Open Source Empowers Inclusive Future——Review of Global Digital Economy Conference and Open Atom Global Open Source Summit (Intel topic)

1. Background of the Summit

From June 11th to 13th, 2023, Beijing, China ushered in a grand event of the Global Digital Economy Conference and the Open Atom Global Open Source Summit. The conference was grandly held in Beijing Beiren Yichuang International Convention and Exhibition Center, providing a platform for exchanges and cooperation for representatives of the digital economy and open source communities from all over the world.

With the theme of "Open Source Empowers an Inclusive Future", this conference focuses on the huge potential and role of open source technology in the digital economy, aiming to promote the development of the global digital economy and realize the vision of an inclusive future.

One of the highlights of the Global Digital Economy Conference is the focus on global trends and innovative developments in the digital economy. Experts, scholars, entrepreneurs and technology leaders from the political, business, academic and technological circles gathered together to share their unique insights and forward-looking thinking on the digital economy. They deeply discussed the latest development trends in digital transformation, artificial intelligence, blockchain, Internet of Things and other fields, and conducted discussions and discussions on how to promote sustainable development in the era of digital economy.

At the same time, the Open Atom Global Open Source Summit, as an important part of this conference, provides a platform for developers and enthusiasts in the open source community to communicate and share. Participants shared their open source projects and experience, and conducted in-depth discussions on the application, innovation and future development of open source technology in the digital economy. The concept and values ​​of open source have been highly recognized and respected at this summit, and all parties jointly discussed how to achieve innovation and inclusiveness in the digital economy through open source empowerment.

Numerous keynote presentations, seminars and workshops were also held during the conference, providing attendees with a wide range of learning and interaction opportunities. The participants actively participated and shared their latest research results, practical cases and successful experiences in the field of digital economy and open source. These exchanges and interactions have inspired new thinking and cooperation opportunities, injecting new impetus into the development of the global digital economy.

2. What contributions has Intel made to the open source field

1 Provide comprehensive support from hardware to software

As a well-known hardware company, Intel (Intel) has shown strong support and investment in the field of open source. At the 2023 Open Atom Global Open Source Summit, Intel is demonstrating its strong support for the Open Source Foundation.

Ms. Liu Jingjuan, assistant to the chairman of the Open Atom Open Source Foundation, said in her opening speech that she is very grateful for Intel's strong support for the Open Atom Open Source Foundation. She mentioned that Intel, as a well-known hardware company, has made great achievements in the field of semiconductor chips, and it is also a company with software services and support as its core. Intel actively embraces open source, adheres to the principle of giving priority to open source, and has more than 20 years of experience in open source investment, investing huge resources in the fields of Windows kernel, cloud native, virtualization, and artificial intelligence. And pointed out that Intel has not only achieved success in hardware, but also has rich experience in open source technology. They established the Intel China Open Source Technology Committee in China to integrate internal resources, use the latest technology and ecological resources, provide Chinese developers with more efficient and professional services, and promote open source ecological construction and industrial development. At the same time, Intel also actively supports the work of the Open Atom Open Source Foundation, joined the governance structure of multiple open source communities, and provided technical and solution support to keep pace with the world's mainstream Linux operating systems.

Intel's open source investment covers multiple areas from underlying basic software to middleware runtimes, frameworks, and applications. They also play an important role in China's open source community, setting up dedicated teams and organizations to provide technical support and solutions. In addition, Intel also supports the development of the x86 platform, and provides technical support and driver code for x86 products.

The Open Atom Open Source Foundation stated that they will continue to strengthen the layout of open source software and hardware, strengthen project reserves and collaboration, and promote the adaptation of open source software and hardware and the formation of an ecological network. They also encourage more enterprises and institutions to participate in the open source cause, and jointly promote the wide application of open source projects in various industries.

Through Intel's strong support for the Open Source Foundation, as well as its investment and participation in the software field, the coordinated development of software and hardware has been promoted, accelerating the pace of progress in software and hardware.

