"Developer-centric software is the next trillion-dollar opportunity in technology, and open source provides a roadmap for realizing this vision."
Experts' assertions make us full of longing for the development of open source software. Open source software is freely available, and its source code is available for everyone to view, debug and even fork. However, freely available open source software does not mean that revenue cannot be generated.
In fact, open source companies continue to successfully IPO, and the current valuation of listed open source software companies is as high as hundreds of billions of dollars. Taking December 3 as the time point, look at the market value or valuation of several major open source software companies, MongoDB (163 Billion dollars), Elastic (11.6 billion dollars), Confluent (4.5 billion dollars), HashiCorp (5.3 billion dollars), Databricks (6.2 billion dollars), etc. Not to mention Red Hat, the originator of the open source company acquired by IBM for $34 billion in 2018 .
How did these companies grow from free projects to billions of dollars? There is no doubt that they have implemented a successful business model, and thus skillfully took a successful business route between keeping developers satisfied and charging corporate customers for quality products and services.
Under normal circumstances, entrepreneurs create a reliable governance model for the open source projects of profit-making companies and cultivate a community of developers to help them spread technology and lay the foundation for their growth. At the same time, it is essential to open core technologies and cloud services, develop business service content, and establish correct business models.
Chinasoft.com selected several successful commercial open source companies in different fields to reveal the secrets of their wealth growth.
01
How MongoDB became NoSQL
Open source database first?
MongoDB Chief Customer Officer Richard Kreuter said in an online interview with reporters: "The original intention of MongoDB database design is to unleash the potential of software and data for developers and the applications they build, so that developers can build modern applications faster and more convenient. "
Developer-centric is the core of MongoDB's success as an open source database.
NoSQL database ranked first
MongoDB, headquartered in New York, was established in 2007. On October 20, 2017, MongoDB was listed on the Nasdaq. The market value reached 16.3 billion USD on December 3, 2020 .
After several explorations, several founders developed a documented database-MongoDB, which has since become the company's name and iconic product. MongoDB's database products have long been ranked No. 1 in NoSQL databases, and MongoDB is also among the top ten databases in the world, including major commercial databases in the world.
MongoDB is a document database that uses the BSON file format of binary JSON.
As a new open source database, MongoDB has three core technical advantages:
First of all, at present, in the database market, in addition to relational databases, there are more than a dozen exclusive database types such as key-value databases and document databases, and they are developing rapidly.
The advantage of MongoDB is the world's first document model database. The document model makes data processing easy, because it is flexible, suitable for a wide range of use cases, and it maps the way developers work in modern, object-oriented programming languages.
The biggest difference from relational SQL commercial databases is that MongoDB is derived from NoSQL and handles unstructured data. The Forrester Wave believes that "MongoDB is transforming from supporting simple modeless applications to becoming a data platform for Fortune 1000 customers to handle critical tasks. In the world, half of the data and analysis technology decision makers have implemented or are implementing NoSQL platforms."
Secondly, MongoDB is a distributed system that can meet the needs of users for rapid expansion. Horizontal expansion, redundancy, and workload isolation are the basic requirements of modern databases, so distributed architecture has become an inevitable choice.
On the contrary, traditional relational databases lack both flexibility and scalability, which is a major obstacle that cannot be avoided for companies seeking to build modern applications with the dual advantages of performance and scale. MongoDB's document-oriented database solves these problems.
Again, MongoDB can run anywhere. Based on MongoDB, it can be developed on a laptop, run in a company data center, or managed in a public cloud.
The complexity of managing applications and infrastructure has grown exponentially. Due to microservices, containers and other related technologies, applications are increasingly subdivided into smaller distributed components.
MongoDB's Atlas cloud database, a fully managed service, can help customers avoid the trouble of dealing with distributed databases on their own.
Responding to competition and revising license agreements
Although MongoDB, as an open source document-based NoSQL database, performs well in all aspects, it once suffered from a problem-cloud computing vendors such as AWS, IBM Cloud, Scalegrid and ObjectRocket all made a lot of money by using open source versions to provide services to customers It's full, but MongoDB is far from achieving the same level of benefits.
Faced with the imbalance of interest encroachment, MongoDB finally chose to modify the open source agreement. In 2018, MongoDB chose to switch its open source license from GNU AGPLv3 to Server Side Public License (SSPL), which is a server-side public license , which will apply to all new versions of its MongoDB community server, as well as all previous patch repair versions .
