.e., a Web-based sales process), its solution can scale to support large environments, according to client references.
End-user experience monitoring, one of the most critical dimensions of APM, is limited within the products, providing simulated synthetic transactions only.
Evolution of the products is significantly slower than the competition, putting ManageEngine further behind competitors releasing advanced functionality.
Details collected during application execution are less granular than other solutions, making the product less of a fit than competitors that better target APM buyers' needs.
Microsoft's focus on unified monitoring of Windows-based server technologies hinges upon the larger System Center management platform. System Center includes the Operations Manager component, formerly known as System Center Operations Manager (SCOM), within the offering focusing on the availability and performance monitoring of associated system, network and application components. Microsoft introduced System Center 2012 to the market in April 2012, integrating the once separate AVIcode APM solution into the core product offering. Since the initial release of System Center 2012, Microsoft has enhanced the offering by increasing the prominence of APM technologies in the core product. With the release of Service Pack 1 (SP1) in January 2013, System Center 2012 introduced APM functionality within standard management packs for SharePoint, and added support for Model-View-Controller (MVC) and Windows Communication Foundation (WCF)-based applications. Microsoft has increased the integration with the System Center Global Service Monitor, a synthetic SaaS offering included with System Center licenses, providing cost savings compared with using third-party synthetic monitoring. Microsoft has begun to push the use of APM within development organizations by creating tighter measurement and triage capabilities for development organizations leveraging Visual Studio Team Foundation Server and Visual Studio development environment footprints. Due to Microsoft's close control and product strategy encompassed within the .NET programming language, it is connecting the development and operations teams more tightly than competitive offerings.
Including APM functionality with server and network monitoring in a single product, this solution can provide the core monitoring needs within Microsoft-centric environments in many enterprise agreements. System Center provides deep functionality.
Microsoft is increasingly stretching the operational nature of APM toward the development workflow, enabling a seamless collaborative environment, driven by increased visibility into production issues. Thus, Microsoft has made organizational changes to continue driving APM in new directions.
Increasing the usage of .NET monitoring within management packs, such as SharePoint, creates additional visibility for those using the popular collaboration platform. Use cases for similar needed capabilities exist for deeper monitoring of Microsoft Dynamics AX, Microsoft BizTalk Server, Microsoft Exchange and Microsoft SQL Server.
Previous to System Center 2012, enhancements of Operations Manager took multiple nonconsecutive years. Microsoft has increased release cadence to yearly, which still lags APM industry norms. Microsoft's exception-centric approach is limited and less focused on transactional monitoring compared with more effective APM vendors.
Support of non-.NET platforms is limited, with only a basic metric collection from Java applications via Java Management Extensions.
New Relic pioneered modern SaaS-delivered APM. The offering has evolved from a basic solution into a comprehensive offering increasingly meeting the demands of more sophisticated APM users. New Relic has been ahead of other vendors for providing fully featured mobile APM, a hotly contested market that will continue to undergo growth and innovation. New Relic's multitenant SaaS platform is growing in size and capability yearly, now collecting over 103 billion (as of June 2013) metrics per day (60 billion in 2012) on more than 1 million servers. This collected data represents the potential to help customers better understand application design, usage and performance. New Relic pioneered the light to no-touch sales model in APM, enabling it to capture a large number of customers in a short amount of time. New Relic supports a large number of languages, including the foundational Java and .NET technologies. It also monitors and provides deep-dive insight into PHP, Ruby and Python applications. The broad language support makes New Relic more suitable for smaller, progressive customers who may need less retention or granularity. New Relic provides a good starting point on an APM journey for SMBs and large enterprises, due to a faster time to value.
New Relic's rapid install and ease of use give it an advantage in initial implementations and quicker time to value than other APM solution providers.
Due to the warehouse of data that New Relic collects, there are many benchmarking possibilities. For example, it has begun with the real-user-focused App Speed Index.
Mobile APM is the next wave of innovation. New Relic has a first-mover advantage and ties it directly into the existing data center and browser-centric product for seamless workflow and use cases.
