Data is growing but with different requirements from processing power, latency or proximity which require different segments to handle the data. As needs shift, growth in data centers will scale accordingly, due to distributed workloads, heavy and smaller compute density for large and midscale companies, and transit needs among these points. This robust demand for specialized compute includes such technologies as augmented reality, virtual reality, intelligent manufacturing, gaming, and more, each with their own needs.
A rapidly growing use of compute at the edge involves augmented reality (AR) and its more in-depth alternative, virtual reality (VR). Augmented reality involves the addition of software inputs to the real world, such as glasses or mobile applications that provide information overlays and labeling, GPS location-based assistance via similar hardware, or the more prospective use of touch screens and projection applications in intelligent manufacturing. Virtual reality is the more immersive version of these applications, with several firms experimenting with headsets to fully involve the user in a software-based environment. While gaming has been the biggest use case to date, other expected variations include virtual conferencing, work environments, and remote learning. While the specifics may vary, all uses require low latency compute close to the end user, as location and speed of use change rapidly. One likely outcome involves smaller data centers located in dense urban areas to ensure rapid traffic to users, with high rack density to offset a lower overall footprint. Greater density will also require more intelligent cooling, with liquid and immersion systems gaining interest for these specialized applications.
While augmented reality and virtual reality may span the working and gaming worlds, the Internet of Things (IoT) also leads to continued workloads from unlikely locations. While an attractive buzz-phrase, IoT simply involves placing chips in previously un-connected devices, from home appliances such as refrigerators and ovens to heavy industrial applications in robots and assembly lines. In home use, these appliances (and security systems, lighting, and so forth) are now accessible on a homeowner’s smartphone, often through a specialized application which in turn needs its own increased bandwidth. A mix of infrastructure solutions is involved, including local cellular networks for the user when away from home, local metro data centers for quick response, and potentially remote data centers for dense processing and analytics. For manufacturing, local in-building networks provide alerts for immediate stoppages or repairs, with heavy analytics processed off-site. All applications are likely to benefit from utilizing a major cloud platform for roll-out, scalability, and uptime.
While the edge is particularly exciting for these use cases, expect growth in many new technologies for entertainment, transportation, manufacturing, and more, as use cases such as augmented reality spill over to other industries. As data consumption grows, capabilities across data centers, fiber, and other infrastructure must scale efficiently.