The license change does not affect regular users who use the community server. According to MongoDB's previous GNU AGPLv3 agreement, companies that want to run MongoDB as a public service must open source their software or need to obtain a commercial license from MongoDB.
Although SSPL is no different from GNU GPLv3, SSPL clearly requires cloud computing companies hosting MongoDB instances to either obtain commercial licenses from MongoDB or open source their service codes to the community.
After revising the license agreement, MongoDB ushered in the spring of development, and its development has accelerated significantly. The cooperation with cloud service providers has been smoother.
To the cloud
At present, cloud computing companies, whether it is foreign AWS, Microsoft Azure, Alibaba Cloud, Tencent Cloud, Huawei Cloud, Qingyun, etc., have vigorously launched cloud database service products.
It has become the consensus of cloud computing companies that databases go to the cloud. MongoDB launched the cloud version of MongoDB-Atlas in 2016. Atlas products can span multiple cloud platforms, bringing more convenience and benefits to customers and providing more choices.
One of the advantages is that MongoDB Atlas can better meet target workloads such as high availability, backup and professional analysis by supporting multi-cloud clusters. With MongoDB Atlas, for the first time, companies can simultaneously deploy fully managed distributed databases on different cloud providers across AWS, Google Cloud, and Microsoft Azure, without the need to increase the operational complexity of managing data replication and cross-cloud migration.
At present, Atlas has settled in AWS, Microsoft Azure, and Google GCP. Users can use MongoDB's solutions in any cloud of their favorite global cloud service providers, and become the main source of revenue for MongoDB, in the first quarter of fiscal 2021. Accounted for 42%.
On this year's MongoDB.Live 2020, MongoDB not only released a new brand-MongoDB Cloud, but also many important announcements, including many new features of Atlas, such as Atlas Search, Atlas Data Lake, Atlas Online Archive, and MongoDB Realm, MongoDB Charts, MongoDB Shell, etc.
Using MongoDB Cloud, developers do not need to switch back and forth between multiple technologies, query languages, and data models, which greatly reduces the learning burden of developers.
In China , through cooperation with MongoDB, Alibaba Cloud will better meet the diverse needs of cloud databases in various industries , so that users on the cloud and community users can benefit, help enterprises achieve digital transformation, and help customers take off.
02
Elastic will play well with open source, big data and search
In the DB-Engines search platform rankings, Elastic's open source project Elasticsearch always ranks first, becoming the largest selection of big data search platforms.
Open source + big data + search
In the enterprise field, the combination of search and cloud computing, big data, Internet of Things and other technologies, as well as open source business models, gave birth to Elastic.
Elastic, which was born in the Netherlands and grew up in the United States, is to grasp search with one hand and open source with the other. Using innovative search engine technology and open source software, it is trusted by millions of developers and thousands of customers worldwide. , To become a leading company in the field of big data search and real-time data processing.
In October 2018, the company was successfully listed on the New York Stock Exchange. On December 3, 2020, Elastic's market value was $11.6 billion.
From Elastic official website
"Unique" products
In addition to the open source Elasticsearch, Elastic has also launched commercial products such as Kibana, Logstash, Beats, and Elastic Stack, which are widely used in the field of big data real-time analysis, including log analysis, indicator monitoring, information security and other fields.
These products can help users explore massive structured and unstructured data, create visual reports on demand, set alarm thresholds for monitoring data, and even automatically identify abnormal conditions by using machine learning technology.
A unique aspect of Elastic is that it satisfies the needs of the big data business in the market. The biggest demand for real-time consumption and real-time application is to quickly search for big data, process it in real time, monitor, warn, and remind data abnormalities. Correlation analysis.
Compared with the industry's search and big data processing fields, Elastic has three irreplaceable advantages:
One is to be able to perform data correlation analysis. At present, the data sources of user applications are diversified, which may come from sensors, Internet of Things, Internet of Vehicles, spatial data, geographic information, time information, etc.;
Data forms and types are diversified, unstructured data occupies most, and data acquisition and conversion tools are not real-time enough;
Data includes multiple dimensions such as time, space, and geography, and the analysis of big data is complex and changeable.
It is difficult for the current database software to perform unified management, processing and early warning of multiple data types. Elastic can manage and process all these data types in real time, perform correlation analysis, and perform early warning and in-depth mining.
The second is fast. In the current big data application, the speed of data processing and analysis is relatively slow. A data analysis operation takes ten or twenty minutes, and the slower one takes up to several hours. Elastic's data search and analysis processing is real-time, with millisecond speeds.
