The last mile in the race to the edge
As previously covered on Techerati, both enterprises and SMBs are increasingly deploying compute at the edge of the network. According to Vapor.IO’s latest State of the Edge report, the shift to the edge is being driven by an increased need for low-latency data processing, an upsurge in remote devices churning out data (175 zettabytes by 2025 to be precise) accompanied by limited bandwidth, and the increasingly impractical cost of cloud services.
In the presence of this growing and lucrative opportunity, a raft of companies are vying to lead the market in the various components of edge infrastructure, which can broadly be divided into edge servers, micro data centre housings and power and service resilience. One such company, largely focusing on the latter segment, is US-based critical infrastructure specialist Vertiv (formerly Emerson), which has a core edge product set comprising a VR Rack, PDUs, a UPS and ACS. Speaking to Techerati ahead of Data Centre World Hong Kong, taking place at the HKCEC May 22-23, Jason Christie, managing director at Geist Asia for Vertiv, discussed the latest developments in the critical infrastructure specialist’s edge strategy.
Edge computing is an evolving and complex concept. Sure, the edge identifies the single perimeter around the enterprise data centre network where a growing amount of compute is being deployed, but edge deployments are manifestly different by virtue of the application they are serving. Christie says Vertiv is seeing hundreds of use cases emerging as IoT devices, sensors, machine learning begin to prove their worth in sectors such as retail, healthcare, manufacturing, and across entire cities.
For a company vying to ‘lead the world’ in edge computing, these modalities must be explored for commonalities, so as to clarify the contours of a roadmap. Through its edge research framework, Vertiv has identified four distinct archetypes: Data-Intensive (e.g. smart building management), Human Latency Sensitive (e.g. smart mirrors and display signs for retail), Machine-2-Machine Latency Sensitive (e.g. smart home sub-units) and Life Critical (e.g. autonomous car and medical equipment safety).
“These four archetypes represent and establish the basis for how we evolve and deliver our strategic roadmap and vision,” Christie says. “Most recently, we have followed up our research by outlining the trajectory, impact and infrastructure requirements of 5G, specifically as it relates to edge computing.”
Common to most edge deployments are their remote, distributed nature, and that they are often unmanned and located in harsh environments. Just like at the core of the network, power and service is of high-importance: infrastructure is needed that is highly autonomous, fail-safe and connected. But the scale of edge sites brings new customer demands and requirements. Christie says Vertiv is looking to dominate the segment by offering globally available, highly integrated and modifiable (yet standardised) systems.
“It starts with global platforms/systems that deliver repeatable form, fit and function anywhere in the world to increase reliability and familiarity within the operations team, secondly it’s rapidly configurable systems based on standard components to solve for potentially 1000s of different deployment scenarios and thirdly it’s ensuring that [edge sites] operate and communicate as a system, not individual components.”
Some applications are more compelling than others. Vertiv is heavily invested in retail, financial institutions and high tech manufacturing, but is waiting on the more advanced (yet headline-grabbing) use cases, such as autonomous vehicles and augmented reality, until the business cases and models reach maturity. “Significant investments won’t happen until the buyer and the seller can be clearly identified,” Christie says.
Another common factor in the growing volume of edge sites is their mission criticality, reinforcing the need for power and service resilience. But Christie clarifies that this does not mean they will be operated and supported in the same way as data centres:
“An edge site will typically be remote and potentially unmanned or not staffed with technically experienced individuals,” says Christie. “So the mission critical nature here is to employ self-healing and internally redundant systems that do not need frequent service truck rolls, can be diagnosed from remote and have some level of predictive failure analysis integrated.”
The average content consumer will not notice the physical foundations of this tectonic shift towards distributed IT. But they will notice their everyday online and mobile experiences reach new levels of speed and sophistication. The likes of Disney and Netflix, that will essentially use these edge sites as colocation facilities, will be able to distribute 4K content at unbeatable speeds. Bolstered location services will enable users to locate their phones to the nearest millimetre. In some cases, such as smart cities, the benefits of speed, precision and bandwidth will converge to ease congestion and improve security:
“The ability to perform real-time or near real-time actions and analytics like facial recognition is game-changing when analysing tens of thousands of faces and matching to dozens of external databases, proving security at entertainment or sports venues without creating massive lines,” Christie says.
Last mile in the race to the edge
For infrastructure providers, the pace of the edge war is accelerating as compelling use cases consolidate on the horizon. A key battle is the one to conquer the power and service resilience segment with edge-focused integrated products. Vertiv has the resources and pedigree to mount a muscular charge, but it faces considerable competition in the form of Schneider Electric. Both have long-predicted that enterprise computing will become more distributed and have altered their business strategies accordingly. Only time will tell who has been more trenchant in their preparations.