The race to the edge
By Samir Shah and Michael Zapata
Race to the edgeEvery data user, whether corporate or consumer, wants content delivered faster. One way to address this has been through the proliferation of edge data centers, spurred by such developments as the growth of connected devices, increasing content distribution and the Internet of Things (IoT).
Designed to bring IT resources closer to where it is being demanded from, edge data centers work to deliver smarter, faster, stronger computing capabilities. In an ideal world, edge data centers should improve connectivity and speeds for users in any urban location. But the ubiquity of video sharing and greater developments in the Industrial Internet of Things (IIoT) are causing greater needs around that connectivity. This article will cover the requirements edge data centers need today to survive the continuing growth of innovation.
Innovation driving content at the edge
Virtual Reality (VR) video requires high-speed, seamless transition of data-dense content. Five to 10 years ago, adoption of Virtual reality (VR) could have been chalked up to the likes of famous film production companies or science fiction. But today, its user-friendly capabilities and widespread adoption via smartphones translates into more VR development and essentially, more demand on data centers to be able to support VR video. In fact, worldwide revenues for the augmented reality and virtual reality (AR/VR) market are forecasted to grow to more than $162 billion in 2020 from $5.2 billion in 2016, with the U.S. in the top 3 global regions driving the growth1.
Existing challenges with 4k VR looking blurry or low resolution due to scaling in a 360-degree viewing area are causing VR platforms, including YouTube, to move toward 8k VR content. This results in a network impact that is four times greater than that of 4k VR videos (8k VR video is comprised of four times the data/pixel count as a 4k stream). As a result, edge data center distribution points will be needed closer to end users to prevent latency and slow performance,
Also challenging is IIoT, which is much more mission critical and data intense than IoT. According to Accenture, IIoT is “a network of physical objects, systems platforms and applications that contain embedded technology to communicate and share intelligence with each other, the external environment and with people2.” The company estimates that IIoT could add US$14.2 trillion to the global economy by 20303. In short, IIoT controls the infrastructure at manufacturing and warehouse sites and allows the network to manage and correct processes based on data received from generators, switch gear, cooling equipment, security systems and other equipment.
Such scenarios all drive the race to the Edge. As a result, edge data centers will require the proper infrastructure and network connectivity to be able to support the growth of data use.
Understanding the innerworkings at the Edge
Most enterprises are not migrating their IT strategy solely to a Cloud model but rather are spreading their applications across a multi-step, data center approach. To identify how edge data centers can support a business’ needs, it is important to understand the four basic building blocks to an Edge platform.
The first is data collection. Here, connectivity throughout the network allows the user to collect and organize relevant data. The next step is visualization of that data. Here, management is informed about what is happening throughout operations and is able to analyze that data and take action. With the capabilities of The Edge, much of this will be able to happen without ongoing human intervention. This leads to block three, control. Automation of data handling at The Edge allows optimized workflow performance. Of course, knowing what might happen in the future based on historic data is a handy management tool. And that is the fourth aspect of an Edge platform, modeling. With predictive algorithms, management can simulate various conditions and generate intelligence on how any change will affect operations.
As data centers move to The Edge over time, they must take into account the efficiency of their footprint. Will they be brick-and-mortar or modular? How will an edge data center meet sustainability and reliability demands differently? What does it need to incorporate given its urban relocation? Will its entire technology be conducive to software-defined monitoring and management?
Edge operators of today
Consider companies like Netflix, Amazon, Google, Facebook, Microsoft, AT&T and the like. Most all are end users and Content Delivery Network (CDN) providers in their own ways. They must invest in The Edge in order to provide content/services (like streaming) to be more effective in remote areas like Tier 2 markets.
A walk through any data center’s infrastructure area can reveal IIoT areas of deliverance. There is potential to link metering, thermostats, parking lot gates, sensors, security, cooling and power to IIoT. This allows end users to leverage IIOT and aspects of a software defined network (SDN) to more efficiently build out their Edge network.
The next step to turn all of this data into usable information is collecting and analyzing it to allow stationary engineers to make smarter decisions with infrastructure. This depends on computer network engineers making smart decisions on when, where, and how to deploy their Edge network.
One secret to success will be standardization across the network. This platform will be able to collect upstream data close to the appliance that is generating it, send it to the cloud on high-speed fiber pipes for instant analysis, and speed it back from The Edge to the end device. It will allow users in Tier 2 or Tier 3 markets all the advantages of Tier 1 access.
Today’s data center deployments at The Edge can also consist of both power and data (IT) modules that are factory-configured and tested. With the proper module control unit, deployments have access to real-time module sensors.
The secret to success will be standardization across the network. It will take expertise to bring the technology that will optimize the network on a platform at The Edge. This platform will be able to collect upstream data close to the appliance that is generating it, send it to the cloud on high-speed fiber pipes for instant analysis, and speed it back from The Edge to the end device.
The end game
Regardless of whether edge data centers manage data from IIoT applications or consumer’s latest episode of Game of Thrones, they will need to identify how to provide all users smarter, faster, stronger computing capabilities across all operations and multiple geographic locations. The end-game will be to provide users with reliable insights into their processes on a software-defined network that is agile, sustainable, secure, and a physical platform that is consistent and scalable.
1. International Data Corporation (IDC), “Worldwide Revenues for Augmented and Virtual Reality Forecast to Reach $162 Billion in 2020, According to IDC,” Aug. 15, 2016
2. Accenture, “Accenture estimates the Industrial Internet of Things (IIoT) could add $14.2 trillion to the global economy by 2030
3. Accenture, “Winning with the Industrial Internet of Things,” 2015