Edge Computing is all about a distributed-cloud architecture designed to reduce latency and increase reliability by storing and processing data closer to the applications that use it. This simple and straightforward concept enables innovations like smart factories, robotic surgery, immersive real-time AR/VR, and advanced multiplayer online gaming – use cases that cannot suffer outages or slowness, and that require high guarantees of speed and availability. Emerging services from AWS and Verizon are bringing the concept to reality, offering zones that support deploying Couchbase on standardized AWS infrastructure at the edge – beyond the traditional Region-based data centers, including onto customers’ premises. Storing and processing data at the edge provides benefits for applications across industries like manufacturing, telco, healthcare, retail, media, and more.

In a fascinating session at our recent Couchbase ConnectONLINE developer conference, panelists from AWS, Verizon, and Couchbase sat down to talk about how Edge Services, 5G, and local data processing work together to enable ultra-low latency for mission-critical apps and services.

Panelists included Brent Eiler, Principal Architect 5G Edge at Verizon, Rob Czarnecki, Principal Product Manager AWS Outposts, Savi Venkatachalapathy, GTM Specialist AWS Wavelength, Pranav Chachra, Sr Product Manager AWS Local Zones, and rounding out the panel was our own Wayne Carter, VP of Engineering at Couchbase.

The session was fast paced and informative, starting off with Wayne explaining how edge computing provides the highest possible speed, reliability and security for applications by moving data storage and processing closer to the clients that use it. With easy to understand diagrams, Wayne described the cloud, edge and client layers in the edge architecture and how they work together to store, process and replicate data, providing high availability and ultra-low latency.

He also covered Couchbase Engineering’s latency tests on AWS Edge Services, in partnership with AWS and Verizon. The tests measured the difference in overall response times when accessing a web-service backed by a Couchbase database deployed in an AWS Availability Zone vs AWS Wavelength Zone, and in an AWS Availability Zone vs AWS Local Zone. The test results were impressive, and proved out the claims of ultra-low latency and single-digit millisecond response times to end-points, showing up to an 82% reduction in latency on Wavelength and a 78% reduction in latency on Local Zones.

With the Couchbase test results putting to rest any disputes about the ability to achieve ultra-low latency at the edge, the discussion shifted to the edge services and their typical use cases.

AWS Edge Services

Savi began on behalf of the AWS team by describing the AWS Cloud Continuum, explaining that organizations are accelerating their digital transformation and want easier development, faster innovation, and efficient scale, and so are looking to the cloud to achieve these things. 

But in some cases, workloads cannot move to the public cloud entirely or right away, due to factors such as industry or region-specific compliance and data sovereignty needs, low latency or local data-processing requirements, or because they need to run close to where the data is consumed. Those applications need to reside at the edge – in on-prem data centers, edge nodes in large metro areas, restaurants, factories and the like, including the edge of the 5G network. But customers who are running at the edge still want the same familiar and trusted AWS services and APIs that they get in the cloud, and for them the AWS Cloud Continuum addresses this directly by extending the cloud with AWS Edge Services. The offerings bring AWS services and infrastructure to wherever the customer needs it, be it on-premises, in a metropolitan edge data center, or on a 5G network. Because AWS is already a familiar environment for many enterprises, AWS Edge Services provide a consistent development, architectural and operational environment for standardized deployment and management. She mentioned that components of the AWS edge services family included AWS Outposts, AWS Wavelength and AWS Local Zones. She also mentioned the AWS Snow Family of edge services, portable edge data centers for remote locales with no internet.

Next the discussion turned to each service specifically – what it does, what makes it different from the other services, and what are some common use cases.

AWS Outposts

Rob explained the Outposts offering, which brings AWS infrastructure directly onto the customers premises. Customers of Outposts typically require an infrastructure that is entirely dedicated to them, for proximity, regulatory or privacy reasons, and with the offering they get AWS services and APIs running on their premises. This allows them to deploy applications wherever they need in order to meet security and ultra-low latency requirements. He emphasized the fact that the Outposts infrastructure runs the same services as on standard availability zones, making migrations easy, development consistent, and maintenance straightforward.

Outposts Use Case

Next Rob talked about a specific use case for Outposts, iGaming. iGaming is essentially online betting, which is becoming increasingly legalized across the country. He explained that the industry is regulated state-by-state, and that local regulators must approve data center locations, because they process bets and transactions. Outposts provide a way for iGaming companies to run data centers where required for compliance, delivering applications in line with regulatory requirements while at the same time providing a low latency experience for the players. And because it’s AWS, the company can more quickly respond to issues and problems because of the familiar and standardized environment. It was a great example of edge computing on Outposts.

AWS Local Zones

Pranav described the Local Zones offering, which provides AWS infrastructure in major metropolitan areas and is designed to serve applications within those areas as an edge data center. He said that in some cases, customers wish to offload owning and operating physical data centers all together, but they still need the benefit of ultra-low latency for users in specific locations, and that’s where Local Zones come in. AWS Local Zones essentially extend standard AWS Availability Zones fro regions to cities, allowing customers in those cities to gain the benefits of AWS for local compute and storage, without managing a physical data center.

Local Zones Use Case

In the discussion, Pranav said that multiplayer online gaming is a very common use case for Local Zones. For gaming companies low-latency is paramount, as anyone who has ever played a game against other online players knows, slow responsiveness is a sure fire way to frustrate gamers. Until the emergence of Local Zones, gaming companies were forced to build and maintain their own co-located data centers around the country, at great cost, to provide the low latency experience they required. Now with AWS Local Zones, these companies are able to quickly deploy edge servers close to gamers on standard AWS infrastructure, removing the expense of operating their own data centers, and delivering a superior gaming experience . 

