AppDynamics, a Cisco company, today unveiled a new platform that will unite all the areas of its application monitoring business. It also helps the company position itself with an AIOps mindset — using artificial intelligence (AI) to automate IT operations — to navigate the challenges of modern environments.
The platform, deemed the Central Nervous System, is propped up by three pillars — visibility, insight, and action — according to Matt Chotin, the senior director of technology strategy at AppDynamics.
“Technology environments are getting too large and too dynamic to manage manually,” he said, referencing IoT, multi-cloud, and containers. “IT needs to change and adopt an AIOps mindset to navigate this challenge. We view AIOps as a philosophy that values prediction over reaction, answers over investigation, and action over analysis.”
The Central Nervous System, aptly named after the human central nervous system, is an extensible, open platform that will leverage third-party systems to ingest data, correlate and analyze across domains, automatically remediate problems, and optimize performance.
As part of the platform’s launch, which Chotin referred to as “ the beginning of a journey,” AppDynamics launched three new products. Two products support the visibility pillar: a serverless agent for Amazon Web Services (AWS) Lambda environments and an application monitoring tool for Cisco ACI. The third product, which falls within the insight pillar, is a machine learning engine.
Chotin noted that while AppDynamics is not yet announcing anything around the action pillar, it has “been doing a lot of work with partners to automate tasks” in areas like incident response, workload optimization, release orchestration. “It’s a journey for us so we will continue to build reference architectures for our customers as well as functionality ourselves to help them on that part of the journey.”
Monitoring for Serverless and Cisco ACI
AppDynamics released a special purpose agent designed specifically for tracking how AWS Lambda — AWS’ serverless computing platform — functions perform in the broader context of an application. “Existing tools today they focus on the individual implications of Lambda functions, but almost operating in isolation,” said Chotin.
One of the features of AppDynamics monitoring is that it tracks business transactions (e.g. login or checkout) and how it ties to application performance.
In the instance of the Lambda, since it and other microservices architectures are being used as part of a larger application, the AppDynamics agent helps users see end to end how a transaction is performing. “Even when it touches all of these different [network] components, some of which may be executing through Lambda,” he said.
This is just the first microservices architecture that AppDynamics will build into its platform, eventually supporting other platforms similar to Lambda.
The company also launched AppDynamics for Cisco ACI, which is Cisco’s service that enables network operators to manage the network by organizing around applications. “Until now, we’ve relied on network engineers to actually manually inform ACI of what the application architecture is,” said Chotin.
AppDynamics for Cisco ACI solves this by automatically applying application context when monitoring ACI environments. It can also map business transactions, and apply root cause analysis, to their correlated network endpoints to understand where errors and dependencies are occurring.
AppDynamics Machine Learning Engine
The final launch from AppDynamics is largely based on its acquisition of artificial intelligence (AI) startup Perspica in October 2017. The product, Cognition Engine, provides anomaly detection using machine learning models. These models reduce the time to identify and fix network issues by learning what normal and healthy behavior looks like in a network so that deviations are easier to detect.
The engine integrates the machine learning engine and streaming analytics from Perspica with its existing monitoring platform. “We’re now able to also use machine learning to isolate the metrics that are deviating from normal and present the top suspects of root cause for any application issues,” said Chotin.