The final post in this four-part series identifying the main archetypes for edge applications and the technology required to support them is Technology Requirements for Local and Regional Hubs. The archetypes were defined as a result of an extensive analysis of established and emerging edge use cases considered to have the greatest impact on businesses and end users. The full report on the archetypes is available here.
The infrastructure required to support these current and established use cases consists of four layers of storage and compute in addition to the communications infrastructure required to move data between the layers.
At the source, there is typically the device that generates or consumes data and a processing endpoint. The device could be a sensor monitoring anything from the powered status of a lamp, the access to a door, the temperature in a room or other desired information. The processing endpoint may be as simple as the PC or tablet a consumer is streaming video to, or could be the microprocessors embedded in automobiles, robots or wearable devices. These components are application-dependent and are typically designed in by the equipment manufacturer or retrofitted to existing devices.
Every archetype will also require a local data hub, which provides storage and processing in close proximity to the source. In some cases, the local hub may be a freestanding data center. More commonly, it will be a rack- or row-based system providing 30 to 300 kW of capacity in an integrated enclosure that can be installed in any environment.
These rack- and row-based enclosure systems integrate communication, compute and storage with appropriate power protection, environmental controls and physical security. For archetypes that require a high degree of availability, such as Machine-to-Machine Latency Sensitive and Life Critical, the local hub should include redundant backup power systems and be equipped to enable remote management and monitoring. Many uses cases will also require data encryption and other security features within the local hub.
For all archetypes except Life Critical, the local hub will require the ability to connect to a metro and/or regional hub, which will provide longer-term data storage and support capabilities such as machine learning.
The metro hub could leverage the existing telco infrastructure, such as underutilized central office facilities, to support the local hub with longer-term data storage and more robust processing capabilities. The regional hub is likely to be a cloud data center operating in the same region as the local hub.
For both the metro and regional hubs, modular designs capable of easily scaling beyond the initial design spec should be considered to account for unexpected surges in demand. These facilities should also be designed to scale in terms of density. Image-intensive applications, such as virtual reality, and processing-intensive applications, such as analytics and machine learning, will likely require rack densities that exceed the typical 10 kW design specification. In virtually all cases, these hubs should provide the same or higher level of redundancy and security as the local hub.
The Next Step
The changes in the compute and storage infrastructure required to support the connected future, particularly at the local level, will be deep. The four archetypes described in these posts will help guide IT decision makers determine the infrastructure and configuration requirements for the many applications discussed as well as those that will emerge in the coming years.