Although the SaaS model is increasing in popularity, some enterprises prefer hybrid deployment models.
While New Relic's granularity is as fine as other products, out-of-the-box reporting takes place through summarizing metrics and high-level transaction traces. New Relic's UI is primarily geared for aggregate metric data. Deeper granularity and drill-down are available, but require extra navigation steps. While this is acceptable for many users, those requiring immediate access to more-granular data may want to consider other solutions.
With high visibility and disruption, New Relic has attracted positive and negative visibility in the media. The latest is an ongoing patent lawsuit that involves New Relic and another vendor, based upon a patent originally awarded to New Relic's CEO and founder while working for Wily Technology, which was acquired by CA Technologies.
OpTier is undergoing a transition from building APM solutions that solve highly complex problems (and, by extension, require a high degree of consulting and customization to tune the product) into a mainstream APM provider. After undergoing management, marketing and technology changes throughout 2012, OpTier found its path had diverged from the market, hence its absence from the 2012 APM Magic Quadrant. It has since refocused on the differentiations it had within its analytics technology, and recently launched a modern SaaS product. This shift was spearheaded by the launch of OpTier Big Data Analytics (BDA) module v.5.5, providing IT operations analytics, and concluded with the release of a SaaS APM offering in April 2013 (OpTier APM v.5.1). These products share a fresh look and feel, and reduce the complexity issues of prior offerings, while remaining transaction-centric. OpTier APM 5.1, delivered on-premises only and was updated, but still is built on the previous interface and technology. We expect these products to come together in a UI shared between the SaaS and on-premises capabilities, which will allow for better focus in terms of development and sales of a unified and modular product offering.
Strong ad hoc analytics technology within the OpTier BDA module allows for ad hoc querying of collected transactional data, enabling discovery of new patterns not detected within typical algorithms. This is especially useful when paired with the component-based context provided by the on-premises APM offering.
OpTier's differentiation in the on-premises APM offering allows for complex business transaction tracking, data and business metric extraction, and tagging for the most demanding use cases.
OpTier has limited customer adoption. Since the BDA and SaaS offerings are new to the market, their success and direction are uncertain as OpTier tries to reinvent itself.
OpTier's SaaS platform is based on established agent technology. The platform has not yet been field tested with high-volume or high-scale situations.
With the complex implementations of the on-premises APM product, customers have expressed the need to re-engage with services when implementing additional application instrumentation, adding often unanticipated cost and complexity to the product life cycle.
Riverbed has solidified itself as a diverse and fully functional performance monitoring business with an impressive amount of expertise and intellectual property.
The FlyScript programmability technology is new technology crossing the Riverbed portfolio, including application delivery controller (ADC), WAN optimization controller (WOC) and other offerings over time. This technology can make programmability and extensibility a differentiation for other Riverbed offerings, driving automation from APM data.
Riverbed's APM technologies are diverse and consist of a layered analytics approach within the products, providing that multiple algorithms in each layer distill data in multiple ways to provide actionable advice.
Products in the portfolio have limited consistent usability or look and feel, making getting up-to-speed time-consuming. Riverbed has invested in UI convergence with APM Xpert Server (dashboards), attempting to unify the offering to that UI. The depth and technical nature of these tools require complex UIs, which will be difficult to rationalize, resulting in a high learning curve for buyers.
Complexity in the Riverbed portfolio is further increased with several products, including network planning, fault monitoring, security and configuration management technology. They detract from the performance-focused goals of the Riverbed Performance Management business unit. Some of these products are not market competitive.
Riverbed's product lines have overlap between Cascade and the AppResponse Xpert modules. Product enhancement, integration and redesign will take time.
We review and adjust our inclusion criteria for Magic Quadrants and MarketScopes as markets change. As a result of these adjustments, the mix of vendors in any Magic Quadrant or MarketScope may change over time. A vendor's appearance in a Magic Quadrant or MarketScope one year and not the next does not necessarily indicate that we have changed our opinion of that vendor. It may be a reflection of a change in the market and, therefore, changed evaluation criteria, or of a change of focus by that vendor.
