Research Report CA APM 9.5 Application Performance Management for Cloud Introduction Clabby Analytics has been following the Application Performance Management (APM) market for several years. In that time, the market and the solutions that serve the market have evolved considerably. Legacy APM tools (such as Wily Introscope acquired by CA Technologies in 2006 and IBM Tivoli Composite Application Manager {ITCAM} for Application Diagnostics) provide very specific root-cause analysis into application performance issues by instrumenting applications and using agents to monitor a wide range of application environments. These tools have loyal customer bases and are still widely used today. Legacy APM tools are rich sources of information but take time to deploy and, in some cases, capture so much information that it is difficult to isolate the source of a performance problem. The following trends have given rise to a new generation of APM tools: 1) Dynamic, composite cloud-based applications Today s applications traverse the virtual and physical as well as public and private clouds, accessing functionality from a variety of sources to piece together the subtransactions and services that make up the application. As a result, predicting application performance is more complex than in the past. 2) Big Data Monitoring tools collect data from literally millions of sources and customers want to find a way to exploit this growing volume of valuable data. 3) Customer demand for new applications and new features to be delivered quickly To keep pace with market demands, businesses must be able to deliver new applications, new features and bug fixes faster without sacrificing reliability. This is difficult for large organizations that have grown to include a wide variety of technologies and processes, as well as a geographically dispersed workforce that has resulted in organizational siloes. 4) SaaS Cloud Delivery Models Customers are looking for solutions that can be purchased as cloud services, replacing a large up-front investment with a pay-as-you-go model, and improving ease-of-use and accelerating time-to-value. 5) Increase in Mobile Applications Smartphone users worldwide will download more than 45 billion apps this year, nearly twice the number of apps that were downloaded in 2011, according to Gartner. An excellent example of an APM solution that has evolved along with the market is CA APM from CA Technologies. In this Research Report, Clabby Analytics will examine the APM market and CA Technologies suite of APM tools, as well as new capabilities in CA APM 9.5 and CA APM-as-a Service. These new features enable CA Technologies to address these market trends and customer challenges with a comprehensive, integrated APM solution that is full-featured, flexible and easy-to-use.
How the APM market has evolved to respond to market trends In order to respond to market dynamics, APM vendors have had to add new capabilities to their solutions or enhance existing solutions. These trends have also led to smaller start-up vendors entering the market. Most vendors are focusing in the following areas: operational analytics, automated integrated solutions to support DevOps, SaaS solutions and easy-touse GUIs. Operational Analytics Today s complex applications must monitor multiple sources to identify and isolate performance issues cloud services, multiple end-point devices as well as the underlying infrastructure including databases, network switches, and servers. Let s use the example of an on-line order and the actions that this transaction will trigger (1) Check availability of the item (2)Verify the customer s payment information (3) Send order confirmation email (4) Set up shipment for the item (5) Bill the customer for the item (6) Send another email with shipment information. In order to report on overall application behavior, APM tools must be capable of aggregating performance data from a variety of sources and using analytics to correlate that data into meaningful information that will enable proactive, predictive management. Some vendors use correlation and statistical engine to self-learn the IT infrastructure, collecting and correlating the data and establishing profiles for what is normal in a given situation and what is abnormal, enabling end-to-end proactive management. The more data that s collected, the smarter these products get. CA s Application Behavior Analytics is an example of this type of APM solution (more on this later). DevOps- Accelerating Application Delivery One of the trends we re seeing in application delivery is around DevOps. A DevOps methodology eliminates operational siloes by improving the communication and collaboration between development and operations, and fully automates the building of development, test and production environments to accelerate application delivery. This simplifies the process of moving software code through the product lifecyle so it can be done quickly and accurately. The goal is to compress cycle times by reducing idle time and wasted time across the lifecycle. By doing this, organizations can reduce a 6-12 month cycle time to potentially weeks or even days. With most development organizations embracing Agile development, more applications are being launched and more changes are being made. DevOps is the latest way for businesses to address new business opportunities with the speedy release of new products and enhancements. DevOps supports what is known as continuous delivery, defined by Wikipedia as software to be developed to a high standard and easily packaged and deployed to test environments, resulting in the ability to rapidly, reliably and repeatedly push out enhancements and bug fixes to customers at low risk and with minimal manual overhead. What does this trend mean for the APM market? It means that APM solutions that must be configured manually will no longer meet the rapid pace of application development. Software developers need feedback on application performance before the application goes into production; APM tools must be designed to work throughout the product lifecycle. As May 2013 2013 Clabby Analytics Page 2