The third is scalability. Traditional big data applications use centralized management. When the amount of data increases, database servers will continue to increase, cost will increase, management and processing difficulty will increase, and it will be difficult to meet user needs.
Elastic uses a distributed structure to support applications of different scales, and can be flexibly expanded according to applications, and can quickly expand to servers of different sizes.
"Unique" business model
Elastic straddles both open source and closed source software business models. Among them, the core product is open source, and users can download or use it freely. The closed source are commercial products, including core commercial plug-ins, high value-added products, etc.
As a search company, Elastic provides an open source Elastic Stack, including Elasticsearch, Kibana, Beats, Logstash (ELKB), as well as commercial functions, including commercial plug-ins and original factory-supported commercial subscriptions, and SaaS product Elastic Cloud, including hosted Elasticsearch services , Application search service, website search service, etc.
Elastic's main business income includes total subscription income (license income License + subscription income Subscription) and service support income (Professional Service); professional service income mainly comes from consulting services, training business, etc.
Elastic chooses open source technology and will insist on open source in the future. A community of more than 100,000 developers, more than 350 million product downloads and more than 7,000 customers have contributed code and a large number of application scenarios to Elastic.
At the same time, based on the characteristics of open source, Elastic has formed a bottom-up sales strategy. Potential customers are the technical personnel who often use Elastic, and gradually penetrate through a series of sales and marketing methods to provide more effective enterprise-level for core customer groups. Service support to stabilize its business model.
03
How does Red Hat, the most successful open source company, love hybrid cloud?
To counter the commercial software in the operating system field-Microsoft Windows, the Linux operating system was born and a new model of open source software was created.
With the help of open source software, foreign companies such as Red Hat and Suse thrive, while domestically produced operating systems based on open source Linux such as Kirin and Red Flag have emerged.
The perfect turn from Linux to cloud services
Red Hat’s early business focus was on the Linux operating system distribution. Although it went through ups and downs, it persisted, and sales revenue exceeded $1 billion at one time.
As the most successful open source technology company, Red Hat’s flagship products include Red Hat Enterprise Linux, Red Hat OpenShift, and Red Hat Advanced Cluster Manager (ACM). Developing one's own products on the basis of open source is Red Hat's most successful business model.
By 2018, Red Hat had become a huge company and was acquired by IBM for a high price of $34 billion. This is the largest acquisition by Giant Blue in 107 years.
The value of $34 billion is not cheap, almost 12 times the revenue of Red Hat in 2017.
However, this move did not surprise the industry. On the one hand, IBM has accepted open source more than 20 years ago; on the other hand, Red Hat has this strength, especially in the field of cloud services. Red Hat knows better than anyone how to get revenue from open source.
The acquisition of Red Hat's Openshift made IBM the number one supplier of hybrid cloud, and the estimated market size of hybrid cloud is as high as trillion US dollars.
Suse, which is also very successful in the Linux operating system, has also been resold many times. In 2018, the Suse Linux business was sold to EQT for $2.535 billion.
Hybrid cloud and open source synergy integration
Both IBM and Red Hat are developing hybrid clouds from different perspectives. RedHat provides a middleware environment through OpenShift, and IBM provides overall solutions for enterprise customers in its cloud products.
Both companies have a long track record in creating and contributing to open source software OSS projects. In particular, Red Hat is the most successful Linux-driven open source company in history.
The acquisition allows both companies to play a greater role in hybrid cloud and software development.
Just as Microsoft acquired GitHub to promote open source activities on Azure, IBM has been trying to keep pace with AWS, Microsoft, and Google in the public cloud market.
In Forrester's evaluation of platforms that are only suitable for public cloud development and enterprise container platforms, IBM and Red Hat performed very well. Through the acquisition of RedHat, IBM will have a leading Kubernetes and container-based cloud native development platform, as well as a broader portfolio of open source middleware and developer tools than any other company. IBM will effectively gain a commanding height in the cloud development platform market.
IBM is using open source software to build the future on the cloud. Like any other software company, earning revenue from OSS is a key survival skill for IBM.
By injecting DNA into the cloud, strategy, and sales organization, RedHat provides IBM with much-needed impetus.
