Nowadays, Data Center Infrastructure Management (DCIM) means different things to different people. No longer limited to managing just power, space and cooling of a data center, today’s DCIM software are also monitoring IT assets like servers, storage and network devices. In the case of Hyperview, our software goes one step further and monitors virtual machines (VMs) too. Soon, you can expect DCIM software to go beyond the data center walls to encompass public cloud and IoT. This is why analysts have coined new definitions like DCM (Data Center Management), DCSO (Data Center Service Optimization) and CIM (Critical Infrastructure Management) to better describe the technology.
The need for auto-discovery became more urgent as DCIM tools started reaching beyond power and cooling. Having to now manage thousands to tens of thousands of IT assets, manual data entry was no longer a viable option. A better method was required to replace this wasteful and error-prone way of capturing data in the system. Enter auto-discovery.
As with DCIM, auto-discovery is susceptible to different interpretations. Hyperview defines auto-discovery as the “ability to detect a device, model it and measure the relevant data points of that device”. A device can be a floor PDU, a network switch or a VM. At first glance, this may look like a simple task–discovering a floor PDU with SNMP is fairly straightforward. However, discovering a stacked switch that is composed of several switches combined into a single virtual chassis is a bit more challenging. For an auto-discovery tool to be effective, it has to understand the device being discovered, and be able to model it as accurately as possible. In order to do so, the tool must be able to understand an expansive list of protocols such as SNMP, IPMI and the like. Any auto-discovery tool that can perform these functions is invaluable to the data center manager. The tool allows a data center operator to get meaningful data on every component of the data center well beyond power, space and cooling. As an example, auto-discovery enables a data center operator to find possible zombie servers simply by querying for servers that have low CPU usage. Getting a list of router firmware levels to plan for an upgrade is another example.
From low disk space on a virtual machine to a failing fan in a blade server chassis, the benefits are clear. Further, the value of accessing all of this intelligence through a single pane-of-glass as a result of auto-discovery makes the tool invaluable in the day-to-day operations of a data center.