AWS Wavelength

Savi discussed AWS Wavelength, which brings AWS to the edge of the 5G network and serves mobile applications with ultra-low latency responsiveness. Wavelength is essentially AWS infrastructure that runs directly on 5G providers networks, which means the data center is accessible from anywhere in the network radius without requiring the internet. She explained that without Wavelength, cloud data centers must be accessed outside the 5G network, across multiple hops, which introduces latency. Running data centers on Wavelength directly in the 5G network removes the internet from the equation, reducing latency and providing a superior experience for mobile app users.

Wavelength Use Case

Savi also said that a common use case for Wavelength is in healthcare, where AI/ML-based image and video processing solutions are emerging. One example involves analyzing dense and high velocity image and video data collected by surgical instruments and medical devices, which is then processed in Wavelength-based data centers for real time diagnosis and treatment recommendations. For the hospital, sending this data to a distant cloud data center and waiting for a response is not an option. Real time responsiveness is required for ML and inference, which analyzes massive amounts of data to make predictions and recommendations that help the physician treat patients, and so they leverage Wavelength and Verizon 5G to get the required low latency. The Wavelength example is impactful, and exemplifies a Multi-Access Edge Computing use case.

Verizon 5G Edge

Brent talked about the network as a critical piece of the edge landscape, knitting everything together. He said that Verizon 5G is an inherent part of Wavelength, and an option for Outposts, and is generally used for mobile use cases, where users are not only vast, but also constantly moving from place to place, while still expecting speedy and reliable responsiveness. They meet these expectations via onsite cellular networking, both 4G and 5G, use of multiple sim profiles, the larger Verizon macro-network, and a new offering from Verizon, the Verizon Edge Discovery Service. This API allows developers to quickly integrate edge data center discovery to their apps, essentially directing clients to connect to the closest and most available Wavelength zone on the network to maintain consistent speed and availability. This discovery an essential component for mobile apps as clients move from place to place.

Verizon 5G Edge Use Case

Brent talked about a fascinating use case in airline baggage handling, where a distributed database was critical to the application because of its multi-location aspects. The application tracks baggage in transit using devices like scanners with embedded databases, which then sync data to local edge data centers in each airport, which in turn sync data to cloud data centers, which then replicate to regional data centers and back down the chain at the next airport, tracking and directing baggage automatically across a huge airport network. The data is vast, for each bag there is a passenger, a manifest, a gate, an arrival and departure time, and much more, multiplied by millions of travelers. But the bags must get to their destination and so the application requires ultra-low latency to track their every move in real time. Thanks to Verizon 5G Edge Services, in combination with a distributed database like Couchbase on AWS Wavelength and AWS Outposts, airlines are able to track bags and make instant route adjustments, ensuring bags get to their destination and passengers are happy.

Couchbase on Edge Services

Wayne talked about how the database is just as critical to edge computing as the infrastructure and network. Apps that are stateful – that need to store data – require a database. And that database needs to be able to operate and synchronize across the entire architecture, not just on a single edge device or end point. Couchbase is edge-ready and supports the edge computing architecture with the ability to run anywhere – in the cloud, at the edge, and even on the device. And Couchbase’ built-in secure synchronization ensures data integrity and makes sure that data is never lost. Couchbase fits perfectly into any of the various edge topologies possible with AWS Edge Services and Verizon 5G, and because it is distributed and comes with SQL, analytics, eventing and synchronization as core features, it provides a consistent and easy developer experience. The combination of AWS, Verizon and Couchbase at the edge brings high guarantees of speed, security and reliability for applications.

Couchbase Edge Use Case

Wayne then talked about a Couchbase customer use case in retail that is a great example of edge computing: Louis Vuitton. They leveraged Couchbase to enable employees in their retail stores worldwide to have instant, reliable access to their product catalog and inventory at sub-millisecond speeds. They also gained 100% availability by taking advantage of Couchbase embedded directly to their application client devices. Prior to Couchbase, they were dealing with data locality and latency issues, and a lack of adoption by staff who saw the initial application as unreliable. They decided to use Couchbase to distribute data to the edge devices from global data centers. Because data processing was embedded to the edge devices, the employees were able to access everything about any given inventory instantly, even when the internet faltered, enabling them to provide a superior experience for their customers. And Couchbase’ built-in synchronization makes sure the data on the device is always up to date. With Couchbase at the edge, Louis Vuitton is able to deliver a consistent, superior experience for their customers worldwide.

Wayne wrapped up the discussion by explaining how Couchbase fits seamlessly into all aspects of the AWS and Verizon 5G Edge Services spectrum, and he made the point that embedding Couchbase directly to edge devices brings the highest guarantees of low latency and reliability: sub-millisecond response times and 100% availability.

Watch the full discussion

Thanks to AWS and Verizon’s edge services, enterprises now have an easy and repeatable on-ramp to edge computing, with a variety of options to meet any use case. And when combined with Couchbase, the edge-ready database, customers are able to innovate and build and deploy applications that are always fast and always on.

Be sure to watch the discussion in its entirety here:

Learn more about Couchbase for Edge Computing

Author

Posted by Mark Gamble, Dir Product & Solutions Mktg, Couchbase

I am a passionate product marketer with a technical and solution consulting background and 20+ years of experience in Enterprise and Open Source technology. I have launched several database and analytic solutions throughout my career, and have worked with customers across a wide variety of industries including Financial Services, Automotive, Hospitality, High-Tech and Healthcare. I have particular expertise in analytics and AI, love all things data, and am an emphatic supporter of data-for-good initiatives.

Leave a reply