The Internet of Things (IoT) is quickly becoming the backbone of today’s digital infrastructure and many modern enterprises, as billions of new connected devices come online each year. What I observe on a daily basis while evaluating data center designs and IT platform resiliency for Uptime Institute, is that enterprise executives today are increasingly dependent on edge computing to provide critical enterprise services. In the age of IoT, most businesses could benefit greatly from a simple checklist to understand the fundamental aspects required to make their edge computing designs more reliable.
Edge computing can range from being a large colocation facility near a population center, to a single rack of critical server, storage and network gear located at a remote location, such as a cell tower or network hub.
The increasingly critical nature of content supported by edge data centers ranges from low criticality consumer content delivery such as Netflix and personal assistants from Google, Amazon, etc., to life-critical commercial content like coordination between self-driving vehicles and more. As such, edge data centers require at least the same design diligence devoted to large scale data centers, if not more. Providing the continuous “always on” or “always available” and “low latency” environment has become progressively more important, and expected.
Resiliency matters
All of us in the data center and IT industry know painfully well that if not properly designed, a single weak link anywhere in the IT stack can cause a service to become unavailable or unusable to the customer.
As a rule-of-thumb to better oversee IT environments, executives and senior managers should focus on the following five areas of edge computing to help ensure end-to-end resiliency:
1. Design and Build – Edge or CDN (content delivery network) data centers have become a critical path component in the overall IT stack delivery model and require at least the same level of focus on design and implementation quality as large enterprise data centers, since many edge data centers are often operating remotely or un-manned. This requires that the components used in all aspects of the edge design and build be qualified and tested as an integral unit, since repair in the field can directly impact the end user and cannot be remediated quickly due to the lack of on-site or even nearby staffing. Uptime Institute’s “Tier Ready Certification” is an important qualification to look for in any remote edge installation.
2. Remote Management and Automation – Facilities and IT designs need special focus because all systems must be able to be operated remotely and with automation in mind (or be completely autonomous). This requires a higher level of on-location monitoring and specialized BAS (Building Automation System) to control the environment, detect and report anomalies, and fail-over to designed-in on-site redundancy when needed. These types of operations and controls systems are complex and require a well-thought-out and high-quality implementation to operate reliably.
3. Network – Communicating with edge data centers is critical to the success of the IoT model, from both end-user and data center operations perspectives. Ensuring connectivity requires redundant networking components, including network link and gear – both physically and logically – that utilize built-in failover capabilities. Also, the logical security of the networking firewalls, DNS and physical security needs to be understood and evaluated. In the future, with the advent of 5G connectivity, remote edge data centers may be mesh network connected, able to link to multiple sites and failover if one is lost, which will eliminate physical networking media concerns. But until then, know and ensure physical edge networking connections are secure.
4. IT Infrastructure – Closely linked with design and build of the data center and certification of the components used, is the ability of the IT infrastructure to maintain both power and cooling at all times. Now, there are many ways this can be accomplished. Power can be provided via redundant utility feeds or onsite generation and storage. Careful selection of the power infrastructure is critical to supporting all elements of IoT edge data center. Cooling also needs to be carefully considered, which means adequately understanding the requirements of IT and infrastructure gear when supported at high load, low load and in all expected ambient conditions.
5. IT Resiliency – The IT servers and storage installed in IoT data centers must also be designed with built-in redundancy and fail-over capabilities, in order to operate continuously for long periods of time with little to no intervention. Data center professionals contemplating their options in this regard should consider ASHRAE class A4 equipment that can withstand higher temperatures and wider temperature excursion than traditional IT gear.
For those dependent upon edge or IoT environments for businesses success, the reliability of the entire system is based upon a thorough design and implementation of each of these five critical component areas. You need to evaluate and begin tracking these components along with your enterprise or cloud data center implementations to help ensure a reliable edge computing environment for your enterprise.
Todd Traver is vice president for IT Optimization and Strategy at Uptime Institute