Google Cloud as Enabler for Enterprise Edge/P5G Networks

Introduction

As Telcos are rolling out 5G network for Consumers and Enterprises, there is growing need for building Edge data centers (DC) for future catering to the requirements of many industry segments like Manufacturing, Retail, Supply Chain, Energy (Utility towers, wind farms, oil rigs,etc ), Smart Cities, Healthcare, Gaming, Farming, etc. Telcos are also planning real-time computing possible in locations where it would not normally be feasible and reduces bottlenecking on the networks and data centers that support edge services.

5G brings a big change in data consumption and growth. The huge volume of data generated by edge devices would overwhelm most Telco networks, hampering all operations on their network and their costs could skyrocket. Telco business priorities have always been market share and they don’t want dissatisfied customers moving their business elsewhere.

While most of the compute, storage and networking requirements can be fulfilled from the GCP public regions, there are multiple valid reasons for Telcos to build Edge Networks:

• Latency — It does matter for many of the applications as any TCP traffic benefits from low latency including HTTP adaptive video. In order to achieve a target of <5ms latency for end customer applications (e.g. Mission critical enterprise applications, Gaming, Healthcare, Connected Ambulance Service, Drone, Robots for Manufacturing and Supply Chain,IoT etc) it is imperative to have Edge DC closer to the enterprises or on enterprise premises. The low latency services not only help Telcos monetize the edge network but also acts as differentiation in the hyper competitive market. The above targets can not be achieved using only Public region DC locations.
• Backhaul Bandwidth savings — As high volume of data is generated at the edge, early break out of the traffic will enable Telcos in saving network capacity compared to backhauling that traffic to central/public region DCs. Building Edge DCs to serve the data closer to where it is being generated helps saving on backhaul network capacity in many of the use cases like Security and Surveillance, Video/Image inspection with AI/ML, 4K Video processing, AR/VR, Streaming services etc. Telcos investment in IoT with huge market requirement and demand for various business sector from Smart Cities, Energy, Manufacturing, Farming, Mines, Oil Rigs,etc huge amount of data will need to be send at short and regular intervals arising network bottlenecks, and degraded data quality that could arise when trying to send information from thousands of sensors, cameras, or other smart devices to a central office all at once. Instead, edge computing enables devices at or near a network’s edge to instantly alert key personnel and equipment to mechanical failures, security threats, and other critical incidents so that swift action can be taken.
• Edge Inferencing — Many of the use cases like Security and Surveillance, Manufacturing quality inspection, Robotics for Warehouse automation, Distributed Apps/Blockchain etc would require edge inference along with AI/ML capabilities closer to the enterprises rather than relying only on centralized DCs.
• Data Localization — Public sector and Smart Cities networks require data localization within the organization or within State/District. This would restrict the application hosting and data processing outside the State geography into Centralized/Public region DCs.

In order to build the Edge network for Enterprises/P5G use cases, this blog will talk about the ecosystem framework bringing 5G capabilities & Cloud scale to deliver applications closer to the end customers/Enterprises.

Challenges & Business problems

Although there is growing demand for P5G/Edge Networks there are many challenges & business problems that need to be addressed by Telcos for their end customers. Few of them are explained below:

• Connectivity & Capacity : Existing 4G & WiFi networks insufficiency for Latency, Reliability and Security requirements.
• Realizing power of Analytics : How to bring in Business & Processes automation within Enterprises to fully realize the power of AI/ML
• Business engineering of use cases : It’s critical to have the justification for the ROI, Capex vs Opex and workout consumption models which are easy to consume by the Enterprise customers.
• Early use cases : Identification of initial use cases for both B2B & B2C
• B2B Use Cases : Visual Analytics [Quality checks, Safety & Security, Remote expert], Digital twins, Industrial IoT, AR/VR [Skills enhancements, Remote troubleshooting], Latency use cases [Robotics, Drones, AGVs] to follow, not in early phase for B2B
• B2C Use cases : CDN, Gaming, Immersive AR/VR [venues, events]Identification of Early adopters (with willingness to pay) for automating at scale within industry verticals [Manufacturing : Predictive maintenance, Quality check, Digital Twins], [Supply Chain : Digital twins, Real time visibility], [Retail : Connected consumers, Connected Supply Chain, AR/VR], [Oil & Gas : Connectivity, Safety & Surveillance], [Mining : Connectivity, Safety & Surveillance, Pilferage], [Ports (Airport & Seaport) Connectivity, Asset monitoring & tracking]

P5G/Edge Network Deployment Models

There can be different deployment models when considering to build Private 5G/Edge Networks for Enterprises as shown in the table below with the option of “Owned & Deployed by Telco” OR “Owned & Deployed by Enterprise”:

Deployment Models

Reference Framework — P5G & Edge Stack

Reference Framework Architecture

The above picture describes the reference framework architecture for P5G/Edge network stack connecting to multitude of devices on one end consuming different applications and services distributed across Telco Network edge and Customer edge/Enterprise Edge. The platform comprises hardware layer with CPUs and/or GPUs, networking and storage depending on the application requirements coupled with OS layer. On top of hardware and OS layers it consists of a horizontal platform layer supporting both VMs and Containers as a service hosting virtualized and containerized applications.

Application layer will differentiate between:

• Enterprise edge : Suitable for Edge services like inference, AR/VR applications
• Telco edge : Suitable for Network services e.g. Telco Network functions, IOT and Security service plus Enterprise services like CDN, Visual Inspection, Business Process Automation etc.
• Connectivity Services : Telco 4G/5G/Fixed Line connectivity services
• Other than above, many vertical solutions can be supported in this framework e,g, Smart Factory, Smart retail, Healthcare, Smart City services etc.

Centralized applications hosted in the core DCs/Public region will provide services at scale like AI/ML models training, Edge Network Operations and Management, Marketplace solutions for enterprises etc.

Delivering on P5G/Enterprise Edge Network Ecosystem

In order to deliver successful outcomes for the Enterprise customers, one of the recommended approaches is where Google acts as ecosystem enabler/partner allowing Telco to operate with control & flexibility over what to deploy, how to deploy, what to sell.

Ecosystem Mapping

The above picture maps the value proposition of both Google and Telco to the reference framework architecture highlighted in Red as Telco play while highlighted in Blue as Google play in building the ecosystem.

This framework allows the Telco having best in class network & customer reach to provide Managed Connectivity, Consulting, Services, Billing, Customer relationship while Google provides best in class products/solutions with open hybrid platform, Data & AI/ML capabilities, Industry solutions capabilities, Security, SRE and other innovations.

This helps in building a joint value proposition with Connectivity, Cloud, Consulting and Best practices for solving specific problems at scale for the enterprise customers.

Conclusion

P5G/Edge Networks powered by Google Distributed Cloud Edge (GDCE), Google’s Data analytics and AI/ML capabilities, Partner (Telcos and ISVs) ecosystem and 1P Google assets/capabilities helps build the horizontal enabler layer delivering multitude of the use cases across many industry verticals. This framework provides open, flexible architecture and deployment models for Enterprises and/or Telcos with future evolution to many new use cases enabled by 3GPP rel 17 and rel 18 features/functionality for Private 5G (P5G) edge networks.

Disclaimer: This is to inform readers that the views, thoughts, and opinions expressed in the text belong solely to the author, and not necessarily to the author’s employer, organization, committee or other group or individual.