2 Development Trend of Edge AI

Regarding the development trend and challenges of edge artificial intelligence, there are three specific stages: edge reasoning stage, edge training stage and edge collaboration stage. In the edge reasoning stage, through the computing power and training models provided by the data center, the reasoning tasks are sent to the front-end devices for execution, realizing edge intelligent applications. However, this model limits the frequency of model updates. In the edge training stage, it is faced with the need to update the model in a timely manner, such as intelligent manufacturing and automatic driving and other fields. Edge training needs to solve the limitations of artificial intelligence operators and the problem of automatic labeling training. The final stage of edge collaboration is the future development direction, where data centers and edge devices are coordinated to achieve more efficient and intelligent edge artificial intelligence applications.

At the same time, the application scenarios of artificial intelligence are becoming more and more extensive, and the emergence of new artificial intelligence, such as ChatGPT, has brought new possibilities for artificial intelligence applications. Intel's work in the field of edge artificial intelligence includes providing products to help developers meet challenges and providing development tools for the community. Overall, the speakers demonstrated Intel's work and technology in this field by introducing the development trends and challenges of edge AI, and emphasized the importance of edge AI and the direction of future development.

However, edge training is not straightforward and needs to address some specific problems. For example, the people performing AI operations on edge devices are typically workers on a production line or drivers of autonomous vehicles who may not have the energy to do human operations. Therefore, more automated labeling and training methods are needed for edge artificial intelligence training. In addition, edge training also faces other challenges. For example, in intelligent manufacturing, the most valuable samples for training are often samples with defects, while in normal production lines, most samples are normal, which needs to be solved in small The problem of efficient training in the sample case. In addition, privacy protection is also a key issue. Edge training needs to consider how to train while protecting privacy. Therefore, technologies such as federated learning have been proposed for edge training.

If these problems are all solved, good edge training means and methods can be realized. The development of edge artificial intelligence does not stop there, and may further develop into the automatic learning stage in the future, that is, the Auto-Learning mode.

• Increased capacity: Advances in modern technology have resulted in substantial increases in computing, communication, and storage capacity, providing the basis for advancing AI.
• Intel's role: As a hardware and data company, Intel provides technical support in computing, communication and storage to promote the development of artificial intelligence.
• Challenges of edge artificial intelligence: Compared with data centers, edge artificial intelligence faces more limitations and challenges, including lower computing power and power consumption requirements, as well as application development and data protection for different platforms.
• Technical solutions: In order to meet these challenges, Intel provides a full range of technical support, including products with different architectures (such as CPU, GPU, FPGA, etc.) to meet different artificial intelligence computing power requirements, such as providing high performance in big data training processor.
• Issues to consider: When developing edge artificial intelligence applications, practical issues such as seamless migration between different platforms and model protection need to be solved. These are factors that must be considered.
  

3 AI Computing Solutions

As a leading company, Intel (Intel) provides a variety of artificial intelligence (AI) computing solutions to meet the needs of various industry applications. These solutions include general-purpose processors and integrated graphics, communication and storage capacity enhancements, AI training platforms, AI reasoning deployment toolkits, and hardware and software innovations. These solutions are widely used in smart education, smart retail, 5G communication, edge artificial intelligence and other fields, providing users with high-performance and high-efficiency solutions.

Below is an overview of Intel's innovations and solutions in AI computing.