Vendors were required to meet the following criteria to be considered for the 2013 APM Magic Quadrant. Since 2012, we increased numerical thresholds, and placed more-stringent requirements around the vendor's ability to prove market execution:
The vendor's APM product must include all five dimensions of APM (end-user experience monitoring, application topology discovery and visualization, user-defined transaction profiling, application component deep dive, and IT operations analytics). The deep-dive monitoring capabilities must include Java or .NET, but also may include one or more key application component types (e.g., database, application server). The solution must include user-defined transaction profiling, IT operations analytics technologies applied to text, and metrics collected by the other four dimensions.
The APM product must provide compiled Java or .NET code instrumentation in a production environment.
Customer references must be located in at least three of the following geographic locations: North America, South America, EMEA, the Asia/Pacific region and/or Japan.
The vendor should have at least 50 customers that use its APM product actively in a production environment.
The vendor references must confirm they are monitoring at least 200 production application server instances in a production environment.
Some features of the APM offering must be available via a SaaS delivery model. This offering must be delivered directly from the vendor.
The product must be shipping to end-user clients for production deployment and designated with general availability by June 2013.
Total revenue (including new licenses, updates, maintenance, subscriptions, SaaS, hosting and technical support) must have exceeded $5 million in 2012.
In addition to these criteria, we evaluate the vendor's ability to cross multiple buying centers, as well as target specific verticals. APM tools have always had a strong market presence in the financial services industry, but the Ability to Execute and the ability to target other verticals are evaluated by the reference clients.
While a vendor may meet the inclusion criteria for the APM Magic Quadrant, placement within the finalized Magic Quadrant depends on its scoring in a number of categories. Ratings in these categories are used to determine final placement on the 2013 APM Magic Quadrant. The 2013 evaluation criteria are based on Completeness of Vision and Ability to Execute.
Product/Service: Gartner evaluates the capabilities, quality, usability, integration and feature set of the solution, including the following functions:
Day-to-day maintenance of the product
Ease and management of deploying new APM
Ease of use and richness of functions within the product
Product deployment options and usability
Integration of overall APM-related portfolio or unified APM offering
Overall Viability (Business Unit, Financial, Strategy and Organization): We consider the vendor's company size, market share and financial performance (such as revenue growth and profitability). We also investigate any investments and ownership, and any other data related to the health of the corporate entity. Our analysis reflects the vendor's capability to ensure the continued vitality of its APM offering.
Sales Execution/Pricing: We evaluate the vendor's capability to provide global sales support that aligns with its marketing messages; its market presence in terms of installed base, new customers, and partnerships; and flexibility and pricing within licensing model options, including packaging that is specific to solution portability.
Market Responsiveness and Track Record: We evaluate the execution in delivering and upgrading products consistently, in a timely fashion, and meeting road map timelines. We also evaluate the vendor's agility in terms of meeting new market demands, and how well the vendor receives customer feedback and quickly builds it into the product.
Marketing Execution: This is a measure of brand and mind share through client, reference and channel partner feedback. We evaluate the degree to which customers and partners have positive identification with the product, and whether the vendor has credibility in this market.
Customer Experience: We evaluate the vendor's reputation in the market, based on customers' feedback regarding their experiences working with the vendor, whether they were glad they chose the vendor's product and whether they planned to continue working with the vendor. Additionally, we look at the various ways in which the vendor can be engaged, including social media, message boards and other support avenues.
Market Understanding: This criterion evaluates vendor capabilities against future market requirements. The market requirements map to the market overview discussion and look for the following functionality:
End-user experience monitoring, including real and synthetic availability testing
Runtime application architecture discovery
User-defined transaction profiling
Application component deep dive
IT operations analytics
Marketing Strategy: We evaluate the vendor's capability to deliver a clear and differentiated message that maps to current and future market demands, and, most importantly, the vendor's commitment to the APM market through its website, advertising programs, social media, collaborative message boards, tradeshows, training and positioning statements.
