In order to be frontrunners, businesses must ensure a robust IT infrastructure for high uptime. To make sure all the elements of the IT infrastructure are up and running, an organization must perform rigorous application performance monitoring. With better application performance monitoring, organizations can enhance their service reliability. Among all the tasks performed by a monitoring system, ADDM (Application Discovery & Dependency Mapping) is the most important. Without ADDM, data experts would not have full observability into their IT infrastructure. Read on to know more about ADDM and the best tool in the market to implement ADDM.

Understanding ADDM

ADDM is an IT process used to discover applications within an IT ecosystem. Once all the software applications within the IT infrastructure are discovered, the external dependencies are discovered. An application software within an IT infrastructure is dependent on several external elements like networks, servers, databases, etc. One software application in an organization can also be dependent on some other application. It identifies all the components of a software ecosystem. To understand the collaborative and comprehensive working of different applications, ADDM is required.

Why is there a need for ADDM?

ADDM is essential for maintaining service availability and optimum performance of software systems. Consider an example of 5 gears that are inter-connected to one another. When the first gear moves, it triggers movement in the second gear. Similarly, other gears trigger movement in their successive gears. Now, if the first gear fails to work, the whole system will cease to work.

It is a similar case with software applications that are dependent on each other. A change in any application can affect the IT elements that are dependent on it. A change in any application could trigger a series of changes in the IT infrastructure. It may also happen that one application failure leads to another due to interdependencies. To enhance service availability and to understand your IT infrastructure in deeper details, ADDM is required. IT experts often use the term ‘Dependency Mapping’ for the process used for uncovering dependencies between IT elements.

Problems with manual mapping

Designated ADDM tools are used by businesses for higher productivity. Organizations that have a vast IT infrastructure are bound to use ADDM tools as manual mapping is not feasible. The issues with manual dependency mapping are as follows:

• The IT infrastructure of organizations is getting bigger and more complex. IT professionals cannot handle the complexity of the IT infrastructure without the help of an automated ADDM tool.
• Due to the increase in the size of IT infrastructure, firms have to hire many IT professionals to complete the ADDM process. Hiring too many IT professionals increases operational costs.
• Round-the-clock ADDM cannot be ensured via manual mapping. If there is a sudden demand for ADDM outside working hours, it may get difficult for the firm. If ADDM is left unnoticed for long, it could jeopardize service reliability and availability.

Types of ADDM and their pros/cons

There are various ways of completing the ADDM process and boosting service availability. Some of the popular traditional techniques for ADDM are as follows:

Sweep and poll

Sweep and poll is an age-old technique for application discovery within the IT ecosystem. IT professionals ping the IP addresses of various applications within the IT infrastructure. The devices that respond to the IP address pinging are discovered. The sweep and poll method uses device group information and ping response to discover IT elements. Using sweep and poll, you can traverse through the entire IT infrastructure from a central node location. It is easy to implement but is not feasible in the current scenario of complex IT infrastructures. The cons of the sweep and poll method are listed below:

• For vast IT infrastructure, sweep and poll can take more time to discover application interdependencies.
• Sweep and poll can leave some critical IT assets undiscovered and hamper the dependency mapping. You may not have full observability into your IT assets with sweep and poll.
• It cannot discover all the interdependencies between your software systems or applications.

Network monitoring

Network monitoring studies the real-time packets that are shared in an IT infrastructure. It works on two levels that are packet-level and flow-level. In flow-level network monitoring, routers in the IT infrastructure send/receive NetFlow records. It can offer real-time application performance monitoring and detect the interdependencies between applications immediately. The cons of network monitoring are listed below:

• Network monitoring works appropriately for WANs but fails for large data centers.
• The flow-level records only notify about the IP address and TCP port of applications. It cannot discover granular interdependencies between the applications.
• Packet capture is expensive and can increase the cost of application performance monitoring for an organization.

To overcome the shortcoming of traditional approaches to ADDM, many organizations deploy a software agent on the client side. The software agent then conducts real-time monitoring for the incoming and outgoing traffic. However, to enhance observability, organizations must place software agents heavily in the IT infrastructure. It will again increase the cost of application performance monitoring for an organization. As more and more businesses are adopting hyper-automation, traditional methods for ADDM are not feasible anymore. AIOps (Artificial Intelligence for IT Operations) is a better solution for ADDM in the current scenario. The pros of using AI for application monitoring are listed below:

• An AIOps based analytics platform can discover all IT entities (network, storage, and applications) in lesser time.
• AIOps can perform effective ADDM even for complex or dynamic IT environments.
• Once AIOps has all the interdependencies between IT components, it uses predictive analytics models to forecast performance issues.
• AIOps-based tools can create topology maps to understand the interdependencies between IT components.

Conclusion

An AIOps based analytics platform can automatically update your CMDB. You can collect configuration data from CMDB and can discover interdependencies in your IT infrastructure. An AIOps based analytics platform will perform round-the-clock ADDM for your organization without any manual interruption.