Presented by Intel
You’ve dreamed about them: Jitter-free AR/VR. Local inferencing for machine learning. Personalized in-store offers for individual shoppers, in real time. And much more. With new advances in 5G and edge technology, it’s time to start planning proof-of-concept pilots.
Launches of next-gen commercial service over the next 24 months promise to bring ultra-low latency and high bandwidth to mobile networks. But for all its speed, 5G alone might not always be fast enough for AI, IoT, and other data-intensive applications forced to traverse distant data centers and clouds. That’s where edge comes in.
A game-changing combo
Combining the two technologies makes it feasible to bring compute, storage, networking, and analytics closer to where data is created and used. 5G and edge will enable the delivery of enormous amounts of bandwidth and compute capabilities and with ultra-reliable low latencies to a user, application, or device like a factory robot, video surveillance camera, or a gaming headset to play the latest release of “Beat Saber.”
That kind of local processing brings several advantages. It eliminates the need to shuttle vast amounts of data across sprawling networks – a huge potential time- and cost-saving in backhaul and service expense. Pairing 5G and edge also opens the door for a host of new services for retail, media, banking, security, smart cities, manufacturing, and many other industries.
Local intelligence drives innovation, 10x market growth
That’s one big reason why the adoption of 5G is forecast to help drive a 32.8% annual growth in the global edge computing market over the next five years. Fueled in part by a geometric increase in edge-connected devices, Gartner estimates that 75% of enterprise data will be created and processed outside centralized data centers or the cloud by 2025.
With commercial 5G dawning, it’s not too soon to get serious about thinking of ways to use this “better-together” duo in your organization. Here are five reasons why:
1. Get clear on the concept and benefits
It’s ironic. While “edge” is one of today’s most talked-about concepts, it’s often vaguely understood. No standard definition has emerged, so “edge” tends to reflect the products, businesses, and orientation of the speaker.
Despite the absence of a precise taxonomy, all definitions share a general theme – moving compute and storage resources
closer to the point of service delivery or origin of the data to deliver the highest return. You can envision it like this:
There are two main types. The on-premise edge locates compute nodes or edge platforms on-premise at an enterprise. The network edge is located off premise, hosted in a localized network or cloud node. The latter is often made up of edge platforms or nodes on wireless, wireline, or cable networks, and in the communications service provider’s central office.
Any of these edge locations can host network and IT-based services as well as virtual network functions. These locations can have additional edge data centers or servers called multi-access edge computing (MEC) platforms to host services.
As for benefits: Broadly speaking, 5G and edge ultimately help unlock the value of data by moving, storing and processing it closer to endpoint devices. Doing so promises to reduce application latency, facilitate rapid data transfer, accelerate service delivery, strengthen data sovereignty, and improve the quality of the user experience. It also should make it easier to strategically locate Artificial Intelligence applications and services where they make the most technical and economic sense, garnering insights from the data being processed at the Edge.
2. Start thinking about business and use cases
Edge and 5G are leading technologies. But classic rules still apply. “At the end of the day,” Navale reminds, “it’s the bottom line that counts.”
Determining just how 5G and edge can bring business value is a basic but crucial first step. Criteria will vary, of course: A communications service provider, IoT service provider, a cloud service provider, and the typical enterprise will have different needs. Where to start?
Planning for edge is a multi-variable equation; it requires balancing the laws of physics, economics, and governments with real-time determinism, performance, power, networking, price constraints, and a host of other factors. The complexity can seem daunting. Yet in the early stages, simplicity helps get the ball rolling.
The strongest business cases start with basic questions. Specifically, identifying applications and workloads likely to yield the greatest benefits from improvements enabled by edge and 5G. Prominent among them:
- Speed: “Where can we support latency-sensitive applications and act on data insights in near-real-time?”
- Scale: “Where can we prep for agile growth by expanding capacity without opening a new data center?”
- Agility: “Where’s the most promising opportunity to easily deliver new services? Or target and shift between localized markets?”
- Reliability: “Where can we best leverage always-on data access and enhanced security, and reduce data in transit?”
What’s your version of a smart beer tap?
