Dimensions of Scale in Performance Management
Networks and data centers, the backbone of any company's IT operations, are growing in complex and often unpredictable ways. New software applications, hardware upgrades, virtualization efforts, and even workforce distribution are all impacting how these systems evolve. And that evolution makes it more difficult to distinguish and identify the variables contributing to problems in performance.
In order to maintain performance visibility, you have to break down issues of scale and address growth across multiple dimensions. We measure scale in performance monitoring in three ways: volume, speed, and access.
1. Volume – Volume covers the number of items you want to monitor. This includes hardware and software, but also . Custom calculations offer summarized data across multiple objects in your network, and synthetic indicators create new metrics for a single object from simple math operations performed on multiple data inputs (e.g. the ratio of packets sent to errors received). The volume of KPIs can be overwhelming, particularly because the more your network grows, the more the number of items you want to monitor tends to increase as well. However, there are ways to manage the volume by using aggregate values that represent a "rollup" of several individual KPIs. Even as the number of individual links in your network grows, for example, the one value that likely matters most is how many total packets are delivered across all of the links put together.
2. Speed – How quickly do you need to receive performance data? Most of our customers need a significant amount of performance data in near-real time, and as systems become more geographically distributed, that data sometimes has to travel long distances or across complicated routes as well. Network degradation or a data center fault can have serious business consequences. The faster you can diagnose a potential problem, however – i.e. the faster you receive performance data – the better chance you have of fixing it before there's an impact on the customer experience. To be able to respond quickly to network anomalies, you need to be able to, access reports and updated dashboards on demand, and receive immediate alerts when standard deviation logic shows a change in network performance.
3. Access – The third dimension of scale is how many users need access to performance data. As a business grows, the number of people who need visibility into performance data, and who may need to pull their own customized or on-demand reports, also rises. This is complicated too by the fact that increasingly employees are geographically spread out, meaning that data has to be accessible from many locations. Scale is as much about widespread access to performance data as it is about how large your systems grow, and how quickly they need to be assessed.
Access is also important for providing self-service reporting to internal business units. Custom portals and dashboards give users a chance to see the data themselves, which reduces calls to help centers, reduces the number of tickets opened, and saves time and money for your business.
In many industries, IT is now such an integral part of business function that growth in performance monitoring has to be considered right alongside the evolution of a network or data center. Downtime, or even less-than-optimal performance, isn't an option. The scalability of your performance monitoring system is critical for preventing problems as IT operations continue to grow.
Tom Griffin is Director of Systems Engineering in EMEA.