1. General-purpose processors and integrated graphics: Intel's general-purpose processors, such as Core and Smart processors, integrate graphics cards and have acceleration capabilities for media encoding, decoding, transcoding, and artificial intelligence inference. These processors are suitable for vertical industry applications such as smart classrooms and smart retail, and provide acceleration capabilities for dedicated AI processing.
2. Improving communication and storage capabilities: Intel provides solutions for 5G access and core networks in terms of communication. For example, the Microsoft Boost acceleration engine is integrated in the Xeon processor to increase the 5G access capacity. In addition, Intel also provides a series of hardware security measures and provides software tools for AI training and reasoning to improve communication and storage capabilities.
3. AI training platform: Intel provides the GIT platform for edge AI training. This platform helps developers quickly find the most effective training samples and conduct secondary training of the model, thereby reducing the number of samples and the demand for computing power.
4. AI Reasoning Deployment Toolkit: Intel provides a deep learning deployment toolkit that can deploy trained network models to different hardware platforms for execution, and supports multiple artificial intelligence frameworks.
5. Hardware innovation: Intel's fourth-generation Xeon scalable processors have general-purpose CPU and AMX matrix expansion capabilities to meet the computing needs of edge artificial intelligence and improve computing power, power consumption, and security. Intel's discrete graphics include vector acceleration, matrix acceleration and other technologies to provide performance improvements for media and artificial intelligence applications.
6. Software platform and tools: Intel provides the GET platform, which uses active learning technology to select appropriate data for edge training, simplifies labeling and sample selection, and improves edge training speed. The platform supports teamwork and is capable of outputting unformatted training results, with other software tools.

In short, Intel's artificial intelligence computing solutions are widely used in fields such as smart education, smart retail, 5G communications, and edge artificial intelligence. They provide processors, graphics cards, communication and storage capacity enhancements, AI training platforms, AI reasoning deployment tool kits, and hardware and software innovations to provide users with high-performance, high-efficiency solutions. These innovations and solutions help to promote the application and development of artificial intelligence in various industries.

4 Contributions in the field of cloud native

Intel continues to introduce new projects and functions in the direction of cloud native technology, bringing many important innovations to the industry.

Mr. Ding Jianfeng is the R&D manager of Intel cloud computing software. He introduced Intel's innovative projects and functions in the field of cloud native. Among them, the work items of the Intel Cloud Native Team are what he highlighted. Intel has optimized every aspect of its cloud-native platform, including kernels, userspace libraries, containerization, and workloads. They have worked with clients in various fields and played a major role in various cloud computing projects, including the open source project OpenStack.

The focus of the Intel team has gradually shifted to the cloud-native field, and the hardware platform has played a role in cloud-native related projects. For example, they are involved in some important projects such as Kata Containers, Cloud Workload Optimization and Kubernetes, etc. Intel teams have made many contributions in areas such as virtualization, containerization, and management.

In this speech, Mr. Ding Jianfeng also introduced an important technology - Service Match, which was introduced from the Alilongli community. He explained in detail the concept of Service Match and its application in the field of cloud native. In addition, Intel has also done a lot of work on performance optimization, serverless, resource management, etc., and mentioned their hardware support around Kubernetes and integration with the OpenShift project.

Mr. Ding Jianfeng emphasized the importance of CPU, memory and storage in computer resources, and mentioned Intel's enhancement projects in CPU resource management, such as the CRM project. He also shared Intel's active work and innovative projects in the field of cloud native.

In addition to the above introduction, Mr. Ding Jianfeng also mentioned some specific innovative projects, showing Intel's actual achievements in the field of cloud native:

1. CNI Enhancement: Multiple CNIs (Container Network Intelface) have been introduced to support multiple NICs and enhanced NIC hardware devices, presented in the form of CNI.
2. Fine-grained isolation: In 8K8S clusters, especially in stand-alone robots, more fine-grained isolation of storage resources and network resources is carried out.
3. Security enhancement: Use Intel hardware platform technologies (such as TDX and SGX) to strengthen confidential content, including key management and data plane enhancement. 4. Performance enhancement: By using the QT hardware acceleration device and new instruction set in the Intel platform, the TLS encryption and decryption of the traffic on the data plane are enhanced to improve performance and save CPU resources.

5 Intel and Tencent launch cross-platform Android

This introduction mainly revolves around the Android development method. The speaker introduced the Android development method and Intel's contribution to the Android cross-platform to the media from a technical point of view.

