By Dr. Thomas Wellinger, Market Manager Data Center R&M
A data center is an ever-expanding, interdependent, difficult-to-understand infrastructure that, once installed, is there to stay. Data center managers must address constant demand for new services and booming data growth. To make the required smart decisions about the infrastructure, they need asset-level visibility into the operations of all their data centers.
Global visibility and KPIs are the first step on the path to effective infrastructure management. As the American professor and management consultant William Edwards Deming once said: “In God we trust; all others bring data.” Driven by the DCIM industry, there have been a lot surveys about ideal KPIs. Usually, these boil down to PUE, total energy cost and delivery cost (€/kWh). More recently, Gartner introduced a different type of metrics. To calculate the IT utilization KPI (IE/CE) these take into account computing capacity, differentiating between energy consumption when all systems are idle (IE = idle energy) and when all systems are under load (CE = computational energy).
In the absence of good information, decisions are based on intuition, trends and incomplete, fractioned data. As a result managers will choose to err on the side of caution and assume the safest choice is over-provisioning. However, that is a very expensive, highly inefficient way to run a data center. An automated infrastructure solution can help provide an up-to-date overview of servers, network, storage, cabling, rack space, port capacity and more.
Key Performance Indicators for data center efficiency
Let’s take a look at a number of KPIs that can help you monitor, analyze and optimize your infrastructure and service offering.
By keeping track of documentation time and average provisioning time, you can lower provisioning and service delivery times. This will certainly result in greater satisfaction amongst internal and external customers. Documentation accuracy - the number of correctly and up-to-date documented items - and Mean Time To Repair (MTTR) are vital indicators which can lead to increasing reliability.
There are also several KPIs essential to increasing infrastructure efficiency and density. Lost capacity can be calculated by examining rack space in relation to limitations caused by factors such as power constraints, whilst Space efficiency looks at floor and rack space usage per area.
Infrastructure efficiency can be further increased through monitoring the utilization degree of switch/router port capacity, looking at the number of provisioned active ports in relation to the total number of present active ports. The number of FTEs supporting the data center in an operational capacity is another interesting indicator that can be related to performance levels. Keeping track of these and making corresponding changes can help significantly optimize operational costs and utilization whilst improving TCO. By tracking the number of unused active and passive ports - or free connection capacities - it becomes possible to optimize on the fly as well as plan ahead. This is also vital to securing capacity.
Besides supporting each of these KPIs, there is another major advantage to introducing Automated Infrastructure Management (AIM). In today’s world, many interdepended aspects of data center operation are outsourced to (multiple) third party companies. Company A might be in charge of network operation, for example, while company B may be responsible for MACs.
In the past, if something went wrong, you would have to speak with someone from your own organization to find out what happened and resolve the issue. What is more, the entire legal responsibility for anything that had an impact was clearly defined.
However: what happens if everything is outsourced, something goes wrong, and you have no way of recording and tracking the individual tasks? Each party will simply blame the others and in the end you might be held accountable.
Good information is hard to find
For data center infrastructures in particular there is huge potential to improve on processes, such as planning, forecasting, creating inventories and MAC processes. A smarter approach will also lead to a highly standardized service catalog as well as constantly high reliability and data integrity of documented actions is an outcome. This is especially important for highly regulated industries such as the financial, pharma or chemical sectors.
Having better visibility is the essential first step in moving towards a more mature approach to infrastructure management. With visibility in place, you can start proactively planning based on predictions and forecasts, and eventually move towards KPI-driven data center management.