AEC-边缘云白皮书（英文版）-2017-25页_1mb

AT&T Edge Cloud (AEC) White Paper Summary

Core Content

This white paper outlines AT&T's vision and strategy for deploying Edge Cloud (AEC) as part of its Domain 2.0 (D2) initiative. It discusses the evolution of cloud computing, the rise of edge computing due to new technologies, and the benefits and challenges of integrating edge computing into the telecommunications infrastructure. The paper also presents AT&T's architectural approach, key requirements, and the role of open source ecosystems in enabling edge computing.

Main Points

1. Cloud Computing Evolution

• Cloud computing has democratized access to computing resources and accelerated disruptive innovation.
• Public cloud providers like Amazon and Google have enabled third-party services to benefit from economies of scale and geographic presence.
• The concept of "cloud" was originally used to describe the shared computing space between providers and end users.

2. Edge Computing Emergence

• Edge computing places processing and storage capabilities closer to the user end of the network, reducing latency and improving QoE.
• It is critical for applications that require real-time processing and communication, such as AR/VR, autonomous vehicles, and IoT.
• Unlike CDNs, which handle static content, edge computing supports dynamic, compute-intensive applications.

3. AT&T's Edge Cloud Architecture

• AEC is an extension of AT&T's Integrated Cloud (AIC) and is designed to support a wide range of services.
• It leverages network virtualization, software-defined networking (SDN), and open source technologies.
• The architecture includes modular, programmable, and scalable components that can be deployed in various locations, including central offices, public buildings, and customer premises.

4. Edge Computing Drivers

• Reducing backhaul traffic by processing data at the edge.
• Maintaining QoE through lower latency and efficient network utilization.
• Decomposing and dis-aggregating access functions to improve flexibility and reduce TCO.
• Enhancing network resiliency by distributing processing between edge and centralized data centers.

5. Edge Location Tradeoffs

• Optimal edge placement requires balancing cost, latency, space, power, security, and reliability.
• AEC uses scientific optimization methods to determine the best edge locations.
• Edge services must be interconnected via APIs to provide a unified experience.

6. Edge Computing Use Cases

• Infrastructure: Supports virtual network functions (VNFs) across wireline and wireless access types.
• Services: Enables services like immersive customer experiences, real-time analytics, and AI-driven applications.
• Edge computing allows for per-user session-based real-time machine learning and AI capabilities.

7. Co-existence of Centralized Cloud and Edge Compute

• While edge computing addresses latency and real-time needs, centralized clouds are still essential for non-real-time applications.
• A balance between edge and centralized compute is necessary to optimize performance, cost, and scalability.
• Common APIs and orchestration platforms (like ONAP) are crucial to ensure seamless service transitions and consistency.

8. Key Requirements for Edge Computing

• Modular, flexible infrastructure that supports various hardware and software configurations.
• Cloud-native applications with lightweight control planes and support for containers, microservices, and serverless computing.
• Automated orchestration and management to reduce OPEX and improve agility.
• Security and reliability at the edge, including secure network connectivity and peripheral management.

9. Virtualization Infrastructure Manager (VIM)

• VIM is a key component of the AEC architecture, enabling orchestration and management of both edge and centralized cloud resources.
• It supports containers, virtual machines, and microservices, and must be cost-effective and thin to accommodate the scale of edge deployments.
• Open source and standardization are vital to enable third-party integration and industry-wide adoption.

10. Scaling Edge Functions Using Cloud Native Computing

• Cloud-native computing enables scalable, flexible, and efficient deployment of edge functions.
• It supports container-based VNFs, which are easier to manage and deploy across different environments.
• The CNCF defines key characteristics such as container-packaged, microservices-oriented, and agile applications that can be dynamically scheduled and scaled.

11. Open Source Ecosystem

• Multiple open source and standard initiatives (e.g., ONAP, OpenStack, ETSI MEC, CNCF) are converging to support edge computing.
• AT&T aims to establish a consortium to foster global commonality and enable next-gen applications.

Conclusion

AT&T's AEC initiative is a strategic move to support the 5G era by leveraging edge computing to meet the demands of real-time, dynamic applications. The paper emphasizes the importance of open source collaboration, modular infrastructure, and cloud-native design to achieve cost-effectiveness, scalability, and security. By integrating edge and centralized cloud resources, AT&T aims to provide a seamless, flexible, and future-ready computing environment for both developers and subscribers.

Key Takeaways

• Edge computing is essential for low-latency, real-time applications.
• AEC is an extension of AT&T's D2 and AIC initiatives, focusing on modularity, automation, and open source.
• Balancing edge and centralized compute is crucial for optimal performance and cost.
• Industry collaboration and standardization are key to enabling edge computing at scale.