"Change is the only constant in life."
This is a quote often attributed to the Greek philosopher, Heraclitus. In the world of application performance monitoring, you know this to be true. Things are always changing. New technologies force you to come up with new ways and processes for doing things. And new challenges force you to develop new methods of solving old problems.
Performance monitoring is an old problem.
There's always been the need to monitor application performance, but not always the ability to do it well. As technology has advanced, we've had to evolve how we monitor applications. Initially, performance monitoring largely involved sending ICMP messages to start troubleshooting a down or slow application.
Applications have gotten much more complex, so this is no longer enough. Now we need to know not just whether an application is broken, but why it broke. So APM has had to evolve over the years for us to get there.
But how did this evolution take place, and what happens next? Let's find out.
Phase 1: The Early APM Days
The concept of application performance monitoring (APM) has existed since the early days of the ARPANET. Ever since the 1970s when the first email was sent, there has been a need to know what's happening with applications. But unfortunately, the capability didn't exist. It had to be created.
Much of that work was accomplished in the 1980s. During that time, there were diagnostic protocols and tools developed to help troubleshoot problems on the network. In the toolbox were things like ICMP, SNMP, and tcpdump. Using monitoring tools that implemented these, you could see how systems were performing. But you really couldn't dig too much into what the applications were doing. It was like a black box.
Since applications were simpler then, the need wasn't as bad.
Then came Java in 1996.
The ability to write code using the Java programming language, and run it on any operating system, changed everything.
The Java Virtual Machine (JVM) introduced another layer of abstraction, reducing application visibility even more. In cases when a whole application, with all of its components, runs inside the JVM, SNMP-based or sniffer monitoring tools could not help much. At this point, these tools are doing infrastructure performance monitoring(IPM).
So APM needed to evolve again.
New agent-based tools were introduced in the late 1990s and early 2000s that would provide visibility inside the JVM. Monitoring the monolithic Java applications of that day became possible.
During this time, the infrastructure wasn't changing that much. The software development process, using the Waterfall model, was pretty slow by today's standards. Because of that, many business applications were installed and running on only a handful of servers.
But that began to change.
Phase 2: Web 2.0 days of APM
The mid to late 2000s brought about changes in software development that lead to things such as service-oriented architectures (SOA). It also marks the start of the Web 2.0 era of the Internet.
During this time, many businesses were moving away from Waterfall to Agile software development process. To do this, many single-tiered and siloed applications were broken up. SOA helped accomplish this because various application components broken up into services could now run on separate systems.
With SOA also came the increased use of XML-based Simple Object Access Protocol (SOAP). While this helped with breaking up applications, XML's chattiness proved to be a performance problem. This meant that you had to be mindful of the distance between your XML clients and servers.
Having these application services run on different systems increased application complexity. The monitoring offered by Phase 1 APM tools was no longer enough. Instead of having one agent that could report metrics data, we now needed multiple agents. Data sent back to the tool needed merging to create a single view of application performance.
Also, visibility all of the communication between services that are on the same servers was a must-have. APM tools were needed that could collect not only Java and .NET transaction data, but also from other languages. APM vendors had to start supporting languages such as PHP, Rails, and Python being used more for software development.
APM tools needed the ability to follow a user transaction from one service to another. They also needed to provide a single view on overall application performance for that transaction. So, Phase 2 APM tools had to give us more.
But ever the constant, more change came ...
Read The Evolving Needs of Application Performance Monitoring - Part 2, covering APM today and in the future.
Industry experts offer thoughtful, insightful, and often controversial predictions on how APM and related technologies will evolve and impact business in 2020. Part 2 covers AIOps, AI and Machine Learning (ML) ...
As the New Year approaches, it is time for APMdigest's 10th annual list of Application Performance Management (APM) predictions. Industry experts offer thoughtful, insightful, and often controversial predictions on how APM and related technologies will evolve and impact business in 2020 ...
Enterprises with services operating in the cloud are overspending by millions due to inefficiencies with their apps and runtime environments, according to a poll conducted by Lead to Market, and commissioned by Opsani. 69 Percent of respondents report regularly overspending on their cloud budget by 25 percent or more, leading to a loss of millions on unnecessary cloud spend ...
For IT professionals responsible for upgrading users to Windows 10, it's crunch time. End of regular support for Windows 7 is nearly here (January 14, 2020) but as many as 59% say that only a portion of their users have been migrated to Windows 10 ...
Application performance monitoring (APM) has become one of the key strategies adopted by IT teams and application owners in today’s era of digital business services. Application downtime has always been considered adverse to business productivity. But in today’s digital economy, what is becoming equally dreadful is application slowdown. When an application is slow, the end user’s experience accessing the application is negatively affected leaving a dent on the business in terms of commercial loss and brand damage ...