Data Center Infrastructure Management (DCIM) systems are a growing category of management products for data centers. While different DCIM products offer different views of the data center infrastructure, all DCIMs consume information from network servers and other components such as power and cooling systems.
Additionally, all DCIM solutions have some level of physical layer information (which is very often manually and inaccurately documented). Automated Infrastructure Management (AIM) systems provide information about the current state of the physical network, indicating which cables are plugged into which ports, and so they function as key providers of information to DCIM products.
There’s a lot of confusion in the market about the relative roles of DCIM and AIM solutions. Let’s take a quick look at how AIM solutions work to make DCIM solutions more effective.
The time span between new generations of technology is growing shorter, and this phenomenon is especially apparent within the data center. Most data centers started as connectivity technology housed in wiring closets, then quickly started requiring entire floors. Today, many data centers are in their own buildings comprised of hundreds of thousands of square feet, housing hundreds of millions of dollars in IT and facilities investments.
As data centers grow exponentially in scope and complexity, DCIM is one of the technologies to emerge. DCIM solutions bring management control to the data center. In a data center, there are three different teams to manage operations: infrastructure, networking, and facilities. Typically these teams operate in their own silos with their own naming conventions. A DCIM pulls this information together through Application Programming Interfaces (APIs) to complete a physical representation of the whole data center. This provides a single, overall view of the data center for such key functions as command and control, asset management, resource management, and forecasting.
For example, a data center manager can use a DCIM tool to determine whether there are servers or space that might be underutilized. This information can help the manager to see if other applications can be virtualized onto these machines to free up servers, or perhaps allow the manager to decommission unused equipment for cost savings. DCIM can also help identify equipment that is consuming too much energy, which, in turn, can help managers determine a better strategy to reduce energy consumption and costs. These are just a couple of examples of what a DCIM solution can do.
But DCIM solutions alone don’t provide complete visibility into the data center infrastructure. What’s missing is documentation about the physical layer – the cables and patch panels that tie the network together. In many data centers, physical layer information (what is connected to what, where) is maintained in spreadsheets or other manual tools that quickly become out of date. Such manual systems are error-prone and labor-intensive – some studies suggest that more than half of a cable technician’s time is spent looking for the location of a problem and documenting that location rather than fixing it.
An AIM system collects information about physical layer infrastructure. Using intelligent cables and patch panels, the AIM system collects specific data about what cables are plugged into which patch panel ports. Many AIM systems also include work order management systems that can cause port LEDs to blink to indicate the location of a problem, thereby saving lots of time for cable technicians. Moreover, the physical layer infrastructure data is always up to date because changes are automatically recorded in the AIM solution’s software.
AIM systems integrate to bring connectivity information into the DCIM solution. In an integrated system, for example, if the DCIM wants server port A connected to switch Port B on patch panel C, the AIM system shows proof that this switch port went to this patch panel.
To integrate an AIM solution into a DCIM solution, APIs are used to exchange data between the two solutions. Data center managers often deploy things uniquely, so there is some customization that has to occur, but the process involves mapping an API to an API. In selecting DCIM and AIM solutions, it’s best to choose DCIM and AIM solutions that are open rather than proprietary to simplify integration.
Currently the International Standards Organization (ISO) is writing a standard for AIM APIs, and those are going to be adopted in standard-compliant products. If all the AIM systems share a common API structure, it will be easier to integrate them. The ISO standard (ISO 18598) should be ratified in 2016.
With DCIM and AIM solutions, data center managers have a complete view of the data center infrastructure that is up to date and easy to read. Using these solutions together, data center managers can be much more effective at planning, control, and resource allocation.