Sales Strategy: We evaluate the vendor's approach to selling APM to multiple buying centers. We also evaluate the vendor's ability to sell in the appropriate distribution channels, including channel sales, inside sales and outside sales.
Offering (Product) Strategy: We evaluate product scalability, usability, functionality and delivery model innovation. We also evaluate the innovation related to delivery of product and services.
Business Model: This is our evaluation of whether the vendor continuously manages a well-balanced business case that demonstrates appropriate funding and alignment of staffing resources to succeed in this market. Delivery methods are also evaluated as business model decisions, including the strength and coherence of on-premises and SaaS solutions.
Vertical/Industry Strategy: We evaluate the targeted approaches in marketing and selling into specific vertical industries. Commonly, APM solutions are bought and targeted toward the financial services, healthcare, retail, manufacturing, media, education, government and technology verticals.
Innovation: This criterion includes product leadership and the ability to deliver APM features and functions that distinguish the vendor from its competitors. Specific considerations include resources available for R&D, and the innovation process
Geographic Strategy: This is our evaluation of the vendor's ability to meet the sales and support requirements of IT organizations worldwide. In this way, we assess the vendor's strategy to penetrate emerging markets.
See Table 2 for detailed weights.
Table 2. Completeness of Vision Evaluation Criteria
Seven aspects characterize vendors that appear in the Leaders quadrant: (1) competitive offerings related to all five dimensions of APM and best-of-breed functionality in two or more of the dimensions; (2) credibility in the monitoring of application domains assembled from heterogeneous sources; (3) effective integration across three or more of the dimensions; (4) the ability to deliver and support APM on a global basis; (5) a consistent track record of innovation; (6) a vision that places APM at the heart of operations and application management; and (7) demonstrated APM SaaS capabilities across multiple functionality dimensions.
Vendors appearing in the Challengers quadrant have competitive offerings in all five dimensions of APM and have often demonstrated their ability to stay ahead of critical market trends like APM SaaS. Their portfolios are, however, restricted with regard to one or more of the following: (1) functional depth in one or more of the dimensions; (2) the environments to which their technologies are applied; or (3) delivery models. These restrictions can keep them from being considered by some large enterprise accounts, despite strong global presence and a long-term strategic commitment to APM.
Three aspects characterize vendors appearing in the Visionaries quadrant: (1) While demonstrating a deep understanding of market requirements, they have been unable to execute consistently on the basis of that understanding; (2) they have demonstrated the ability to maintain the loyalty of the customer base among demanding submarkets, such as financial services and telecommunications; and (3) they have demonstrated aggressive and innovative forays into the IT operations analytics dimension of APM.
The effective deployment of APM technologies and services depends upon the acceptance of four fundamental premises regarding IT operations that run counter to the grain of many traditional practices:
Monitoring (and, indeed, all aspects of IT operations management) must become primarily application-centric, and, within that application centricity, the primary emphasis must be on ensuring a high quality of end-user or customer experience. That is not to say that monitoring infrastructure components (e.g., servers, storage, networks, and the virtual fabric) becomes unimportant. Rather, processes must be reshaped to ensure that the data collected during the course of monitoring such components must always be analyzed for what it reveals about how the behavior of those components impact the end-to-end performance of the applications that exercise those components. Applications must become the windows through which the infrastructure and its performance are observed.
The assumption that system events take place within a relatively static infrastructure that can be effectively described in advance of the monitoring by a topology or graph must be replaced by the idea that most relationships among system components are highly dynamic and frequently change during the course of an application's execution — sometimes in response to that execution and sometimes due to external factors. It would be better to replace the entire notion of a static infrastructure and application topology or graph with the notion of a pace-layered evolving structure, in which any relationships among components are probabilistic.
