It's Time to Rethink Application Performance Management

By ExtraHop Networks CEO Jesse Rothstein

When Raja Mukerji and I founded ExtraHop Networks in 2007, we had a vision for a powerful and elegant approach to application performance management (APM). We realized that legacy APM tools had become untenable due to exponential increases in IT complexity. The ExtraHop Application Delivery Assurance system was based on the premise that the health and performance information needed to manage business-critical applications is available on the wire. Previously, this valuable information was discarded, and due to gains in processing power and storage capacity, we were able to engineer an innovative solution that enabled a truly application-aware, network-based approach to APM.

The ExtraHop system offers enterprise scalability to analyze hundreds of thousands of transactions each second—up to a sustained 10Gbps—across tens of thousands of devices. It provides visibility across tiers, including network, web, database, and storage. Sometimes our customers are even more intrigued by what we don't do—the ExtraHop system imposes no performance overhead, requires zero configuration, does not consume network bandwidth through synthetic transactions or polling, and does not install any host-based instrumentation. The result is a solution with powerful APM capabilities (what we do) without the drawbacks of traditional approaches (what we don't do).

It's time for IT Operations and Application Support teams to reconsider old assumptions about APM. I took a close look at these assumptions in an article published in Network World (and syndicated on CIO.com and Computerworld) this week, "5 Reasons to Avoid Agents for Application Performance Management."

The ExtraHop system provides the health and performance metrics needed to identify and fix application performance issues and supports a broad array of application protocols in real time:

• Web. Real-time HTTP analysis and SSL decryption enable the ExtraHop system to monitor the health of web applications and alert on performance problems. Details include URIs, errors, and transaction processing times.
• Database. Transaction analysis for IBM DB2, Informix, Microsoft SQL Server, MySQL, Oracle, PostgreSQL, and Sybase enables IT teams to monitor database servers and detect specific slow queries.
• Storage. CIFS and NFS analysis reveal access patterns and errors by method, file, or user on the NAS.
• Network. The ExtraHop system provides advanced TCP analysis capable of detecting esoteric TCP performance issues.

At ExtraHop, we've focused on how APM should be instead of repackaging old technologies to solve new challenges. APM products that rely on host-based instrumentation such as performance agents are difficult to deploy and maintain and fail to understand the impact of the network. IT teams should not have to worry about software compatibility issues, consumption of system resources, or perturbing the very applications that are being monitored. There is a better way.

Would you like to find out more? Please give us a call at 1-877-333-9872 or leave your details in our contact form.