IBM spins off, Red Hat becomes the core of the new company
In 2014, Satya Nadella (Satya Nadella) took over from Steve Ballmer (Steve Ballmer) as the CEO of Microsoft, and quickly turned to cloud technology, creating a new myth that Microsoft's market value exceeds one trillion US dollars. IBM's Krishna (Krishna) took over IBM from Ginny Rometty in 2019 and quickly acquired Red Hat for $34 billion, driving IBM's transition to the cloud.
When IBM decided to split the company into two companies in October of this year, the new IBM separated from the IBM consulting business could develop a story around Red Hat. Red Hat is the leader of the Linux operating system and its OpenShift cloud management software. With the exception of VMware, Red Hat has few competitors in the breadth of its hybrid cloud solutions.
Krishna launched a series of specific actions, including strengthening the two major strategies of IBM-hybrid cloud and artificial intelligence. IBM and Red Hat established Linux, containers and Kubernetes as new standards, making Red Hat OpenShift a hybrid cloud. The default selection.
With the help of the cloud native tools provided by Red Hat, customers can quickly develop various applications and provide services that meet specific business scenarios.
04
Big data company Databricks quickly integrates into machine learning
Big data is the driving force for almost every enterprise to promote change. It not only created new formats, but also innovated existing industries. Big data makes it possible for companies like Didi Chuxing or Airbnb, and enables existing companies like Google, Facebook, and Amazon to use data to create tens of billions of dollars in business.
Develop a basic platform for Spark big data applications
Databricks, a big data and data analysis startup based in San Francisco, USA, was founded by Ghodsi and his colleagues at the University of California, Berkeley in 2013. Since then, the company has raised nearly US$900 million, more than doubling the company’s valuation to US$6.2 billion.
In 2018, Databricks generated more than $100 million in revenue from customers such as Nielsen, Overstock, and Shell. It plans to seek overseas development in Asia, Europe, the Middle East, Africa and Latin America.
Spark started as an open source project in Berkeley in 2009 and is an important open source project for big data applications.
Databricks provides a Spark-based cloud hosting platform that allows customers to implement their entire big data pipeline in one environment-from data extraction, data transformation, interactive processing to data products. It uses machine learning, graphics processing, and building and running data products to provide interactive visualization that can unlock the value of its data.
Databricks Cloud’s advanced cluster management capabilities enable enterprises to start, resize and tear down clusters in seconds. Most importantly, its rich tool set enables companies to interactively query and visualize data and build interactive dashboards.
"The purpose of our establishment of the company is to continue to accelerate the improvement of Spark, increase the program's functionality, stability and contribution to the open source community. On the other hand is to put it in an easy-to-use software package, that is, Databricks cloud service."
Databricks CEO Ali Ghodsi
Enter machine learning quickly
Databricks is committed to commercializing Apache Spark and quickly entering the field of machine learning, developing MLflow, an open source toolkit for data scientists to manage the life cycle of machine learning models.
Unlike traditional software development, machine learning relies on a large number of tools. For each stage involved in building a model, data scientists use at least six tools. Before determining the right toolbox and framework, each stage requires extensive testing. The decentralization of tools and the need for rapid iteration make machine learning extremely complex.
Databricks' MLflow aims to reduce complexity through an abstraction layer for dialogue with various tools and frameworks. Individual data scientists and even large teams involved in building machine learning models can use the toolkit effectively.
MLflow solves the three basic challenges of building and managing ML models:
1) Insight into how each parameter and hyperparameter affect the model;
2) A consistent way of experimenting when evolving the model;
3) Simplify the reasoning model in multiple environments.
Cooperation with cloud service providers
At present, some commercial open source companies are trying to cooperate with cloud service providers: if you can't beat the cloud service providers, then join them. These open source commercial companies did not try to fight against cloud service providers, but found a way to create mutually beneficial business models.
A good example of this strategy is the collaboration between Databricks and Microsoft Azure. Databricks users can purchase their products directly through their Azure accounts. Essentially, Azure has become a distributor of Databricks products to increase the revenue of both parties.
Azure also gains the benefit of selling more computing infrastructure and storage through each new Databricks cluster on Azure. Of course, the challenge of this model is to have sufficient leverage to negotiate fair revenue sharing agreements.
Because Databricks is a very successful open source company with a large and loyal developer community, it has sufficient market share and influence to negotiate with Microsoft.
Happy families are similar, and unfortunate families have their own misfortunes. This rule has also been verified in the field of open source software. Successful open source business software companies are similar. Unique products and services, a win-win business model, and a robust development community are all the foundations of their success.