Posted on Reading Time: 4 minuteson
MEF recently hosted a webinar on How Edge AI Services Benefit the Retail Sector, featuring presenters from AT&T Business, NTT Communications, Sparkle, and analysts from GSMA Intelligence, STL Partners, and AvidThink. The webinar examined how retailers could benefit from edge computing and AI. I had the pleasure of representing AvidThink.
- Will Eborall, Assistant VP, Edge Solutions Product Marketing Management, AT&T Business.
- Yasuhiro Takenaka, Senior Manager of IoT platform Services, NTT Communications.
- Daniele Mancuso, Chief Product Management, Sparkle Group and Director, MEF.
- Roy Chua, Founder & Principal, AvidThink.
- Dalia Adib, Principal Consultant, and Edge Computing Lead, STL Partners.
- Sylwia Kechiche, Principal Analyst, IoT & Enterprise, GSMA Intelligence.
- Kevin Vachon, MEF COO.
The webinar included highlights from MEF’s February 2021 Proof of Concept Showcase demonstrations from MEF members, insights from the analysts, and a live Q&A with AT&T, NTT, and Sparkle on edge computing services. As the audience learned, there are significant opportunities for edge AI. However, service providers and partners must do an equal amount of heavy lifting to bring these services to market.
To get the full experience, watch the full How Edge AI Services Benefit the Retail Sector webinar on demand. In the meantime, I’ll try to whet your appetite by conveying a few key takeaways in this short article.
Different Edges for Different Retail Situations
There’s not a single edge but a continuum of edges that represent distributed cloud locations. To keep things simple, there are two major edge locations of importance to the retailer:
- The network edge (within the carrier’s network), and
- The customer premises edge (at the retail/warehouse location).
Network edges are usually located in metropolitan areas with high population or business density. Retailer access to these network-edge locations can be via 4G LTE, 5G mobile connections, or a wireline (fiber) connection.
Depending on the retailer, they may choose to use the network edge when customer premises are constrained (small footprint), ephemeral (pop-up stores), or otherwise inappropriate to host on-site computing equipment. On-premises are an option when edge computing needs the lowest latency and reliability, independent of the vagaries of the wide area network (WAN). The webinar covered AT&T’s perspective on how to best approach the edge selection decision. Daniele of Sparkle also shared extensive views of the multiple types of edges (telcos, on-premises, and hyperscalers) and their role in retail.
Regardless of edge type, fellow analyst Dalia from STL observed that edge computing brings value in multiple use cases. In retail, especially, given the limited amount of time to impact a customer, having real-time interactions is critical. The edge provides that low level of latency and high level of interactivity needed to improve the customer experience.
Other use cases for AI at the edge include customer queue management, inventory scanning and management, customer self-service recommendation engines, and many IoT applications.
IoT Applications Key to Edge for Retailers
Speaking of IoT, multiple panelists made it clear that IoT is a crucial element of retailer strategies. Fellow analyst, Slywia from GSMA shared that 63% of retailers have adopted IoT as part of their digital transformation. And 52% of retailers view latency as a key to unlocking transformation and achieving value through IoT. IoT connects the physical and digital world for retailers, and the use of AI and machine learning (ML) at the edge can extract real-time insights from that IoT data.
Will Eborall of AT&T presented PoC 133 SD-WAN and 5G with Network Slicing, powered by AT&T Business, Ericsson, and VMware, which demonstrated a retail grab-and-go application. A combination of IoT technology with fast, local, edge processing enabled a convenient and seamless self-service customer experience. One could imagine extending this PoC to use edge-powered AI computer vision to determine which product had been selected as part of the purchase.
Orchestration is Critical
While the demonstrations and proofs of concept looked impressively simple, the reality is that there’s significant work behind the scenes to bring edge AI services to the market. Wenyu Shen of NTT Communications shared the details of a PoC 131 Orchestrated Multi-Edge Clouds, demonstrating the power of orchestrated multi-edge clouds. This PoC brought together NTT Communications, Netcracker, Juniper, Microsoft Azure, and ADVA. It showed how a service provider could orchestrate an end-to-end platform from customer premises to the cloud, across a resilient and optimized virtualized network (SD-WAN), running on top of multiple physical connections (WiFi, 5G, MPLS). The demo showed a unified application architecture that stretched from universal CPE edge platforms (on-premises) to public clouds.
In addition to the orchestration of the underlying platforms, there will be complexity in tying in the varied vertical retail applications. Yasuhiro shared NTT’s perspectives on how the carrier approaches the go-to-market decision across different geographic regions and his thoughts on vertical integration delivery versus a horizontal platform approach.
Coordination and Standardization Accelerates Service Deployments
No matter what the go-to-market, the end goal for retailers is to improve the customer experience. Edge services offered by the carrier, incorporating AI/ML, and tying into IoT, can be a significant digital enabler for these retailers. However, to get to where retailers can consume edge services without hassle, there needs to be a coordinated effort between all the partners in the ecosystem. The multi-edge clouds PoC 131 uncovered the large number of players necessary to bring these edge services to market.
MEF can play the role of orchestrator across these ecosystem companies and act as a hub where multiple parties unite to make these PoCs a reality. It provides a forum to share experiences and learn from each other. Daniele shared his views on the importance of MEF and the various standards that MEF members are currently working on orchestration (LSO and more) to SD-WAN and the edge.
This PoC was presented at the MEF PoC Showcase in Q1 2021. Watch the Presentations on YouTube
About MEF’s PoC Program & Showcase
The MEF 3.0 Proof of Concept program effectively fosters innovation, seeds new MEF standards and projects, and accelerates our existing work within the ICT industry by providing our members—including service providers, technology suppliers, and other stakeholders within the ICT industry—the opportunity to collaborate on MEF 3.0-based use cases throughout the year.
Initiated by MEF members and facilitated by MEF staff, each MEF 3.0 PoC receives a unique, identifying number that remains unchanged as its title and messaging evolves over the life of the project.
PoC work is highlighted in public showcases and award presentations that explore individual Proofs of Concept. Learn more about these enabling technologies and the MEF 3.0 PoC Program.