邊緣運算，5G，及AI - 新的競爭空間:邊緣運算和5G的情境中AI的新機會的探求

報告涵蓋了企業訊息和通信技術(ICT)行業的新競爭領域。諸如Kubernetes之類的技術使混合和分佈式雲端基礎架構變得更易於管理和操作敏捷。新興的物聯網(IoT)導致了互聯感測器以及機器人等互聯效應器的激增。同時，5G無線網路已開始出現在企業內部。這些網路可顯著降低延遲，從而在工業控制系統，機器人技術，自動駕駛汽車和增強/虛擬現實(AR/VR)中實現新應用。因此，將應用程式遷移到更接近最終用戶以充分受益於較低的延遲是競爭的必要。

本報告提供邊緣運算，5G，及AI的新機會的相關調查，邊緣運算，5G，及AI的互相的相關性，新興企業，使用案例，商務及運用模式，策略等資訊。

目錄

摘要整理

最尖端的崛起

AI採用的成長要素的邊緣運算

競爭空間的理解

• 企業
• 模式

AI市場的差異

• 忠實的維護，洩漏的危險性
• 等待時間的重要性
• 演算法的重要性

結論、建議

• 硬體設備供應商
• 雲端服務供應商
• CSP
• 軟體供應商

This Omdia report covers a new arena of competition in the enterprise information and communications technology (ICT) industry. Technologies such as Kubernetes have made hybrid and distributed cloud infrastructures more manageable and operationally agile. The emerging Internet of Things (IoT) has led to a proliferation of connected sensors and also connected effectors such as robots. Meanwhile, 5G wireless networks have begun to appear within enterprises. These networks provide dramatically lower latency, enabling new applications in industrial control systems, robotics, autonomous vehicles, and augmented/ virtual reality (AR/VR). As a result, it is a competitive necessity for applications to migrate closer to the end user to fully benefit from lower latency.

The combination of these trends with one more-artificial intelligence (AI)-means that the edge is becoming to the 2020s what the cloud was to the 2010s: the strategic focus of competition in the ICT sector. The applications that enterprises want to deploy at the edge are ones that benefit from the spectacular improvements in AI and machine learning (ML) technologies seen in the 2010s-for example, machine vision, robotics, and time-series data analysis. Having proven itself in the cloud, AI is coming to the edge.

Players from multiple markets are being drawn to this new opportunity. Communications service providers (CSPs), the network vendors that support them, and a new breed of alternative service providers all hope that their expertise in wireless will permit them to own the edge. Semiconductor makers and server OEMs see a major new line of business for their existing server products. Hyperscale cloud providers see a potential disruption of their business model built on centralized data centers and managed services and hope to preempt it by leading the charge to the edge.

Key Questions Addressed:

• Which players are competing in the emerging 5G edge space?

• How do AI, private 5G networks, and edge computing interrelate?

• What are the typical use cases and what role does AI play in them?

• What are the various business and operating models, and which players prefer which models?

• What strategies are the major cloud providers pursuing?

Who Needs This Report?

• Artificial intelligence and machine learning practitioners

• Communications service providers

• Enterprises in vertical industries

• Semiconductor vendors

• Investor community

Table of Contents

Executive summary

Rise of the cutting edge

Edge computing as a driver of AI adoption

Understanding the competitive space

• The players
• The models

How this space differs from the broader AI market

• "Data is the new oil": Jealously guarded, dangerous if leaked?
• Latency is king, so hot chips will only get us so far
• Algorithms may be less important than adaptation

Conclusions and recommendations

• For hardware vendors
• For cloud service provider
• For CSPs
• For software vendors

Table & Figures

• Major applications and edge AI opportunities

• AI chipsets for the edge are set to boom

• Edge servers are likely to be the key category in the enterprise

• Chipset revenue for edge servers

• Global enterprise LTE and 5G projects by type

• Global enterprise LTE and 5G projects by industry

• Players converge on the new opportunity space

• Business and delivery model scenarios

• AI business models by software vendor revenue, 2018-25 forecast

• Enterprises choose the multi-party option for 5G projects

• Four scenarios for the edge AI operating model