1. Android development method: Android is an open source system advocated by Google. Its development method is to close the source first, conduct internal testing and adaptation with partners, and then release it on the reference platform. Intel has always been one of the contributors to Android core development, and can obtain access rights as soon as a new version is released, and adapt and optimize it.
2. Android ecosystem: Android is not just an operating system, but an application ecosystem. Every year, millions of developers develop various mobile applications for Android, and these applications have a long life cycle. Mobile applications are not limited to mobile phones and tablets, and developers are constantly exploring new scenarios for interacting with end devices.
3. End device innovation: Intel hopes to create new business opportunities on end devices through the Android platform. Although Intel's voice is low in the mobile phone and tablet markets, the Android ecosystem is still vibrant in the field of end devices, and developers are constantly looking for new ways to interact with users.
4. Innovation on the PC side: Despite the slowdown in the mobile and tablet markets, the PC side still has huge potential. Intel is committed to driving innovation on the PC side on the Android platform. Through the Android system, developers can extend mobile applications to PC devices to achieve more diverse user experiences and functions. This innovation of combining mobile applications with PC platforms can bring users more convenient and diverse computing experiences.
5. Smart home field: With the rapid development of the Internet of Things, smart home has become a hot field. The Android platform provides an open development environment and a rich application ecosystem, which facilitates the development and application of smart home devices. Intel hopes to cooperate with developers in the smart home field through the use of the Android platform to promote the innovation and popularization of smart home devices and provide a smarter, more convenient and connected home experience.

At the same time, the spokesperson pointed out that PC is still a huge market, with nearly 300 million PCs shipped every year, about 60 to 70 million new PCs shipped in China every year, and the number of active PC users is about several hundred million. Internationally, Google's Chromebooks have supported running Android apps on them for several years. Two years ago, Microsoft introduced Windows Subsystem for Android (WSA) in Windows 11, allowing users to install and run Android apps on Windows systems. Last year, Google released Google Play for Games, bringing its Google Play platform to Windows, enabling users to play Android games on Windows.

The spokesperson mentioned that there are a large number of users in China who like to play casual games. Google saw this business opportunity and introduced Google Play into the Windows system. However, there is currently no Android-on-PC solution in China, which is a huge market opportunity. The speaker introduced the product jointly launched by Intel and Tencent——Tencent's PC App Store, which uses Intel's technology to realize the seamless experience of running Android applications on Windows. Users can open the application market on Windows, select and download their favorite Android applications, and experience the same experience as on mobile phones.

In addition, the speaker also talked about the development trend of smart large screens, which have high definition and powerful AI capabilities. They introduced the 8K cloud screen jointly launched by H3C and Intel, which uses Intel's computing power advantages and ecological maturity to provide high-performance end device experience. In general, Android operates in different interactive scenarios, and has broad market opportunities in terms of high-performance requirements and x86 ecological integration.

6 There are many projects in the field of cloud computing

In the cooperation with Ali and China Mobile, there have been many cases. One of the cases is a cooperation case related to service match with Ali, and the other is a case of network strengthening cooperation with China Mobile. These cooperation cases have played an important role in promoting innovation and improving service quality.

In the field of cloud native, various network components and technical points are involved. These include QTA acceleration, ITS acceleration and the collaborative work of IPV6 and IPV4. The application of these technologies enables the cloud-native system to have higher performance and better compatibility, and improves the overall network efficiency and reliability.

In the field of K8S, the production work environment has been strengthened through cooperation with mobile, and good results have been achieved. This cooperation helps to improve production efficiency and optimize workflow, creating a better working environment for enterprises.

In the cooperation with VMware, the direction is strengthened through EPPF technology, and the performance is improved by using the bypass of TCPIP. This cooperation promotes the development of network technology and improves the operating efficiency and user experience of the system.

Longli Cloud Native is an evolution of resource node resource management, covering cluster management, stand-alone components, underlying operating system, and stand-alone resource management and control. This evolutionary architecture provides cloud native systems with better resource management and scheduling capabilities. Improve system reliability and performance.

The Dragon Lizard cloud-native distribution architecture includes two parts: cluster service and stand-alone service. Cluster services provide container service capabilities, while stand-alone services provide container, network, storage, and stand-alone resource management and control capabilities. This architectural design makes the system more flexible and scalable.

