Verizon pushes the boundaries of 5G edge computing with MongoDB
Today, mobile devices are more powerful than they’ve ever been and as a result, they’re demanding more from the network they’re connected to. But it’s not only mobile phones asking more of the network. Whether it’s a smart stadium full of fans or a 5G-enabled digital billboard, the Internet of Things (IoT) means every endpoint needs access to the right data at the right time. The next exciting challenge in the story of 5G is making sure there’s a network that can deliver this.
That’s why the network edge has become one of the most explosive cloud computing opportunities in recent years. As mobile contactless experiences become the norm, and as businesses move ever-faster to digital platforms and services, edge computing is positioned as an oftentimes faster, cheaper, and more reliable alternative for data processing and computing at scale.
Traditionally, cell data is routed via cloud regions in the United States and, no matter where you are, your data will be routed through one of them. The challenge is that even though 5G promises faster speeds and lower latency, the infrastructure itself – and its limited proximity to end users – thus acts as a bottleneck. If companies are rolling out ever-richer apps, such as augmented and virtual reality, the data connection needs to be more dynamic. Verizon wanted to enable data that can follow the user on the edge and avoid the need to go through one of the regional zones, thus eliminating the bottleneck.
“Verizon 5G Edge is a mobile edge computing platform, which embeds popular hyperscaler compute and storage, such as AWS Wavelength, at the edge of our 4G and 5G networks so application builders can extend existing workloads using the same popular services they know and love,” explains Robert Belson, Principal Engineer, Corporate Strategy, Verizon. “However, certain services, such as databases, are not natively supported, which is where Atlas and Realm come into play by creating unprecedented flexibility for the developer and the end customer.”
By embedding AWS compute and storage services in data centers of Verizon, AWS Wavelength extends its infrastructure to Verizon networks at service access points on the edge of the 5G network. This approach enables traffic to reach application servers running in Wavelength Zones without leaving the mobile network, reducing the latency that results from multiple trips to the Internet, and helping developers unlock the potential of 5G for their applications.
“Our journey began when we installed MongoDB Community on a virtual machine at the edge of an AWS Wavelength zone and measured how fast it would take to send and receive data, compared to the existing network,” adds Belson. “And what we found in that race with over a thousand examples was a 47% improvement.”
To build an enterprise-ready reference architecture, the team determined that a comprehensive data platform was needed. Atlas Functions was integrated with the Verizon edge discovery service to help direct 5G mobile clients to the topologically closest database instance across a customer’s edge deployment.
At the same time, Verizon is overlaying a data persistence layer with Realm, helping enable ultra-immersive, personalized experiences with the seamlessness to extend to the network edge. Together, they have created a reference architecture with full automation which can be stood up in minutes at each of the 19 AWS Wavelength Zones.
Verizon is also using Atlas Device Sync and Realm to ensure the seamless synchronization of data between devices, the cloud and edge of network, online and offline. To ensure that the edge persistence layer could complement existing MongoDB Atlas deployments, regardless of size, Verizon selected Atlas Device Sync to abstract away the complexity of conflict resolution and cluster linking: “We can take a fleet of virtual machines across all 19 AWS Wavelength zones and run Realm on each of those instances with Device Sync. What I love about Device Sync is, as a developer, I don't have to deal with conflict resolution. It's just done for me,” says Belson. “And we think that’s an incredibly powerful tool for developers.”
When fully rolled out, data will follow consumers according to location, not which of the zones they are tethered to, keeping latency to a minimum. At the same time, enabling data analysis on the edge, rather than performing it centrally, will reduce networking and storage costs. Additionally, downtime in one zone can be easily mitigated, therefore maximizing overall availability. All these benefits set the stage for the next generation of rich apps for consumers and in sectors such as manufacturing, logistics, and agri-tech.
For example, fleets of robots equipped with 5G modems could seamlessly upload data in real time to MongoDB at the edge where it can be processed and analyzed for anomaly detection, fleet analytics, and other safety use cases. One use case which has Belson particularly enthusiastic is an app for American Sign Language. Sign-Speak is bridging communication gaps for the Deaf and Hard of Hearing communities by translating American Sign Language into the spoken word — and vice versa — using machine learning.
“I'm really excited about the work that we're doing with the Sign-Speak team to bridge the communication gap for the deaf and hard of hearing community by having a real-time American Sign Language Conversion Tool. This is a powerful mobile app that we can bring to MongoDB on the edge to help not only run those machine learning models, but also take care of all the model updates and metadata changes while running all of that in the context of a Kubernetes cluster on Verizon 5G Edge,” comments Belson. “So, you have a whole tech stack at play, you have the data, you have the compute, you have the end fleet of devices and the cloud itself. And we think those are the use cases that are most exciting, extending an existing architecture to the edge to boost its performance.”
Together, MongoDB and Verizon are unlocking a next generation mobile application experience and creating a force multiplier for innovation, with the goal of extending the life cycle of devices and creating a return on investment for enterprises who harness the power of the network edge.
“So what better way to build applications at the edge than working with MongoDB to make software development easy while we make the network easy? The future of Verizon 5G Edge and MongoDB has never been brighter. We're focusing on developer education. We're doubling down on working with customers,” concludes Belson. “Most importantly, we're taking advantage of the entirety of MongoDB’s portfolio, alongside our network APIs, to create a more uniform application experience for mobile and IoT users alike. We're super excited to continue our collaboration with MongoDB and look forward to taking advantage of data at the edge.”
Robert Belson, Principal engineer, Corporate Strategy, Verizon