The focus on capturing instances of predefined, fault-signaling event types must give way to a focus on continuously observing system behavior in the hopes of steering that behavior away from manifestations that negatively impact end users or customers. Indeed, the very concept of an event, in so far as it is retained, should be modified to mean departure from normal system behavior.
The walls traditionally segregating application development from IT operations need to be knocked down, or at least perforated. This movement, known as DevOps, provides the foundational cultural changes needed to accomplish this change in operations, especially with regard to application understanding. The data gathered and analyzed by APM technologies can greatly assist the tasks performed by application developers and testers; that community's expertise will be required to effectively interpret APM data.
Even if these premises are accepted in full, it is still not advisable to try to deploy all five dimensions of APM across the entire application portfolio. Given the cost and complexity of APM technologies, an enterprise should first focus on, at most, the top 10 business-critical applications for those, concentrating initially on end-user experience monitoring. Once that is mastered, other dimensions may be taken up. Finally, once an enterprise becomes familiar with APM, it can contemplate extending end-user experience monitoring and application performance analytics to the top 40 or 50 business-critical applications. It is unlikely that further expansion, either in terms of functional dimension or portfolio coverage, will deliver a positive ROI.
APM has been an implicit part of IT operations since the mid-1990s. Vendors such as Patrol, EcoSystems Software, Mercury Interactive and Candle (acquired, in turn, by BMC Software, Compuware, HP and IBM) and service providers, such as Keynote, provided critical monitoring functionality to enterprises during the period in which classical client/server application architectures began to be set aside in favor of multitier Web-oriented constructs.
Since 2005, there has been an increasingly explicit acknowledgment of, and high-profile focus on, the APM process among global enterprises. Gartner estimates that 20% of the largest global organizations are trying to reconstruct the whole of their IT operational process frameworks to monitor and manage applications, rather than the infrastructure, in a central place. The factor most responsible for the increased attention on the APM process and the tools and services supporting comes from the business side of the enterprise, which, during the past decade, has fundamentally changed its attitude toward IT in general. Line-of-business and C-level executives generally recognize that IT is not just infrastructure that supports background workflows, but also, and more fundamentally, is a direct generator of revenue and a key enabler of strategy.
Due to their heightened appreciation of IT, executives are more willing and have the ability to influence IT operations professionals. Most executives say that what is important about IT is the application portfolio that directly enables them to accomplish their business goals. Unfortunately, just when executives have become enthusiastic about imposing an application-centric view of the world on IT operations, applications have become far more difficult to monitor. In general, architectures have become more modular, redundant, distributed and dynamic, often laying down the particular twists and turns that an execution path could take at the latest possible moment.
The combined impact of modularity, redundancy, distributedness and dynamism has undermined the effectiveness of the technologies and techniques that traditionally supported the APM process. Products such as those offered by Patrol and EcoSystems Software represented minor variations on classical event correlation and analysis and metric collection. A comparatively small number of thresholds indicating unacceptable levels of resource consumption were set in advance. If those thresholds were transgressed during the course of an application's execution, then the transgression would be recorded and sent to a screen or data store.
Because application code was developed, stored and processed in large contiguous blocks, it was, in theory at least, possible to infer the application's overall state of health from a few threshold transgression signals. It was generally assumed that such a sparse set of threshold transgression signals correlated reasonably well with application availability and latency as experienced by end users. If anything further was needed, especially when users accessed applications over a variably performing Internet, monitoring could be supplemented by technologies from vendors such as Candle or service providers such as Keynote, which supported the launch of stereotyped scripts from strategically placed software robots standing in for typical end users.
Modularity, redundancy, distributedness and dynamism, however, combined to ensure that:
Data gathered from the execution of one region of application code provided very little information with regard to what has happening at other times in other regions.
Thresholds could be meaningfully set around the cause versus the effect.
Stereotypical synthetic transactions could not be defined in advance, because user interactions with an application became increasingly varied, and availability and response time characteristics were increasingly sensitive to even small differences in space and time.