CRI-RM is a container runtime interface proxy component, which is used for dynamic resource partitioning on nodes, and is compatible with the resource management and control of Dragon Lizard. Through the application of CRI-RM, dynamic allocation and management of container resources can be realized, which improves resource utilization and system elasticity.

Anolisbox is a single-node resource management solution for cloud-native scenarios, with features such as standardization, planning, and strategy. It adopts the CI agent mode, and in the running mode, the underlying operating system and containers can be configured according to priority labels or other resource setting labels. This solution provides better resource management and optimization capabilities, improving overall system performance and manageability.

7 Create a community ecology with domestic operating systems

There are many domestic operating system manufacturers and communities participating in the construction and development of the operating system ecosystem. Among them, the Linux operating system is the most important upstream community, and many organizations and individuals around the world participate in the construction of its code, including Intel engineers. However, for enterprise users, it may not be practical to directly use the upstream community version of the Linux operating system, because its update frequency is relatively fast, and enterprises need a relatively stable operating system environment to run applications.

Therefore, some manufacturers in China provide Linux-based operating systems, such as OSA, Ubuntu, Kirin, etc., and they provide operating system solutions suitable for different user groups. In addition, several influential operating system communities have emerged, including the Dragon Lizard community, the Euler community, and the open cloud web community, among which the open cloud web community has a greater influence in the field of server operating systems.

These communities operate with the support of the foundation and jointly build the domestic operating system ecosystem. Users can choose foreign products such as Red Hat, or choose the operating system provided by domestic manufacturers, or directly use the open source version for modification and customization. In recent years, the domestic operating system ecology has developed rapidly, and communityization has become an obvious trend.

In the domestic operating system community, Intel actively participates in the construction and development of the community. Intel plays a role in the operation and governance of the community, and has established the support and contribution of Intel architecture in the dragon lizard community, Euler community and open cloud web community. In addition to platform support, Intel brings other technologies and optimizations into these communities, working with foundations and native OSVs (operating system distributions).

In addition, Intel has also established a cooperative relationship with Chinese OSV manufacturers to incorporate its platform support into its release version, so that users can get immediate support. Kirin and Hongqi are one of the two largest commercial OSV manufacturers in China. They have comprehensive strategic cooperation with Intel to integrate Intel's platform support into their distributions.

Generally speaking, the domestic operating system ecology is booming, and communityization has become an important trend. Many manufacturers and communities have actively participated in it, building a diverse ecosystem of operating system solutions.

3. Conclusion

After reviewing the Open Atom Global Open Source Summit, I feel optimistic about the future of open source technology. The summit showcased the innovative power and collaborative spirit of the open source community, as well as the great potential of open source technologies in the digital economy.

The summit brought together members of the global open source community, who shared their projects and experiences, and demonstrated the application of open source technologies in artificial intelligence, Internet of Things and other fields. These innovative solutions open up possibilities for the future.

However, we still face challenges. We need to strengthen the cooperation and communication of the global open source community and build an open, inclusive and collaborative environment. Promoting open source technology is also key, educating more people about its concepts and benefits, and providing training and support.

Thanks to various open source communities such as Dragon Lizard Community, Euler Community and other communities, and thanks to Intel, CSDN, Huawei, Tencent Cloud, OpenHarmony, Alibaba Cloud, Open Source Foundation, Kirin and other open source communities and members for their support. Their efforts and contributions have promoted the development of open source technology, and let us see the hope that open source can empower the future of inclusiveness.

All in all, the Open Atom Global Open Source Summit demonstrated the potential and value of open source technologies, while also reminding us of the tasks and challenges we face. Through global cooperation and exchanges, promotion and popularization of open source technologies, and attention to the sustainable development of the ecosystem, we can realize greater value of open source technologies in the digital economy and create a prosperous and sustainable future. Let us work together to create a new chapter of open source empowering the future of Pratt & Whitney!