a result, they must be easy to use and maintain, and able to identify and diagnose problems quickly so they can be fixed without impacting time-to-market. APM as a service With the growth of the overall software-as-a-service (SaaS) market (Gartner reports that SaaS market will reach $22.1 billion by 2015), many customers large and small are looking for APM tools that can be consumed via a SaaS model. Security issues have held some customers back (in financial services for example) over concerns that customer account numbers and other sensitive information could be compromised. However, for most large enterprises, APM-as- a-service is a viable option at least for some subset of applications and encryption and other security features allay most fears about security. With APM-as-aservice, customers don t have to install and maintain the APM application, they are able to monitor a wide-range of applications across both cloud and non-cloud environments, and it can be managed from any web-based console throughout the organization. Easy-to-use web-based Dashboards and support for Mobile Devices Easy-to-use web-based GUIs are the norm today in both APM-as-a-service solutions as well as on-premise. Many legacy APM solutions have been updated with application oriented dashboards with in-context drill-down, role-based customization, and bundled reports and templates as well as support for mobile devices and mobile applications. CA APM A Closer Look Clabby Analytics published a report in April 2011, CA Technologies in Application Performance Management http://www.clabbyanalytics.com/uploads/caapmsysviewfinal.pdf. In that report, we highlighted CA APM s ability to monitor application performance and provide deep-dive root cause analysis, as well as view and trace all transactions across both mainframe and distributed environments. The CA APM Solution also includes the CAAPM Cloud Monitor which provides end-to-end transaction response time visibility into cloud, mobile and web applications. It monitors synthetic transactions across the hybrid environment to accelerate problem diagnosis. It is available in an on-premise or SaaS version. CA Technologies has a new solution, CA APM-as-a-Service, that is full-featured enterprise-scale, and capable of monitoring billions of transactions in the datacenter, public and private clouds, and hybrid environments. Like its on-premise version, it includes enduser experience monitoring, deep application diagnostics and web performance monitoring. In this Research Report, we will focus on three features available in CA APM 9.5: Application Behavior Analytics, a new web-based user interface, and Browser Response Time Monitor (BRTM). Application Behavior Analytics After a demo of this feature, I was really impressed with how easy it is to use and how quickly problems could be identified and diagnosed. In a few mouse clicks root cause was identified, and it was also easy to see the impact of the root cause on various aspects of the application. Application Behavior Analytics is included in CA APM 9.5, and requires that another analysis server (can be virtual or dedicated) is installed. Application Behavior Analytics has an auto-configurable state-of-the-art learning engine that uses May 2013 2013 Clabby Analytics Page 3
multivariate techniques to analyze all raw metrics data that is collected and stored in the APM Smartstor to find anomalous application behavior across multiple application components and even across multiple applications. The data is correlated and analyzed over time to understand what is normal and what is abnormal in a specific situation. According to CA, it can correlate up to 100,000 metrics every 15 seconds. Application Behavior Analytics is self-learning and goes beyond traditional threshold violation, and runs without user-defined policies or rules. For example, if two metrics typically have a positive correlation and suddenly show that they are inversely correlated, this will be flagged as abnormal. The engine is constantly analyzing time-stamped data and updating its understanding of what is normal. With the statistical learning engine, you never have to delete learning or data and start over the engine is constantly adapting and relearning. It can alert to a first-ever event or report collectively on a pattern change. Every anomaly is given a rating of from 1-100, and anomalies are ranked by these ratings. If a change in the environment is made, it will be detected, and obvious changes will be reported in less than an hour. Over time, this anomaly score will go down as the environment learns the new normal typically within 2-3 days. This capability frees IT resources from day-to-day management allow a focus on revenue-generating activities as well as improve quality of service and end-user experience by reducing mean time to repair and identifying and solving problems proactively. When an alert is received on the APM WebView homepage, an Analysis Workbench tab (shown below) displays metrics from multiple sources (agents, passive, third party) analyzed together, and examined for unusual interactions or patterns. The tab also provides contextual linkage to existing APM visualization. (see Figure 1, below). Figure 1 Analysis Workbench Source: CA Technologies May 2013 2013 Clabby Analytics Page 4
New Web-based User Interface According to CA Technologies, The APM WebView Homepage is targeted to the triager and diagnoser, with features such as charts and graphs, bookmarking, alerts and a UI performance monitor. It provides an intuitive user experience via cross-browser HTML5 and JavaScript-based user interfaces, making it easier to monitor, triage and diagnose application performance problems from many popular Web browsers. The new layout makes it easier to display alerts to IT operators, allowing them to quickly see potential problems across the entire application environment. (see Figure 2, below) Figure 2 APM WebView HomePage Source: CA Technologies Browser Response Time Monitor (BRTM) BRTM is a feature that complements the capabilities of Customer Experience Manager (packaged in CA APM) and is included in the 9.5 release. While Customer Experience Manager is integrated with APM deep dive diagnostics and transaction tracing for a deeper understanding of application issues, BRTM leverages CA APM agents to summarize transaction performance (through out-of-the-box dashboards) at the browser tier, monitoring transaction response time down to the last mile across browser, network, and application. BRTM gives visibility into how the Web browser renders web pages and how much time it time it takes, providing: Visibility into the end-user experience that starts from the browser Visibility into the browser side (client) activity of the transaction Visibility into the quality of service provided by a content delivery network (CDN) This feature helps administrators see the entire user transaction to identify slow-running components with the goal of improving the end-user experience. May 2013 2013 Clabby Analytics Page 5
