- Introduce yourself
Hello, friends of GIAC. My name is Zhou Jing, and my name is Zhixian. Currently, I am in charge of the construction and landing of the edge cloud container platform in the Edge Cloud Native team of Aliyun. These years, I have been engaged in the backend system architecture of high availability, high concurrency, servitization, ServiceMesh, resource cloud, service container, cloud native and other related work, I am a back-end development veteran.
In 2011, I founded the entrepreneurship technology community “iDev Entrepreneurship Technology Salon” together with the then kaixin001 boss. It was not called MeetUp at that time, so the network ID has always been iDevz. It was the first time for me to get in touch with high availability and high concurrency. I also got to know many industry leaders and had many opportunities to learn and improve.
In 2012, I came into contact with OpenResty in the process of architecture selection. In the real sense, I practiced big concurrency in my work. In the following years, I have done a lot of things around OpenResty, and witnessed and participated in the early construction and development process of OpenResty community, and felt the power of community for the first time. When I first went to Sina in 2012, I played with Lua every day by myself, and developed an OpenResty framework like PHP Yaf (the representative work of Brother Bird) in order to get people to the Lua stack, and finally successfully promoted the large-scale architecture upgrade of Sina Mobile for many times.
In 2016, I transferred to Weibo internally and was responsible for the construction and implementation of the cross-language service of Weibo. At that time, it was not called Mesh. When we started, there was no ready-made Mesh solution, Linkerd was not well known, and Istio was a later thing.
In 2017, the integration of microblog cross-language service system for WeiboMesh opened the road of open source construction and landing of WeiboMesh. After that, I understand that cloud computing has greatly squeezed the living space of traditional back-end architects. The technologies I am fascinated by, such as high availability, large concurrency, and servitization, are very mature and have been productized by cloud vendors. However, considering my own interests, I simply go to cloud vendors to continue the leading edge.
Therefore, in 2020, I decided to come to Aliyun and chose a direction that I think is the most promising. I mainly consider two points. I hope I can dig deep in a field and do something more meaningful in the future. On the other hand, I think edge computing is a strategic technology that may affect industries and industries in a real sense compared with the traditional consumer Internet. It is also the technology that I know so far is most likely to disrupt and change people’s production and life.
In the following days, I hope I can communicate with you and learn more. Thank you.
- Tell me about the project you have been working on in the last year. What is its technical value? How is the industry developing? What are the technical highlights and challenges of your project?
It is just over a year since I came to AliYun Edge Cloud, during which I have been designing and building Aliyun Edge cloud container platform, and promoting the integration of CDN and edge computing as well as the landing of CDN on ENS project.
Since last year and especially this year, it has been evident that 5G edge computing has been gaining popularity. Edge computing is also more and more people contact. However, it is also clear to all that the current coverage and commercial scale of 5G has not risen, and there are not many cases of application landing. However, both in terms of national planning and social development, marginal computing is widely recognized as a core strategic direction.
In my personal understanding, There are two core advantages for Ali Cloud. One is that Ali Cloud has been building the flying Edge operating system base from IaaS level since 2016 and exploring various possibilities of edge computing. In addition, With the precipitation of Alibaba CDN for more than ten years, ali Cloud has a very deep accumulation of technology and resources. And Ali Cloud is the only one I know that has really achieved the convergence of CDN and edge computing. Another is the huge business ecology of Alibaba economy. We have a variety of super-scale edge business scenarios, through which we can build and polish our edge cloud platform.
As FAR as I know, edge cloud is a distributed cloud composed of large-scale geographically dispersed edge nodes that can be controlled remotely, secure, trusted, standard and easy to use. My main work is to cloud the edge node resources of the whole domain based on the basic resource base of Ali Cloud edge cloud, and build a multi-rent and isolated IPaaS edge cloud native basic platform based on this distributed and heterogeneous cloud base. Support CDN products and services, cloud communication, view computing, cloud games, cloud desktop, cloud intelligent terminal and other scenarios.
There are several interesting and challenging points in this. For example, how to abstractly manage heterogeneous, distributed, multi-cloud systems at various levels through a unified base plane? How to provide users with reliable and rich edge cloud native services when the edge single-node resources and network are limited? And most importantly, how to ensure that the overall architectural design that meets all these needs is basically evolvable and iterable? A series of problems faced by the edge are all caused by the extensive coverage of edge cloud node resources, miniaturization, complex network environment, diverse resource networks and other characteristics. The edge heterogeneous resources, the unreliability of cloud and network, and other factors extend the two core technical topics of edge cloud: Integration and synergy will always run through the whole process of our platform construction. In order to solve a series of edge core problems, we designed and built aliyun edge cloud computing platform.
- What issues do you usually focus on when developing a technical solution? How to ensure the smooth implementation of technical solutions?
