Click to listen to the 10th issue of "Big Coffee Interview": Thinking about open source governance issues from the perspective of OSPO
(It is recommended to play under WiFi)
Open Source Rainforest: Please introduce yourself briefly
I am Bian Sikang, the person in charge of Ant Open Source Office. I have been doing software research and development for more than ten years, and have accumulated a lot of development and product experience. When I joined Ant, Ant took open source as an important direction of technical strategy exploration. At that time, I took over this project and established OSPO. It has been almost two years now. This is my only experience in open source. I may grow up with many open source communities. The types of talents who get up are different.
Open Source Rainforest: So is Ant’s open source office working in a project-based way?
It can be said that a more accurate way of speaking should be a strategy initiative. My department (Technology Strategy Development Department) will make corresponding long-term strategic judgments based on industry technology trends, and explore corresponding fields based on strategic judgments to determine whether there are potential business opportunities in this field. When we first explored open source, we looked at it more in a more restrained way. We did not consider what we could do through open source from the company level, and whether we should take open source as an important strategy from the company level and carry out systematic research. The promotion (of course, the later development has achieved this).
Open Source Rainforest: Is the OSPO you understand and the OSPO of Ant different from the OSPO of most other companies?
I think maybe every OSPO is different. For example, if the company is using open source, contributing to external sources, and the project is open source, it may need a systematic way to manage the business areas involved; or the company encounters some compliance risks and security risks. Driven by pain points, it brings demand, and then grows into a different open source office according to its own business scenarios.
OSPO actually has common parts, and TODO Group has already had a more detailed dismantling of this issue. We believe that OSPO is responsible for three things: one is the risk domain , how to ensure the compliance and security of the open source field is basically guaranteed? The second is excellent practice and business effects . We pay attention to the internal practice of an open source project and engineering culture, as well as the business exploration effect we want to achieve through this practice. Most of the attention is focused on areas such as our technical efficiency and public domain development model that are closely aligned with risk performance, and how we develop products in the public domain to obtain anti-fragile projects that naturally grow out of the open ecosystem; the third is business Development , when the project develops to the point where it can be closely integrated with the external community and form a benign community, how will it relate to the community, the foundation, and the external business ecology? This kind of positive thinking on how to use open source to promote business development, I think, is a development model of OSPO, and Ant is also doing it, and will make dynamic adjustments according to the problems it faces every year.
Open Source Rainforest: Is there an overall principle behind the continuous adjustment of Ant OSPO?
Ant's open source can be summed up in two words: pragmatism and trust . The main business scenarios of Ant: payment, international finance, To B, etc., the logic behind it is trust, and open source itself is a very specific and effective way that can bring trust. Another point is pragmatism. I think pragmatism is a very specific thing. Ant’s pursuit of technology is very simple. We hope that our technology can solve the real problems of the industry. If the solution is not bad, we will open source these technologies and do ecological development. I feel that maybe these technologies can have better development under the open ecology, and can help everyone better solve common problems in the industry.
The typical ones are SOFAStack and subsequent OceanBase. SOFAStack is a cloud middleware that grew out of the Double 11 business scenario. The subsequent OceanBase is based on the payment business scenario of Ant. No matter how high the traffic is, the technology formed by the payment data error cannot occur. It is also a ten-year-old sword. Very challenging.
Open Source Rainforest: OSPO has so much work, how does Ant OSPO prioritize?
Ant OSPO has been established for two years, who will be responsible for the important and urgent items in the first year, and how to formulate the personnel meeting mechanism process. We need to build a people meeting mechanism process that can help us solve the problems we encounter, bring some different insights, and generate potential best practices that can positively affect the entire industry. In the early exploration, we found that there is a lot of content to be developed, and it is not as simple as imagined. For example, how to do project review, how to do internal organizational culture promotion, etc. Based on this, some problems have been formed that need to be solved in the first place, and we have also made these things stable in the first year. The second year was mainly about open source introduction compliance risks, open source tool issues, and internal project incubation and development issues discovered in actual combat. Because many things cannot be planned in advance, we will make dynamic adjustments based on our exploration findings and the status of the development cycle of the open source office.
Open Source Rainforest: How to understand expectation management, and how to do a good job in leadership expectation management?
Managing expectations has many dimensions. Before doing expectation management, we first need to clarify what the core issues the company needs to solve, such as the compliance tools we just talked about, and some fundamental issues at the project governance level.
Open source is an open ecosystem. After completing the current tasks, what business direction can we explore next? This matter needs to be constantly focused with the leader. In this process, we need to make a choice, such as a developer ecology or a ToB partner ecology? What type of incubation services to provide based on certain projects? Is it the guidance of putting forward value propositions, or the theoretical precipitation of comprehensive courses? Do it yourself or work with the community? If you cooperate with the community, what kind of community will you cooperate with? Rather than managing leadership expectations, it is better to help leaders and our management team provide strategic decision-making support, so that everyone can make better next steps based on what they have seen so far.
