In many ways, the internet of things offers the perfect use case for virtualization — an elastic and highly adaptable environment and framework capable of adjusting to the dramatic lulls and spikes that characterize the data streams between machines. In much the same way, IoT, with what is shaping up to be an innumerable collection of connected devices, requires the dramatic scale that most traditional networks can’t provide in a cost-effective way.
Of course, IoT isn’t the only driver of virtualization. The ability to dynamically allocate CPU capacity, memory and disk space to applications as needed is a powerful benefit in and of itself. Not surprisingly, for those just beginning their journey with virtualization, it can feel like performance, capacity and scale are unlimited — and in many ways, that’s understandable.
Even so, as IoT continues to add additional devices and data to these same networks, an inescapable reality becomes clear: It’s just as important to monitor the health and performance of your virtualized network and components as other parts of your infrastructure. In fact, reliable network monitoring plays an important and crucial role in dynamically assigning the very capabilities and capacities virtualization makes possible.
It’s also no secret that any network failure in a virtualized environment can have a dramatic impact on the applications in question, and for that reason alone, network monitoring is required to ensure that system administrators, network engineers and IT teams know as soon as any issue arises. That, of course, isn’t all. Following are some of the many reasons it’s important to monitor your virtual assets, as well as specific things you’ll want to keep an eye on to ensure that your IoT initiatives function smoothly with the virtualized network you ultimately put in place.
- Use monitoring to ensure that you plan for the virtual assets your IoT initiative will demand. Increasingly, smaller organizations are looking to virtualization and the many benefits it offers for the first time as their IoT initiatives grow in scale and scope. Network monitoring technologies and capabilities should be considered a necessary investment for organizations at this important stage for a simple reason: It’s critical to understand what different applications demand of your network resources. Virtualizing systems without knowing the CPU and memory load, dish usage and network usage is very risky. Surprises are never good in any network, and virtualization is not a one-size-fits-all proposition. It’s important to plan your efforts based on the facts, not intuition.
- Use monitoring to optimally assign resources. With virtualization, you always want to find the right balance. You don’t want to assign too few virtual machines to a host — and in that way waste resources — but you also don’t want to overload any virtual servers — and in that way slow or disrupt the performance of all of the systems and applications running off of it. Network monitoring technology enables you to see in real time how these virtualized resources are being used during periods of peak usage and lulls — information you can use to assign just enough of the resources at your disposal, but not too much.
- Use monitoring to maintain quality of service. Today’s network monitoring technology not only enables you to see what’s happening in your network right now, but it also provides you with access to a historical record and view that can be used not only when troubleshooting any problems that arise (you don’t have to replicate or recreate scenarios to see what happened when a real historical record is available and easily accessed), but also compare how any changes to the network ultimately impacted the quality of service. Virtualization is no exception. Monitoring before and after you begin and complete your virtualization deployment will enable you to demonstrate just how much the virtualization effort improved the service experienced by users.
Fortunately network monitoring technology has advanced with virtualization, and today there are many sensors and technologies available that provide a real-time view not only of what’s happening in these networks, but also simultaneously in the traditional infrastructure and data centers they augment. Some of the many performance metrics and capacities that can be monitored in virtual networks include CPU usage as a percent of guests, CPU usage for each hypervisor, total CPU usage, read and write speeds, the number of packets sent and received by bytes or time period (such as seconds), network usage, disk usage, available disk capacity, active memory, consumed memory, the number virtual machines running, load average and the health of any host hardware, including temperature, power, fan rotations per minute and battery voltage. These are only to name a few.
A bright future, of course, lies ahead for IoT as we find new ways to make life better with connections that put more information than ever at our fingertips. But just as these connected devices will enable us to act with greater knowledge, the virtualized networks that make them possible will require far greater diligence on the part of networking professionals who must now, more than ever, take steps to ensure that they know exactly what’s happening across their networks at any given point in time.
All IoT Agenda network contributors are responsible for the content and accuracy of their posts. Opinions are of the writers and do not necessarily convey the thoughts of IoT Agenda.