Another helpful early-stage exercise: Familiarizing yourself with a growing range of applications taking advantage of the low-latency, ultra reliability and high bandwidth of edge and 5G. They include:
- Industrial: Predictive maintenance, automated robotics, safety and security, smart defect detection using AI inferencing, and process control
- Retail: Multi-factor AI inference, opportunistic pop-up stores, predictive inventory management, and personalized shopping, signage, and POS in real time
- Media: Seamless cloud gaming, 4K, 8K and volumetric visual-workload processing, lifelike immersive media, video content streaming, Content delivery from the edge that improves the user experience
As always, looking both within and outside your industry can yield exciting possibilities worth exploring. Some might be edge versions of innovative real-time cloud apps like this smart beer keg. In all, Navale advises realism in planning. “There may be cases where your TCO is still better in the cloud,” she says.
“But,” she hastens to add, “edge is so new, we don’t know what will be created. Remember: a few years ago, when the iPhone and the App Store appeared, we never knew that Uber was going to be invented. What new edge services will drive economic output or value to the supply chain? Right now, we’re only scratching the surface.”
3. Plan with the (new) big picture in mind
Advances in 5G and edge are part of a larger transformation: the ongoing, global “cloudification”, convergence, and consolidation of networks, workloads, and service providers. These tectonic shifts promise new technology and vendor choices for handling data-centric services and 5G-era workloads.
Service providers and enterprises alike are accelerating adoption of virtualized cloud architectures, models, economics, and technologies. Observes Navale: “Major, multiple convergences are taking place all at once – networking, edge, 5G, IOT, the mobile core infrastructure, and media and IT.” Their common goal: no less than reinvention of both private and core mobile networks.
In the process, traditional lines are blurring between enterprises, comms service providers, and cloud service providers. AT&T’s collaboration with Microsoft Azure and AWS is a good example. “We’re starting to see these companies all coming together,” Navale notes.
Against this backdrop, smart planners will be wise to stay on the lookout for new, better ways to get new and old jobs done.
4. Look for “Easy” buttons
While the possibilities of 5G and edge are exciting, the complexity can seem overwhelming, like 3D chess. Industry “easy buttons” can help organizations build and deploy the technology combo more simply and quickly.
For example, Intel has taken a leading role in developing services and solutions to help manage edge and 5G complexity. It offers pre-integrated, pre-verified, and optimized solutions for specific edge workloads as part of the Intel® Select Solutions program, as well as network optimized CPUs available through its second-generation Intel® Xeon® Scalable processor family. Intel also has invested heavily in specialized software platforms, middleware, open and inference toolkits, and a global network developer ecosystem for edge and 5G.
According to Navale: “All of this is geared toward making sure that we have the best possible type of platforms that enable our customers to optimize their TCO, their time to market, and run their edge workloads in the most performant per dollar, per watt optimized manner, without having to deal with all the complexity of 5G or LTE.”
Big communications service and cloud providers are building on these platform capabilities and optimizations to offer customers similar easy-rollout packages and solutions. For instance, the Akraino Edge Stack project, hosted by the Linux Foundation, which is working to simplify software integration for edge computing included participants Intel, AT&T, Ericsson, Nokia, Huawei, Red Hat, and many others.
5. Don’t wait – start with 4G
While 5G makes edge more powerful, it’s not required. Edge computing can be done on today’s 4G networks to increase the responsiveness of ever demanding applications like AI.
A trial by China Unicom, Intel, Nokia, and Tencent Cloud provided proof of concept for edge deployed on 4G networks. New offerings like the MEC solution from Nokia, powered by the Intel Xeon Scalable processor family, enable 5G-like services using today’s 4G infrastructure.
Edge computing promises to be one of the most exciting, transformative uses of 5G, enabling a new wave of business and societal innovation, especially with AI. Growth in intelligent apps, cloud infrastructure loads, and swelling data are poised to drive ten-fold growth, to a $16-billion annual market, by 2025. Network functions virtualization (NFV) and software-defined networks (SDNs) were just the beginning; the next frontier is at hand. Start smart to get an edge.
Sponsored articles are content produced by a company that is either paying for the post or has a business relationship with VentureBeat, and they’re always clearly marked. Content produced by our editorial team is never influenced by advertisers or sponsors in any way. For more information, contact [email protected].
Source: Read Full Article