How to Overcome Limitations of Conventional Discovery in the Cloud

Following a few years of fits and starts, enterprise use of the cloud is growing exponentially, and wider deployment is coming as more companies adopt cloud-first strategies. IT leaders are looking for better ways to manage on-premises IT assets with newer cloud-based assets. Common early cloud use cases (e.g., development/test) aren’t mission-critical and don’t require rigorous asset management. This is changing as organizations move production workloads to the cloud. As IT migrates to the cloud, major obstacles surface. One is a lack of comprehensive visibility. In this eWEEK slide show, Walker White, vice president of data platform at Flexera, offers his expertise about the limits of conventional discovery and best practices for cloud asset management.

Conventional discovery tools are able to identify only instance-level data at the operating-system level. They are unable to capture information about the applications running on the cloud instance OS.

Conventional discovery tools are generally unable to identify specific applications running on cloud instances. When they do look at cloud instances, they may not recognize data beyond the instance level to the applications. If they are able to recognize at the application level, only the most common applications may be included while others may be missed, leaving a gap in discovery data.

Conventional discovery tools for the cloud typically identify instance data only on the cloud and not the applications. They are also often in a format that doesn’t allow a full view when combined with on-premises data that has been identified by on-premises discovery tools. In addition, what may be discovered often contains duplicate, irrelevant and incomplete data.

Cloud platform services are typically used for development and testing. Quick access to cloud-based virtual machines makes it easy for developers and test engineers to quickly “spin up” and “spin down” new machines to test software. Unlike development and test workloads that are short-lived, production workloads are subject to rigorous service-level agreements (SLAs), vendor audits and regulatory review for companies that operate in highly regulated industries, such as financial services, government and health care.

With the inability to dive deeper than the OS level and recognize applications that have been migrated to the cloud, the challenge is even more daunting with the inability to recognize the thousands of Linux-specific commercial applications that may have been migrated. Without a standard trusted and comprehensive source of all of the available applications, ensuring complete and sufficient coverage is impossible.

To gain a true picture of IT application assets, cloud discovery tools must be able to read and recognize all applications that have migrated to the cloud IaaS. OS-level data is not sufficient.

Native integration with the cloud vendor toolset can ensure all of the available information from the IaaS platform can be accounted for, including short-lived instances.

Native integration with Amazon EC2 Systems Manager enables the capability to collect even short-lived or ephemeral instances.

Discovered data often contains duplicated, irrelevant and incomplete information. Data that is discovered and normalized against a de facto standard provides a clean, complete and comprehensive view of asset data by removing all duplicated, irrelevant and incomplete data.

Clean, normalized data enriched with market information such as end-of-life (EOL) data, end-of-support (EOS) data and vulnerability information can enable more insights about IT asset data to enable better decisions for software licensing, procurement, software audits, IT service management, cyber-security risks and much more.