Before a technical solution to promote must ensure that the system design is based on a deep understanding of business, for the details of the status quo of touch panel, the overall scheme have short term plans, the beginning of the system design point and make careful think about possible problems related to the extension point of design, and related supporting, stability and observability) rely on carding, form a complete set of construction in the process of advancing also cannot little. These are the premise that the technical solution can be promoted efficiently and high quality.
I understand the process of technical solution landing, more is the process of step-by-step verification and focus of the solution, the smooth progress information and phased review is particularly important, I personally think that a good solution and system must not be designed. Architect based on business (demand), the present situation, the allocation of resources, environment topology) and future (short medium and long term development plan) give a clever and extensible architecture is the basis of the scheme fall to the ground, the stability of the system security and practical ground rhythm is the key to ensure smooth implementation technology solutions, long-term communication and checking mechanism in the process, Regularly help all parties involved to find their own position, find the problems of the current system promotion and solve them accordingly, so as to promote the overall promotion into a virtuous circle, and the smooth implementation will become natural. However, there are many difficulties in this process, especially the persistence of the design and implementation of the mechanism. Time is one of the most terrible things, a good blend of the ups and downs, the trifles and the milestones.
- What are some of the challenges that architects face in the latest wave of technological change? How to address these challenges?
My initial career choice to become a back-end architect was based on the idea that a back-end architect would only need to understand the business and dig deep into the technology. That is a developer’s most modest expectation. I want to solve all kinds of interesting technical problems every day with machines and code.
But that’s clearly not the case. Especially when I changed my job recently, I thought about starting a business together with my friends. At the beginning, I also participated in various matters involving early-stage venture investment. During the whole process of looking for money, I gradually understood the logic behind the pursuit of profit by capital. To sum up, every organization and individual should be clear about their core values and what value they have created and what pain points they have solved for the society. For example, The core of Deppon Logistics must be the whole logistics business, and IT facilities are only his auxiliary management tools. The core value of Sina News is to provide users with more real-time, thousands of consulting services. As long as these services are stable and reliable, they should meet the business needs.
I’m glad I years always focus on the backend architecture in this direction, from the monomer was applied to micro service architecture, from the physical machine, more room for deploying the physical architecture to distributed cloud cloud native architecture, cloudy resources from the RPC to witness ServiceMesh etc and depth of service involved in all kinds of architecture, the transformation of technology, during the process of iteration. I understand that there are two driving forces behind this series of technological changes. One is that profit-driven capital drives organizations to pursue cost reduction and efficiency; the other is that the social division of labor becomes more and more detailed. It was a natural choice.
In this context, I think the most important things for architects are the following: one is to abandon technical idealism and always understand that all technology is for the business, the technology is for the business, and the business is for the business; One is in the current cloud computing, cloud native technology popularization, under the tide of pratt &whitney and to keep pace with The Times, based on the deep thinking of the core value of your own organization, clarify the technical construction of whether organizational core competence, timely switch, based on cloud computing, cloud native, or a hybrid cloud is the best path to design your own architecture, by good architecture design in order to save operating costs, Improve service maintenance efficiency. First, keep a keen sense of technology and keep learning. Only by constantly refreshing their knowledge system and finding their own positioning can they better realize their own value.
- What are some of the questions you often consider when selecting a technology?
I think the most core point is the business understanding, demand mining and status quo comb. This is the premise of technology selection. What is the nature of the problem I want to solve? The other thing is what is the abstract model of the technology or solution related to the selection to the essential problem THAT I’m trying to solve? What are the capabilities of the current solution? How does it help us solve our problems? What are the possible side effects of the introduction of corresponding solutions and technologies? Is there a simple solution? What is the cost of introduction? Is the overall selection conclusion guaranteed to be relatively simple, maintainable, scalable, and iterable? Is the stability capability easy to guarantee?
- Which project or technology in cloud native do you believe in, and why?
I engaged in the field of edge cloud native itself, although the edge of cloud native is using cloud native concept on the edge of the edge of cloud build native technology infrastructure and ecological industry, it is more than just a cloud native, but a lot of technologies are interlinked, and a lot of technology we are also willing to try even to the edge of reuse, So it seems to me that there are just too many high-value projects and technologies in the cloud native space.
If I have to discuss the most optimistic, first of all, what core problems do I think this project or technology has solved? Did it also solve some of the pain points in my scene? Do these issues have, or are likely to have, profound implications for the entire cloud-native system?
For example, when we are building an edge cloud container platform, considering that we need to provide multi-rent isolated K8S service for tenants, and at the same time, we also need to manage many W devices among thousands of IDC in the whole network at the same time, a single K8S cluster cannot handle it. The VirtualCluster project gave us a lot of ideas about large-scale and multi-tenancy of K8S, which I understand is the current best practice for multi-cluster management. Another example is OpenYurt, which solves a series of problems such as off-grid autonomy and unit-based management encountered in the production process of K8S in the edge environment. These technologies are all being productized for external services in Aliyun. In my opinion, being selected and productized by cloud vendors is the biggest recognition of this project or technology, because the productization process requires cloud vendors to invest a lot of research and maintenance costs, and no organization will pay for worthless things.
