What is application dependency mapping?
Application dependency mapping is an accurate map of an entire IT ecosystem, including infrastructure, applications and their dependencies.
The need to map applications dependencies is becoming widely recognised. Unfortunately, what isn’t nearly as universally understood is how to effectively perform this vitally important task. While there are many different techniques for application dependency mapping, each producing different results, there are four primary approaches that every business should be aware of.
1. Sweep and poll
This is one of the oldest and lightest-weight methods. It sweeps and pings IP addresses, gathers the type of device that was pinged, and provides information about what is running on the server and the applications it is using. There are typically blueprints of the layout that it uses to search for information and identify the pieces that frame out the larger structure.
The advantages of the sweep-and-poll method is that it’s relatively easy to perform and allows users to sweep an entire network from a single location. The downside is that dynamic, complex environments can affect the sweep accuracy, and a single datacentre sweep can take a long time.
2. Network monitoring
Network traffic patterns are analysed either at the packet-level using packet capture or using NetFlow at the flow level.
Network monitoring occurs in real time, making it easy to detect dependency changes as they happen. Also, because network monitoring does not refer to pre-built blueprints, this method is effective within less understood systems. Disadvantages of network monitoring include issues related to scaling and duplicate flow records, and problems with differentiating between different application-level dependencies.
3. Agent on server
Agents provide real-time monitoring of incoming and outgoing traffic. This allows them to find and understand the components while recognising changes to any statuses as the topology changes.
Agents within the server provide the benefit of real-time monitoring and are capable of easily distinguishing between multiple applications running on the same IP address. However, an organisation will need to place agents on every relevant, connected application, technology and server to monitor it, potentially increasing the cost significantly.
4. Application dependency mapping
This leverages orchestration platforms. It deploys and maintains each underlying application component. Consequently, the orchestration understands, at all times, which individual components compose an application.