Gong Peiyuan, Vice President and Executive Partner of Gartner: Analyzing the five key areas of "new infrastructure"

Source: Communication World, author: Liu Tingyi

Recently, the country's move to vigorously develop new infrastructure construction (new infrastructure) has attracted widespread attention from all walks of life. New infrastructure refers to infrastructure construction that focuses on the technology end, focusing on the new generation of information infrastructure represented by 5G, data centers, artificial intelligence and industrial Internet, as well as the digital and intelligent transformation of traditional "railway, road and air" infrastructure using digital technology.

Ma Yuan, a researcher at the Enterprise Research Institute of the Development Research Center of the State Council, said that focusing on "new infrastructure" is an effective means to respond to the impact of the epidemic, promote consumption, and stabilize growth in the current situation. It is also a key move in the long run to build a foundation for the innovative development of the digital economy and seek future international competitive advantages.

Against this backdrop, Gong Peiyuan, vice president and executive partner at Gartner, offers insights and suggestions to chief information officers and IT leaders in five key areas.

1

5G technology: a full assessment

5G is a new generation of cellular technology that will transform the mobile service delivery model from consumer-centric to business- and consumer-centric. Currently, 5G technology is at the "peak of inflated expectations", and the international standards body for 5G, the Third Generation Partnership Project (3GPP), is developing in sync with the regulatory process, spectrum allocation, and deployment by telecom operators.

Gong Peiyuan believes that telecom operators, as current users and holders of licensed spectrum, are the main mobile service providers for enterprises in the 5G development stage. At the same time, LTE technology will continue to exist and lay a solid foundation for the initial deployment of 5G.

Therefore, Gong Peiyuan believes that plans for adopting 5G should be made by thinking about cellular network services and use case requirements. When starting a plan, first of all, we should no longer regard cellular networks as services that are only for consumers, nor should we develop applications just for mobile infrastructure. Secondly, we should understand how the combination of RF, network software, and slicing technology affects the required architectural changes, start with use cases, and establish 5G connection services. Thirdly, do not adopt 5G blindly, understand the technologies that affect the dependencies at all levels of 5G, and develop realistic expectations. Finally, evaluate whether 5G can serve as a continuation of SD-WAN and LTE or IoT-type application services in the early stages. At the same time, actively communicate with telecom operators to understand how these services are integrated with business-centric 5G services.

2

Wireless Charging for Electric Vehicles: New Opportunities

Gong Peiyuan believes that charging systems for residential and community parked vehicles do not require new technology, the existing system only needs to offer the right price.

In this regard, Gong Peiyuan pointed out that in order to gain market share in the growing automotive wireless charging, technology and service market, existing technologies can be used in residential and community applications to gain time-to-market advantages.

At the same time, target local government departments as a customer base for early charging infrastructure investment, join industry groups and standard-setting organizations, and prepare for the arrival of industry standards.

3

Data Center Infrastructure: What to Do in 2020

How to maintain and optimize the existing data center infrastructure? Gong Peiyuan believes that data center infrastructure can be managed, transformed and improved through automation and machine learning.

When planning future data center infrastructure, it is necessary to develop an infrastructure vision that covers data center compute, storage, networking, and data center design.

In terms of emerging technologies in the data center, an intelligent infrastructure strategy can be implemented with the following characteristics: software-defined, composable, agile, programmable, scalable, elastic, responsive, and able to adapt to change.

The role of data centers in a hybrid environment balances the operations of traditional data centers with the impact of serverless technologies, data centers, hosting service providers, cloud and edge computing.

4

Core technologies of artificial intelligence: 2020 predictions

It is well known that AI infrastructure strategies used for AI project pilots will not be easy to scale to production due to challenges in skills, technology, and infrastructure integration. AI inference engines will be deployed in various locations, including edge, traditional data centers, and public clouds, which will promote the need for full-platform deployment solutions.

Technical debt and infrastructure complexity associated with production AI pipelines will become a difficult task for IT leaders in most enterprises. Gong Peiyuan said that the need for real-time response is driving the need to place analytics near the data collection point and in edge systems or endpoint devices.

In this regard, Gong Peiyuan suggested that IT leaders who want to expand from AI proof-of-concept to production and achieve output growth can consider the following aspects: First, design and demonstrate a prototype of a custom reference architecture for streaming data analysis infrastructure to accelerate the application of AI in production. Second, use containers to encapsulate machine learning models and simplify model management to create an inference engine deployment process while applying AI to production. Third, accelerate the implementation of AI through the strategic use of cloud services and suppliers with scalable AI infrastructure capabilities. Fourth, determine the best location for deep neural network analysis by quantifying the type and amount of data to be collected, as well as the impact of communication bandwidth, latency, and availability.

5

Global Industrial Internet of Things: Market Opportunity Analysis

In 2018, the Industrial Internet of Things (IIoT) software platform achieved profitability for the first time, but profitability has remained challenging since then. Gong Peiyuan believes that compared with other IoT stacks, Industrial IoT applications have relatively more stable growth rates and profit margins, mainly due to the constant changes in purchasing centers and IoT-related purchasing requirements.

However, the growth and profits of industrial IoT services are still limited by the following two factors: suppliers focus on narrow professional fields; and suppliers are unable to create attractive product catalogs for repeat revenue products. Gong Peiyuan admitted that connectivity and hardware (HW) are the basic elements of IoT solutions and are the most important profit and revenue sources of industrial IoT respectively.

In order to create innovative industrial IoT products, Gong Peiyuan made the following suggestions to the technical general manager: First, maintain growth by expanding the customer environment through the development of industrial IoT platforms with specific industry characteristics and functions. Second, develop and differentiate platform-independent industrial IoT applications that can "transform" industry expertise and business wisdom into products for basic use cases in the target industry and generate revenue. Third, ensure annual recurring revenue by creating an asset-based managed service portfolio. Such services can provide an Information Technology Infrastructure Library (ITIL) approach to the operation and service management of IoT solutions and IoT assets. Fourth, increase the growth rate while reducing sales costs by targeting industrial enterprises that create networked industrial products and services (which require embedded or bundled hardware and networking capabilities at the point of sale).