A new approach to APM technology was required. Between 2005 and 2008, the outlines of such an approach began to appear, in the way that technology buyers conceptualized the APM problem space, and how vendors were augmenting their product portfolios. The fundamental problem was that applications built according to modern architectural principles needed to be monitored in a holistic, end-to-end manner. Detail remained important, but it had to be set into a well-understood overall picture of system behavior. To that end, five distinct dimensions of, or perspectives on, end-to-end application performance have been assembled by market participants, each one essential and complementary to all the others.
The technologies underlying each of these dimensions are typically deployed by different communities within an enterprise, and the dimensions reflect the concerns of different stakeholders. There is, nonetheless, a high-level, circular workflow that weaves the five dimensions together through five steps. For example:
Step 1. End-user experience monitoring picks up a problem as it impacts the application's user.
Step 2. The application's runtime architecture is generated and/or surveyed to establish the potential scope of the problem.
Step 3. User-defined transactions are examined as they flow across some subset of the possible paths exhibited by the runtime architecture graph, to ascertain what nodes in that graph are the sources of the problem impacting the end user.
Step 4. Deep-dive monitoring of those nodes is carried out in the context of the results of the previous three steps.
Step 5. Analytics are brought to bear — on the one hand, to establish root cause in the midst of the vast volumes of data generated in the first four steps, and, on the other hand, to better anticipate and prepare for end-user experience problems that could emerge.
In 2013, shifts in the APM market are based on execution of differentiations and delivery models. We expect this shift to continue, because the delivery models will enable analytics to better answer and solve user questions in an evolved manner. Many vendors are reworking their APM processes or not launching SaaS services. We have seen vendors not meet the criteria this year, as well as differing levels of execution by large incumbent players. Key differentiators have been simplification not only of the product interfaces and usability, but also of the portfolio, which increases the value of the products. These design and architecture criteria were important factors in this research. The importance of these elements will continue to achieve broader appeal to solve complex problems. By year-end 2013, the global spend for APM software licenses and first-year service contracts will grow to approximately $2.12 billion, which represents a 5.4% growth over the $2.01 billion spent in 2012. This 5.4% growth rate represents a deceleration from the 5.8% growth rate experienced between 2011 and 2012.
This deceleration can be traced to five factors:
APM technology portfolios shifted from being the targets of early adopter investment to mainstream investments. They can no longer defy the gravity of weak economies in the U.S., Europe and Japan. In general, investment in operations management technology has decelerated in response to macroeconomic problems. Because it is now a central component of IT operations management, APM on-premises technology investments will also track the business cycle. Post-2013, this will continue to slow spending on APM.
Most of the APM portfolios offered to global enterprises have been assembled from multiple acquisitions and, as a result, are often poorly integrated, difficult to implement and expensive to administer. This problem afflicts many operations management domains. However, the natural complexity and cross-silo nature of APM make the portfolios seem more of a maintenance problem than portfolios assembled for NPM or virtual fabric performance monitoring. While early adopters tolerated the inelegance of these complex portfolios, mainstream operations management teams are balking at the burdens they impose. By the end of 2014, vendors that remain competitive will deliver radically simplified portfolios, at least by the standards established from 2007 to 2012. Most of the other vendors will not even be considered viable alternatives.
The APM issue that is causing global enterprises the most difficulty is capturing end-user experience and end-to-end latency data when applications are accessed from mobile platforms. Smartphones and tablets are less than completely transparent compared with standard APM technologies. The blind spots multiply when Ajax and HTML5 are thrown into the mix. While established vendors and startups in this space are spending considerable money and effort to develop robust solutions to this problem, no vendor or approach has, to date, been established as a standard. As a result, many potential buyers will postpone a significant investment in a comprehensive APM solution until after a standard is established. By 2014, there will be early attempts to integrate robust monitoring of a complex, eventful process into overall APM.
