Starrocks responded yesterday: answers to questions about starrocks

Time:2021-10-20

This article is reproduced to StarRocks WeChat official account.https://mp.weixin.qq.com/s/r8…

It was released yesterdayDisputes about starrocks open sourceAfter the article, starrocks contacted us and asked us to help publish its reply to relevant disputes. We do not comment on the full text, but only represent its neutral publication in the community. The following is the original content of its response:

The following is the original text of starrocks’ response

First of all, I would like to thank you for your attention to us. It is necessary to give you an answer about our situation.

We are very grateful to the Apache foundation and the Apache Doris community for creating such an excellent analytical database product. We are also very grateful to everyone who has contributed to Apache Doris. The Doris project has gone through 13 years from Doris to Palo to contributing to the Apache community and returning to Apache Doris. In these 13 years, starrocks’s little partners have also worked hard and contributed countless time and energy to Apache Doris. Without the Apache Doris project, there would be no starrocks today. Similarly, without these small partners, there would be no Apache Doris today.

The main reason why Apache Doris was not mentioned before is that I don’t want to continue to be confused with the Apache Doris project. At the beginning of entrepreneurship, almost all of our team members came from Apache Doris, so we referred to the previous practice of prestodb when naming the project, named dorisdb, and registered trademarks in China, the United States and Japan. At that time, I didn’t notice that this would affect the graduation of the Apache Doris project. In the second quarter of this year, friends of Apache foundation formally discussed this matter with us. Although all our actions are legal and compliant, in order not to affect the Apache Doris project, we promise to solve the problem in the third quarter of this year. Now, we have fulfilled our commitments.

Here, we solemnly declare that from dorisdb to starrocks, from trademark to code agreement, our work is carried out completely under the framework of laws and regulations, and we welcome the attention and supervision of people from all walks of life. If it constitutes a material infringement on any aspect, we are willing to bear all legal liabilities. In the absence of actual evidence, we will also resort to legal procedures for people and the media who make irresponsible remarks and damage our reputation.

At the beginning of 2020, we observed that with the deepening of data-driven, users’ analysis speed needs to be faster and faster, and the traditional architecture can no longer effectively solve these needs. In the history of analytical database, the storage layer has experienced the technical innovation from row storage to column storage. In the future, the computing layer will begin to upgrade and innovate from row computing to column computing (vectorization Computing). “Independent research and development to build a new generation of comprehensive vectorized computing engine“. This is the starting point of our project; “Helping customers achieve great achievements” is the original intention of our entrepreneurship.

As we all know, an analytical database has three major parts: optimizer, actuator and storage engine. At present, we have rebuilt two large components from scratch: optimizer and actuator, and completely abandoned the original row by row kernel. We will release a new version of the real-time inventory engine recently, and have also made a major reconstruction of the original inventory engine. Therefore, although starrocks uses the Apache Doris framework, its kernel has been completely independently developed. The original Apache Doris code in the current project retains the original Apache license 2. Only the code completely developed by US adopts elastic license 2. This is also a common practice in the industry.

So,You can use our products like elasticsearch!

Today, starrocks’ goal has been upgraded to “a new generation of Lakehouse with stream batch integration” “. this is not consistent with the goal of the Apache Doris project. This is one of the reasons why we did not merge the code back into the Apache Doris project. The times are like the wind and the data are like the sea. In such a great era, starrocks hopes to work with all parties to focus on the needs of customers and create a new era of data analysis!