- Please talk about some of the hot technologies that you have been paying attention to recently, and what you think about them
I have been paying more attention to DDD, WebAssembly and Dapr recently. DDD domain driven design may be a powerful framework methodology to support the real large-scale implementation of edge computing in industry and industry in the future. However, the contradiction between the professional and vertical technical knowledge of the domain and the iterative evolution of IT infrastructure has gradually become prominent. DDD provides domain experts and IT technical experts with more explicit and specific communication primitive, framework support.
WebAssembly + Dapr is considered the next generation cloud native runtime, and I’m a believer in that. A cross-language, portable, lightweight, high-performance, security-isolated runtime will provide powerful runtime support for a variety of edge heterogeneous scenarios.
- Please introduce the topic or topic you are responsible for in this GIAC speech
I am very glad to be invited to attend this GIAC Conference as the producer of the edge computing special session, because 5G edge computing is still in the early stage of initial implementation. In terms of topic selection, ONE of my main starting points is to invite guests from different sectors, starting from their respective positions (fields). From different perspectives to interpret the current 5G edge computing development status and the overall situation of the landing.
It is a great honor to invite relevant senior experts from production, research, operators and cloud computing manufacturers to share with you. Here, please allow me to introduce you and share with you our guests and topics.
First of all, we are honored to invite Dr. Wang Zhe, a senior engineer of China Academy of Information and Communication Technology (CAICT). Caict has always been the leader and guide of 5G edge computing standards, and as an authority to guide the localization of international standards, caICT’s understanding of 5G edge computing will promote and guide the better development of the industry. Dr Wang Zhe to bring us at the conference on the edge of computing development situation and the application practice, help people grasp the edge of computing the overall development situation, understand the edge calculation the present application situation and the key technology in the Internet industry development trend, understanding of the first edge of the computing industry in our country project “edge calculation standard parts plan” the latest progress and future plans.
The capabilities provided by 5G, such as high-performance wireless data connection and local data streaming, provide unlimited possibilities for the business scenarios supported by edge computing, and edge computing itself is one of the core business scenarios of 5G. The combination of 5G and edge computing is an important attempt of the deep fusion of CT and IT. The implementation of 5G edge computing in the industry is inseparable from the promotion of operators’ traditional basic network construction, but more inseparable from the combination of CT and IT technology, network edge cloud collaboration and other breakthroughs. The honor to invite to a researcher with the institute, director of the China mobile In the song dynasty Titled “5 g edge to bring computing business scenarios, technical system and its application to explore” share, help you to understand 5 g edge computing and industry application potential of combining site, understand the operators 5 g edge computing system, explore more cooperation direction and cooperation based on this model.
If the INSTITUTE brings standardization and directional guidance, and the operators bring the practice of 5G overall infrastructure construction, then the real cases of the industry will certainly bring more reference and reference value to everyone. We have the honor to invite Mr. Shan Jinxiao, a technical expert in the digital center of China Merchants Group Co., LTD., to bring us the theme of “Practice and thinking on the construction of edge computing in large traditional Enterprises”. Focus on sharing the experience and challenges of China Merchants Group, a large diversified traditional enterprise, in the process of digital transformation of cloud, network, edge and end collaborative construction, discuss the value and significance of edge computing in the digital transformation of traditional enterprises, and share the practical experience of construction. Let us have a deeper understanding of the technology architecture ideas in the digital transformation of large traditional enterprises, understand the application of distributed hybrid cloud in large diversified enterprises and the exploration and practice of edge computing in large enterprises.
Finally, I bring to you a titled “ali cloud edge cloud vlsi container platform practice technology sharing, ali edge cloud as the industry only has realized the traditional CDN business and edge cloud fusion, distributed heterogeneous cloudy fusion, distributed cloud standard is easy to use, this share mainly focus on the platform of distributed heterogeneous cloudy cloud and the fusion, And some best practices of the edge cloud native IPaaS platform and business landing process based on this construction. Finally, briefly share the current Ali edge cloud has some noteworthy edge cloud capabilities. Help you better landing edge cloud.
- Any comments on this GIAC
Although this is the first time to participate in the conference as a producer, but I would have is an old friend of GIAC assembly and a high availability architecture community, from following a high availability architecture community have to upgrade to the now together hand in hand together, hope GIAC assembly and a high availability architecture community even better, hope to open source technology community in China is becoming more and more active, the new generation, To contribute more to our domestic technical environment, I hope GIAC will bring more quality content to you.
Welcome to the Edge Computing Sub-forum of the GIAC Global Internet Architecture Conference on 31 July 2021.
PS: If you want to know more about Ali Cloud edge cloud business, you can nail search Jiaqun 35469210