Shulou(Shulou.com)06/03 Report--

Last November, a piece of news from CNBC reported that Wall Street giant Goldman Sachs wanted to share Alloy, a technology platform it has invested in for 14 years, and a language designed specifically for it, free and open source, to other Wall Street institutions. Earlier last year, Goldman Sachs also claimed to have contributed some of its trading and risk control-related code to GitHub (rival JPMorgan Chase has opened up quorum, a blockchain-related technology, to GitHub).

Cdn.xitu.io/2020/4/14/171766c09907cf82?imageView2/0/w/1280/h/960/format/webp/ignore-error/1 ">

Goldman Sachs seems to be serious about open source technology. In addition to joining the Eclipse Foundation, a well-known open source organization in the open source world, Goldman Sachs participated in the commercialization of container technology in 2015 (when technologies such as docker were just beginning to be known), opened up some technologies to GitHub in 2017, and even contributed a loose software license of Apache 2.0 (a license that is very friendly to the open source community) in one of their Java technology frameworks.

It is no exaggeration to say that open source software changes the world-whether in an App or in a trading system, with countless open source code bases and open source tools. Now who dares to say that their software is built from scratch? As Don Duet, head of technology at Goldman Sachs, puts it, from a technical point of view, "open source permeates everything we do."

The necessity of Open Source of basic Software

Why do we need to open source and participate in open source to engage in financial technology? I would like to contribute 50 cents with my past experience.

In Morgan Stanley. It was a year when IT had fewer than a thousand people, was located at 750th Avenue, and engineers were still using Sun and HP workstations. Morgan Stanley has its own infrastructure team to develop UI development framework Morgan Stanley Toolkit (MSToolkit) across operating systems (Solaris, AIX, HPUX, Windows), plug-ins for the first generation of Web application server Netscape, and many other cool technologies that do not seem to be directly related to the securities business but support the upper business, and the level of design is less than that of professional software companies at that time. In fact, it is not uncommon for large financial institutions of that era to develop their own transaction middleware, graph computing engines and even specialized computer languages.

It can be said that, driven by profits, technologies that can help make money may be invested as core competencies. Compared with the Internet upstart technology companies later, Wall Street's IT is not lagging behind at all. The embarrassment, however, is that these closed, internal-only technologies can easily lose their vitality: if some of the business projects that fund a technology are cut, the technology is likely to be wiped out, and the market is not good. This kind of technology is also the first to bear the brunt. With too few audiences and too narrow application scenarios, these technologies quickly lose their first-mover advantage.

At the Yahoo Research Institute. Like many giant Internet companies, Yahoo has its own system of technologies, including object storage, message middleware, internal wiki, JavaScript development framework, mobile development tools, Web servers, and even similar package management and distribution tools in Linux. These technologies may be very advanced in the early days because there are no equivalents on the Internet, but with the development of the online open source movement, these closed technologies are becoming more and more non-mainstream, and newcomers tend to freak out. People who stay in a closed technology environment for a long time are also easy to be brainwashed by tools, do not know the outside world, leave this environment to go out to look for a job, interview do not know. In the end, these technologies have been abandoned by internal people and become extinct, and people are more willing to integrate into the open source world. The technology that still exists behind Yahoo is Hadoop, which has affected the development of the entire big data field, proving that open source software has strong vitality and outlives its inventors.

In a domestic securities company. Personally, I have always believed that the logical architecture of software must clearly reflect layering and abhor the realization of being a "big pile". To give two examples: the first is to develop a social customer relationship management and service platform for brokerage business, involving infrastructure such as instant messaging tools and rule engines, because at the beginning the scenario is very simple. So we are "hands-on" reinventing the wheel to make them in the system directly as part of the business function? Or explicitly decouple them as general-purpose modules? If it is regarded as an independent module, it is self-developed? Or do you adopt a third-party plan? The actual situation at that time was that without suitable third-party technology (that is, not long before Wechat came out, and mobile IM was still a new thing), our own R & D probably dug a big hole, which would be difficult to maintain by a special team in the future; directly as an application function to meet short-term business needs, it would lead to very poor malleability and forward-looking flexibility and no elegance in technology at all.

Another example that is easier to understand is the trading system. in order to build a solid foundation of the trading system, it is obvious that we must have abstract, highly universal and thoughtful message middleware that has been repeatedly demonstrated in the trading scenario. Open source message middleware on the Internet is often not designed for securities business, can not be used or extremely awkward to use. Independent research and development? You must have a strong team, a determination to maintain optimization over a long period of time, "sponsorship" from the business unit, and support from the company on IT strategy. Or I won't be able to do it. Many of the so-called "self-developed" trading systems will fail because the basic technical layer of "indigenous steelmaking" fails.

Not only business application-oriented financial institutions, including technology companies, will inevitably need to involve a variety of basic technology frameworks, technology libraries, and underlying tools in the process of developing their own technology products, which are often "chicken ribs". If you do your own research, it's a bit "unprofessional", and your team is often very unprofessional in these areas and can't be maintained in the long run. In fact, the open source path taken by Goldman Sachs is to save costs and make efforts.

The pain Point of IT Research and Development in Financial institutions

The IT research and development of securities firms and banks, "Zhengye" must be to support business innovation, do can produce differentiated competition applications. In this process, there should be "something to do, something not to do": infrastructure and general technology framework, try to rely on third-party, business-related logic, as far as possible to develop their own. But the real world is not so ideal.

Using third-party closed technology, you may have to struggle with these problems: the technology of traditional big companies is very closed, and they do not have an open interface (that is a money-making tool-charge for one). Their ancient technology architecture can not carry plug-in customization, they do not respond to your personalized demands, it is very uncomfortable to use. You are worried about the technology of small companies, for fear that they will change the direction of their products or even terminate their product lines, or they may be acquired, merged or even shut down. In short, you are afraid of being "kidnapped" based on closed-source technology.