The scope of APM as a service has broadened considerably. Historically restricted to synthetic transaction-based monitoring of end-user experience, a growing number of service providers can deliver all five functional dimensions. Many IT operations teams are deploying service offerings, rather than making the larger on-premises technology investments, due primarily to reduced complexity and increased flexibility. This price compression is one factor. Another is that the overall pricing of APM technologies continues to decline as these products mature and competition increases. In 2014, APM SaaS will mature, and most enterprises will consider it a viable, and often preferable, delivery model.
While application performance data-specific analytics capabilities have always been considered essential to a comprehensive APM solution, interest has increased in cross-domain IT operations analytics platforms. Vendors such as ExtraHop Networks, Netuitive and Splunk come from different starting points in IT operations analytics, have targeted application performance as a key problem domain and have aggressive sales strategies, combined with ambiguous marketing. Consequently, some potential customers see these vendors as offering an alternative (rather than complementary) approach to the traditional five-dimensional APM. The cost and complexity associated with the traditional approach makes the idea of tackling APM with analytics alone attractive. While IT operations analytics will continue to play an important role in enterprise IT operations management strategies, the majority of organizations will realize that it delivers the most value when deployed as a complement, not as a replacement, to APM and other monitoring disciplines.
As long as macroeconomic conditions do not get worse, the deceleration in APM spending growth will be a temporary phenomenon.
APM as a technology category is closely related to, and frequently confused with, five other technology categories. Unfortunately, vendors have often exploited and compounded the market confusion. The related categories are:
Application management — APM technologies are a proper subset of application management technologies, which include application development, testing, quality and release management technologies, as well as application project portfolio management technologies.
Application-aware network management — Network behavior is a significant contributor to end-to-end application performance, and network administrators need to understand the application context within which network services are executing in order to contribute to the search for service problem root causes and to better prioritize their management activities. This has led to the emergence of a set of technologies that, in many ways, resembles and sometimes overlap with APM. Many enterprises are increasingly expecting such technologies to meet the needs of application operations and network administration. Nonetheless, network administration and application operations needs remain distinct. Even if some technologies (e.g., packet capture appliances) are shared, the ways that the two communities use them are fundamentally different.
Business service management (BSM) — Business services are collections of IT functionality defined and presented in terms intelligible to business users or customers, and governed by SLAs stated and measured in terms that are relevant to business or customer concerns. Most BSM technologies have focused on mapping infrastructure events (i.e., warnings or instances of resource consumption or average latency threshold transgression) to business events defined in terms of business service functionality, and the resultant parallel monitoring of business service and infrastructure events. Thus, BSM is compatible with APM. Although business services that interest users and customers in many large enterprises are applications, we find an increasing number of these enterprises looking at APM as an initial step on a longer road to BSM. Another way of relating BSM to APM is that business services decompose into user-defined transaction types, the execution of instances that are monitored by APM technologies.
Business process monitoring (BPM) — BPM technologies enable users to monitor the execution of a modeled business process flow. Most business processes in large enterprises contain some paths that are executed by applications. It is conceivable for a BPM platform to hand off the monitoring of those application-enabled paths to an APM system, particularly those dimensions associated with end-user experience monitoring and user-defined transaction profiling. In practice, few BPM platform implementations capitalize on that possibility, although numerous APM vendors (e.g., HP, Oracle and Progress Software) have added some BPM functionality to their APM portfolios.
Business transaction management (BTM) — BTM remains a popular term to describe APM offerings that are restricted to user-defined transaction profiling and packet-based real-user experience monitoring. Many users, particularly those in operations management as opposed to those who are application-support-focused, find that conjunction of dimensions to be an attractive starting point for APM; therefore, the term denotes a useful concept. The qualifier "business" is misleading, however, because it assumes that such technologies are, out of the box, concerned with business transactions such as bank account updating or ticket buying. Instead, they are simply concerned with any named set of user activities at an application interface. The qualifier "management" is misleading, because it presupposes that the technologies do more than exhibit how a user activity at an application interface started a series of events to make their way across an application stack and supporting infrastructure.