Another thing to manage is the expectations of all business parties. In fact, many effects of open source are indirect and long-term, and it is not possible to achieve great development immediately after one injection. At the same time, enterprises' investment in open source must also be long-term. If they pursue short-term interests just because the state attaches great importance to it, they may not necessarily achieve good business results. Therefore, the importance of managing the expectations of partners and the expectations of management leaders, I think, is on the same dimension.
Open Source Rainforest: How to view KPI open source, how to avoid open source KPI?
I think the essence of KPI is "body temperature measurement", and the business health is judged based on the data obtained from the actual measurement. If we go one level higher, I think there may be two problems behind the problem: First, should we have a goal when doing open source, and how should this goal be formulated? The second is how to measure the business results of achieving the goals? The so-called open source goals may be divided into two categories: one is the goal of the development of open source projects themselves, and the other is the goals of organizations/institutions like OSPO when they realize open source business.
So should open source have a purpose? My feedback is: Not only should there be, but it should be very clear. The first step in our internal open source governance is to close all applications for external open source projects, because some projects have no goals in the early stage of development, resulting in deformation of subsequent community actions or implementation plans, so we will Before a new project is open-sourced, there is a very clear goal-setting requirement, which sets the goals of the open-source project in advance. As for how to set goals for OSPO business, this will actually involve whether the business itself is short-term or long-term. We hope that in the next two to three years, the company's open source awareness and open source water level will be raised to a certain level. We hope that everyone has a certain basic understanding of open source. There may be a set of courses that cover 100% or 80% of the development population. To formulate such goals, I think it is necessary.
If the project has a very clear development goal and a very clear KPI based on the goal, then we should formulate a KPI logic, but what is the transparent information behind the KPI logic, and how can we make good use of this information? is the real key behind the scenes. When we do business, we often have the logic of counting and counting. We should avoid the formulation of some numbers without goals, but whether the number you get can really help you achieve the goal you actually want to achieve? I think In many cases, it is debatable, and this is also a pit that some project teams often step on in the early days.
Open Source Rainforest: What does open source mean for the creation of technical products? How to create technical products in an open ecosystem?
I recommend everyone to read the book "Silicon Valley Ecosystem". This book does a good job of dismantling why Silicon Valley is Silicon Valley and why innovation in Silicon Valley can continue. Although it is about innovation in Silicon Valley, I think Everywhere is open source. How to cooperate between projects to produce the effect of 1+1>2, so as to avoid creating wheels in the public domain, is a core concept emphasized in all "Open Source Bibles". Under the open source ecology, certain things must be abstracted, so that everyone does not have to repeat the construction of some low-value parts. Once this part forms a factual consensus at the open source level, we can focus on solving the next one. The problem, this is the process that produces innovation.
Taking Kubernetes as an example, Kubernetes provides a more elegant solution for cluster orchestration. After years of accumulation in the community, it is also recognized by everyone as a very good solution that can be designed on it to solve the "next problem". Another example is Open Telemetry. Open Telemetry is an open standard that provides relatively clear definitions for Trace, Metrics, and Logs. The definition itself may not include reference design, but with such an open standard, we can In this way, the consumption caused by different data formats or abstract method standards is greatly reduced.
Pulling up from the technical dimension, the "technical product" itself is an elegant abstraction. After I have abstracted some technical content, I want to make it into a product status, so as to solve certain types of problems. Thinking in this dimension, we have also seen some problems in the ecology. If this If everyone does things well enough, then in fact, there will be generalized use for certain types of products, so as to solve a certain type of problem well. But we will see a lot of big and comprehensive things in the domestic ecology, not to say that big and comprehensive things are not good, but to say that large and comprehensive things may not be able to interact with the open source community at the product design level in an open ecology , product iteration occurs naturally. How to form a relatively smooth upstream and downstream in the ecology, so that everyone can rely on it, so that the community can refine some specific points, reach a relatively advanced product consensus, and solve the next problem based on the product consensus. This is very important.
Open Source Rainforest: Do you have any good opinions or suggestions on Open Source Rainforest?
I very much agree with the top-level logic of the open source rainforest as a whole. The diversity of species grown in an open ecology and the sustainable ecological health brought by the diversity will be better than a closed environment. At present, the high-quality content in the ecology, as well as the content and coverage of open source governance and open source culture are missing, or far from reaching a saturated state. It is hoped that Open Source Rainforest can regularly launch some publicity and traction for content or offline communication in certain fields, continue to invest, and improve the open source technology and application level of the entire industry.
Open Source Rainforest builds a knowledge system around the three aspects of open source general knowledge, open source use, and open source contribution. It is willing to systematically share long-term accumulated experience with enterprises, provide cooperation in teams, mechanisms, and projects, and promote enterprises to use open source more efficiently. , Contribute to open source, improve the level of open source technology and application in the whole industry.
The content of Open Source Rainforest has been open-sourced and hosted at https://github.com/opensource-rainforest . You are welcome to contribute content in the form of Pull Request, discuss in the form of Issue, and jointly maintain the content of Open Source Rainforest.
Welcome to pay attention to the "Open Source Rainforest" public account to get the latest and most complete news.