By Jon Trout, President of Compass EdgePoint Solutions
If you have ever implemented a management system for your data center, you probably remember the process had more moving parts than the average car engine. As the need for low latency data delivery continues to grow, so will the number and distribution of an organization’s data center locations. You may be saying to yourself, “Man, can you imagine how complex it is going to be to implement a monitoring and management solution for distributed infrastructures?” If so, you see where I am going, and you are right. Multiply everything you need to do for a single data center by multiple sites, and you quickly discover that there are some things you will need to work out on the front end before you implement a distributed infrastructure management system.
A review of a recent Compass Edge Solutions’ project illustrates the types of considerations you will need to address. Recently, we were approached by a customer who was rolling out a large fiber to the home network and wanted a single platform to monitor environmental factors, power, IT and physical security for a few hundred of their high-density telecommunications shelters. After immediately agreeing to take the job, we sat down to reflect on what we had committed ourselves to and how to complete the project. The first step was to sit down with the customer to define the parameters of the solution.
Planning for any project, not just a management system, means starting with the end goal in mind. Determining our customer’s future system capabilities required them to answer a number of questions including:
- How many locations did they want to monitor?
- Did every site need to be accessed physically and virtually?
- How did they want specific information to be displayed?
- What components, at each site, require control from the central console (for example turning an HVAC unit on or off remotely) versus on-site?
After determining the desired system capabilities, we proceeded to examine the customer’s physical and virtual architecture to assess its capacity to support all of the desired applications. During this phase of assessment, it is critical to include all involved parties. For example, in our evaluation, we found that support for low bandwidth devices was on an internal network, and the remote site cameras utilized a production network. Unfortunately, the network security team identified a few security issues associated with operating the cameras on an external network. This is the type of potential problem that must be addressed upfront by all project stakeholders.
Since change is a constant in any data center, we spent a great deal of time with the customer identifying any differences between site configurations, and what their plans were for any upgrades or refreshes. Depending on the type of change, specific information may be required to update the system’s display of the shelter’s configuration. A new generator, for example, makes a significant difference. Performing a detailed site analysis was instrumental in developing formal procedures for customer site updates to ensure the integrity of the system.
Site alarms announce that there is something amiss at one or more sites, but since not all alarms are of equal severity, determining a “ranking order” for these alerts is imperative. In any distributed infrastructure the triage for alarms reflects what problems require immediate attention as opposed to those that can be addressed later. Developing an alarm hierarchy with the customer will reduce its ongoing operational costs by limiting the number of “truck rolls” that their personnel has to make to resolve a problem.
It is an unassailable fact that we will see rapid growth in edge and distributed infrastructures in the coming years to move information as close to the end user as possible. That this growth will fuel the need for comprehensive management and monitoring system implementations is also indisputable. As our customer experience demonstrates, there is no such thing as over-planning to implement management functionality for a distributed infrastructure.
About the Author:
Jon Trout is the President of Compass Edge Solutions, which provides solutions for edge computing and distributed IT infrastructure. Compass Edge Solutions is a part of Compass Datacenters, which provides solutions from the core to the edge. Trout is one of the industry’s foremost experts on edge computing and integration of distributed facilities. To contact him, email email@example.com.