The deployed global power footprint of the edge IT and data center facilities is forecast to reach 102,000 megawatts by 2028, with 68 percent of the deployments being on the infrastructure edge and the remaining 32 percent on the device edge.
source: State of the Edge report
The by now classic need for edge computing is latency performance for applications and use cases operating at machine speed, not human speed, at huge scale, on a highly-automated basis and involving a mix of media including compute-intensive virtual and augmented reality.
A drone flying at 60 miles-per-hour travels 100 yards in four seconds. Avoiding collisions may require decision-support from an edge server and a delay of 100 milliseconds could cause it to crash into something 10 feet away, for example.
source: State of the Edge report
One might say the rise of edge computing is a deepening of the content delivery network widely used to speed up application performance.
Of course, there are “many edges,” and one particular form is of interest to connectivity providers. Computing done directly on devices, by enterprises on their own platforms, on premises, or locally by third parties are not directly business opportunities for most connectivity providers, except as use of those devices requires wide area network and local access connectivity.
The infrastructure edge, on the other hand, might directly involve the connectivity provider as a supplier of facilities, full applications or required WAN access services, thus creating a direct new value proposition, role and revenue source.
Potential roles include supplying real estate (rack space, power, security, air conditioning), private network access or actual applications and computing services.
No comments:
Post a Comment