In our last blog post, we discussed why the architecture of the desktop application monitoring solution is important in delivering reporting data. This week, we’ll focus on one component of an application monitoring solution architecture – Agents – and why we believe they are fundamental to a modern architecture.
First, our assumptions:
- Real-time monitoring is fundamental to real-time usage visibility
- Desktop application execution is changing from desktop to Software-as-a-Service (SaaS)/Cloud
- Usage monitoring will extend from individual applications to collaborations
- Usage monitoring will extend from individual applications to hybrid environments such as execution on the desktop with data stored in the cloud
- We are focusing on applications that are primarily individually user oriented, not enterprise applications such as SAP or Oracle
To gather data about desktop application usage, the solution has to be able to obtain the data. Back in the good old days, usage data gathering was relatively easy – application licenses were tied to a desktop, or they were metered from a server. Putting a monitor either on a desktop or license server solved the problem, as these were essentially the two points of access to data collection, and both were typically “owned” by the enterprise. Larger organizations tended to license applications with server-based licensing, so desktop monitoring solutions gravitated to reading server logs to get usage information.
Internet-based applications change the paradigm in several ways:
- Licenses delivered through the cloud don’t come off a local server
- Browser-based applications don’t have desktop executables to monitor
- Cloud-based, hybrid applications may charge on metrics such as data volume, with usage information available via API calls
To make matters worse, the shift to Bring Your Own Device (BYOD) to access cloud-based applications essentially takes all usage data access out of IT’s hands. Those monitoring solutions that are built around monitoring server logs will quickly find themselves monitoring a rapidly shrinking pool of applications.
So what are the options for monitoring the new/future applications?
We see a couple of options:
- Enhance a server-based monitoring approach to include a network sniffer
- Enhance a server-based monitoring approach to access APIs from each SaaS vendor to be monitored
- Put an Agent on each device to be monitored
Cetrus chose Option 3 for a number of reasons:
- Server-based approaches are inherently limited by the data and timing granularity of the server – namely, the server will report which machine requested what application and when and when that license was returned. Server logs don’t provide any actual usage information. Enterprises have to create data marts to supplement server-log-based approaches to know who the user is, what group the user belongs to, etc., to provide user -centric or group comparison data.
- Enhancing server approaches with a network monitor creates two different data sources requiring IT to build a data mart for consolidation. The enterprise still has the data granularity problem with server logs. Providing a real time view of usage with disparities on reporting frequency can cause confusion as to what was really occurring, such as apparently being active in multiple applications at the same time. Providing a real-time view of usage with disparities in reporting frequency.
- BYOD use outside of the office/network make it difficult to monitor SaaS applications, further impacting the quality and completeness of usage data
Given the complexity and the increasing number of points of application usage and application generation, the data collection point that covers the most application types will be the simplest to manage. We believe that point of data collection is an Agent on each device.
Agents can monitor applications running on the desktop. They can monitor applications running in the cloud. They can monitor hybrid applications. The single point of usage ensures that no backend data consolidation from multiple data feeds has to take place. Pulling attribute information from other data sources occurs when a usage change happens, so that associating usage to a project can occur in real time, or identifying which license server provided a license is immediately reflected for reporting. Only an Agent-based platform can provide “to the second” visibility of actual application usage regardless of where the user or application is.
Cetrus Process Meter is the only Agent-based application monitoring solution. We built the architecture to grow with where the industry is going rather than where it’s been.
Give Process Meter a try. Send an email to email@example.com, or call +1 877–238-7871 and we’ll set you up with a no-charge trial account.