If you do your own research, you may also struggle with these problems: for some underlying technologies that are not directly related to the business, your team either lacks expertise to control or is unable to concentrate on doing it well. once launched, it is tantamount to digging a hole for yourself. and often face questions about the cost budget. Using open source technology, there is often no choice that is in line with the characteristics of the industry and directly available for financial scenarios, so it is expensive to learn to transform (once you transform yourself, you are likely to fall into the pit of long-term maintenance).

Unless you are Goldman Sachs, in some business scenarios where some basic technologies are needed but there is no choice in the market, develop on your own, and then share it with the open source community for maintenance and utilization by the same industry and even non-industry technicians. avoid the fate of "stillbirth".

It is not very useful to buy source code when purchasing closed source software.

Many institutional procurement systems like to ask for source code-buy from them for big companies and take from them for small companies. But personally, I don't think it makes any sense.

First of all, the practice of "buyout source code" is really, very old-fashioned. The source code is not fixed and will not change after a "buyout", because manufacturers may continue to upgrade, fix defects, and optimize in the future. You buy a "snapshot" of the same thing, based on it. It will soon be separated from the original version of the factory, manufacturers will not be able to maintain for you, and you will not be able to enjoy their optimization, refactoring and upgrading in the service industry.

Secondly, "buy code" this kind of thing, a considerable part of the money is spent wrongly, in fact, they do not have the resources or ability to maintain, that is, to buy an "insurance" bar.

Even if you really want the source code, it is not a "buyout", but a purchase of services-such as getting access to part of the manufacturer's common code base. Within the time frame of the service contract, the latest code can be extracted at any time and can be built by itself through the tools and guidelines provided by the manufacturer. In a word, if you modify it, you can submit it back to the developer's code base for merger. Is to use source code version management tools and best practices to manage industry-level, cross-agency assistance.

Of course, some developers do not have the ability to deliver and manage this source code. You want to "buy out"? Make a compressed package and send it to you as an email attachment, or find a network disk to share temporarily, or send it with QQ. After that, the code completely bid farewell to the main library and can no longer be synchronized.

Change the way software is developed, sold, and delivered

More and more software companies, business models are based on open source ecology. It is appropriate for companies that provide basic software technology to financial institutions to do so. How to calculate "basic software technology"? It is the basic technical layer that has the generality and pertinence of the industry, meets the commonness of financial business applications, and summarizes and abstracts from the business scenarios of many financial institutions. It is often first oriented to the IT R & D personnel of financial institutions for customization and secondary development to support higher-level business applications. Take our company's instant messaging technology as an example, it can be privatized, equipped with compliance storage and proof engine, client SDK can be arbitrarily embedded in any bank and securities firm App, provide hundreds of interfaces and events for the integration of financial business applications, it itself is a development platform. This is the basic technology tailored to the needs of the industry.

The basic technology is suitable for open source, and it is a new software production cooperation model, which has the following commercial competitive advantages:

Attract and utilize talents through the open source community. The most promising and useful open source projects tend to attract top engineers. This is actually a use of "leverage" and can also promote the rapid iteration of open source products in the community.

It helps to get customers. As mentioned above, many financial institutions also want to obtain the source code when purchasing the system, and the logic behind this is to worry that the manufacturer is closed so that they cannot be customized; and the prevention of being "kidnapped" by a certain technology-in case it suddenly ceases to exist and cannot be easily replaced, while the minute-by-minute and money-linked business system is built on it, this risk is too great. Sensitive basic technologies, such as instant messaging tools, whose source code is open and auditable, also reassure some institutions about information security and data privacy protection. Open source has now become a guarantee of confidence.

Help to build a business ecology. Through the opening of intellectual property rights, customers can adopt it at a lower threshold, gradually form a market scale, form a community, establish upstream and downstream partners, and have the opportunity to become a de facto standard (de facto standard). Kafka is a good example. Unwittingly, many financial institutions have adopted it, and technicians even use its technical concepts as a vocabulary to describe problems and solutions.

For software companies built on open source, R & D is a community-based collaboration, led not only by their own R & D personnel, but also by customer developers (such as reporting defects or even directly repairing and submitting code merging). There are even Internet technology fans who agree with this technology.

The way of selling has also changed, and organizations are more likely to get information, word-of-mouth and cases about products online, and communicate with software companies at any time through the community or other social channels. Sales staff can accompany the online operation of demo, two-way interaction. The threshold for software experience is very low, for example, we use a dual license system, the community version is free and uses a non-commercial license, while the enterprise version uses a Copyleft ("copyright", see the next article introduction) license or commercial software license. The community version allows financial institutions to achieve prototype verification of POC with the lowest threshold and even directly use it for commercial use.

In terms of delivery, we deliver the source code through the developer community, part of the code is open to the Internet, and anyone can obtain it directly from GitHub; some is open to institutional customers, through which the developer community can access the source code base and drag authorized branches. The finished products of the software system are all based on the cloud's native technical architecture and 100% containerization, and users pull images through the image warehouse for automatic deployment. It's like a parts warehouse for a carmaker, where IT workers press a button (run a script) to download and automatically assemble the relevant doors, tires, seats, steering wheels, engines, housings, and so on.

There are a variety of open source business models, including open-core and hybrid, etc. We will share in the next issue of "Open Source Business Model promotes the Development of Financial Technology Ecology" to explore how open source technology can be established in the financial industry and promote the development of financial technology.

Article source: Fantai geek

About Fantai geek: help financial institutions and even any other industry organization that wants to have similar technology, build "debris" distribution centers, reduce management costs, improve the efficiency of research and development, form your own digital ecology, and establish true digital connections with customers and partners.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.