CA APM in a Broader Context One of the things that sets apart CA APM is that it is fully integrated with CA Technologies other Service Assurance solutions. The following products work in conjunction with CA APM. (see Figure 3, below). Figure 3 CA APM in the Big Picture Source:CA Technologies 2013 CA Executive Insight for Service Assurance is designed for the mobile platform, connecting executives to real-time views of how the health and performance of key IT services will impact business. It includes a new Web services query tool that gathers social media data, enabling executives to see how users feel about a particular application including any customer issues or complaints as well as positive feedback which may be airing via social media. CA Service Operations Insight is designed to correlate and analyze information from your infrastructure, application performance and other IT management tools in real time to accurately visualize services, calculate service quality and to pinpoint what impacts quality and what puts it at risk. CA Catalyst delivers the Unified Service Model (USM) that makes IT management products from CA Technologies and other vendors easier to integrate by exposing a consistent programming model and a common data model. It also provides a set of services to standardize the implementation of various forms of integration-related functionality. CA LISA will be enhanced with technology from the recently acquired Nolio, a leader in continuous application delivery, which will beef up their DevOps offerings by extending the capabilities of CA LISA to include CA Release Automation. CA Lisa Release Automation improves the ability to move application software efficiently and reliably through the development process and into production. May 2013 2013 Clabby Analytics Page 6
CA APM 9.5- Application Performance Management for Cloud Customer Feedback At the recent CA World event, I met with one CA APM customer and attended a break-out session hosted by another. I was interested to learn that CA APM s Introscope, which has been described by some as complex to instrument, install and manage, was viewed as none of those things by either of these customers. Lance Neal, Global Program Manager at Lexmark said that, in his opinion, the work required to instrument applications was minimal for the depth of Java-specific root-cause application information provided, and the scalability CA APM provided to support their 170 global sites and 12000 employees. Not only that, Neal felt that CA was well equipped to support their 2013 IT goals of (1) 1 st customer (of their own solution to improve feedback to R&D) (2) Evolution to analytics (3) Nimble IT (accelerated delivery/rapid scalability) and (4) Bring your own device (BYOD), based on CA s strategic emphasis on Dev/Ops, mobility, and analytics. Lexmark is not only a customer of CA APM, but of other products in the Service Assurance suite as well, and Neal spoke highly of the integration between products all visible through a single pane of glass connected through Service Operations Insight. Summary Observations Having the opportunity to speak with both customers and CA Technologies product managers at the CA World event in May 2013 provided several new insights for me. First, I discovered that CA APM, with its Introscope roots, has many loyal customers who have been using the product for years with no concerns about its complexity. For them, the deep dive diagnostics and vast scalability far outweigh perceived disadvantages of a longer setup and deployment time. Second, CA Technologies is enhancing CA APM with dynamic instrumentation in CA APM 9.5 that will automate many of the functions for a quicker time- to-value and more flexibility to support a DevOps methodology for faster application delivery. Third, in a demo of Application Behavior Analytics, I saw self-learning analytics collect and correlate many Big Data sources and look at metrics in the context of one another automating problem diagnosis and improving the productivity of IT staff. The new GUI is intuitive, has in-context drill-down and is consistent across products. CA APM-as-a-Service will make these capabilities available in a new pay-as-you-go model which will enable CA Technologies to address customer opportunities that start small but have the potential to grow quickly, and need the instant scalability that a SaaS solution provides. To me, with CA APM 9.5, CA Technologies has clearly covered all the bases enabling existing customers to support new growth initiatives in the area of mobile and analytics, for example. By the same token, CA Technologies has overcome many of the objections that may have caused potential customers to consider smaller start-up vendors that offered APM-as-a-service, or a specific expertise in analytics, transaction tracking or customer experience monitoring. For businesses looking for an APM solution that can support their needs now and well into the future, CA APM 9.5 deserves a close look. Clabby Analytics http://www.clabbyanalytics.com Telephone: 001 (207) 846-6662 2013 Clabby Analytics All rights reserved May, 2013 Clabby Analytics is an independent technology research and analysis organization. Unlike many other research firms, we advocate certain positions and encourage our readers to find counter opinions then balance both points-of-view in order to decide on a course of action. Other research and analysis conducted by Clabby Analytics can be found at: www.clabbyanalytics.com.