Interview with Michael Stonebreaker
Russell Garland (Russell Garland), WSJ
Michael Stonebraker, when creating a new company, does not strive for big profits, Michael Stonebraker , instead, he is thinking about developing an idea that can revolutionize an entire industry.
From Wikipedia: Stonebriker is a database expert and professor at the Massachusetts Institute of Technology . He is also an entrepreneur, co-founder of eight companies.
')
Some of these startups have already been acquired, including the very first, Ingres Corp. , and, for example, Vertica is currently acquired by Hewlett-Packard, the amount of the transaction has not been disclosed. One of several posts currently is Technology Director (CTO) at Paradigm4 Inc. , a secret startup developing analytics for massive datasets.
Stonebraker worked with a number of well-known venture capital firms, including Accel Partners, Bessemer Venture Partners, Highland Capital Partners, Kleiner Perkins Caufield & Byers, New Enterprise Associates and Sigma Partners .
Michael is one of the leading thinkers of the next wave of innovation in data warehousing (both in management and in analytics), which is called “ big data ”. He recently moderated a discussion on this issue at an event organized by the Massachusetts Council of Technology Leaders ( the Massachusetts Technology Leadership Council ).
We talked with Mike Stonebreaker about entrepreneurship and his projects. Here is the edited
Question: What makes you create startups?
Answer: Let me give you a historical background. The first company I founded was Ingres Corp. - there developed the eponymous popular relational DBMS. I was an assistant professor at the University of California at Berkeley and read an original article on Ted Codd 's relational databases in June 1970. Immediately, there was a gap between the followers of Ted Codd, who thought that relational databases were the best discovery since the invention of the wheel, and the traditional majority who said: “We cannot implement this effectively, because we do not understand modern languages. ” It became obvious that we needed a serious implementation of these ideas.
In 1972, we started working on Ingres in Berkeley as a research prototype and by 1975 a lot worked. Most research projects with the achievement of the performance of the prototype declare the success of the implementation and move on to the next idea. For some reason, we continued to get a really workable Ingres database. Around 1977 or 1978, Ingres worked very well, and we had about 100 external users. Arizona State University even agreed to use Ingres for their student accounting databases, but the whole project went to the dump when it became clear that by that time there was no COBOL programming language for Unix, and they used only COBOL, and this was the collapse of the project.
It became clear that if we want to do something different from the model one — get a serious introduction to real use of Ingres — we cannot do it with an academic prototype. Then we said, “Eureka!”, Let's create our company. This led to the creation of Ingres Corp. in 1980. Ingres worked really well.
After you have started (company), then you understand that if you have new ideas, if you want them to see the light, you either have to find a company to get the idea up or you have to start your own company. And my experience shows that it is many times easier, at least for me, to organize my own company. The general model is such that when I have new ideas and it seems to me that they have commercial prospects, then I would like to get a good working capacity inside the university to ensure the attractiveness of venture capitalists who are ready to invest.
Question: How do you decide which venture capital firms are right for you?
Answer: Venture capitalists need to get more than just money. First and foremost, what I want is for someone to work hard to make something successful. Secondly, I need someone who has really good contacts and connections. Most venture capitalists are really smart. And finally, I want to deal with someone I can get along with.
Question: How much do you think about the possible profit at the start of the company?
Answer: I never think. In the case of Ingres, the relational database management system, there were new ideas, and I was fascinated by them. They have experienced more than 25 years in the database market. It gives me great satisfaction. This leads to a paradigm shift. This is what fundamentally motivates me - change for the better.
Question: What advice would you give to entrepreneurs who want to start their own business?
Answer: What you need is a good idea and the backbone of the team, in order to move the idea forward, you need to have at least one sales or marketing director and super-competent programmers working on implementation. I find the venture capital community very competent in assessing these three characteristics. But if you do not have all the three components, I think it will be very, very difficult to get funding.
The biggest plus, as I see it, is that I now have experience, and it will be much easier for me to do the next project. But if you do not have a “credit history”, then the venture capital community will look askance at you. It's like in the book / movie "Trick-22" ( Catch 22 ).
Another thing I want to add is that the software of modern systems is becoming more and more complex. Federal government agencies, which usually provide money for research, are now virtually unwilling to invest in a commercial profile infrastructure. So if you want to get funds that really work, you have no choice but to look for it outside the traditional channels of research funding.
Q: Now you have CTO positions in two companies, VoltDB and Paradigm4 . Are these your main projects at the moment?
Answer: Yes. It all started with Vertica . It became clear that you need to make data warehouse systems work quickly, organize storage in columns, where you need to organize column-to-column data, it works tremendously faster than line-by-line storage, where data is stored line-by-line. This made me think about “vertical” tools for the market, since in the data storage market, such vertical tools will blow conventional tools to pieces. VoltDB is a tool for the vertical market, it focuses on the operational processing of transactions. Paradigm4 is a specially-created tool for the vertical market, which is focused on the segments of integrated analytics and science, and uses the characteristics of this market to get a realization that is completely different from both Vertica and VoltDB .
I am a passionate fan of specially-built tools, because “One size cannot be suitable for all cases”, and that if you decide to take advantage of the characteristics of specific vertical markets, you can get a solution at a speed of one to two orders of magnitude faster than using standard solutions. Current - I affectionately call them "elephants" - the main suppliers of relational databases, "cut one size fits all", providing a standard tool for all occasions.
Drastic changes occur, because if you asked a question a few years ago what people want from analytics, they said: “I want to know the average salary of employees grouped by departments”. But in many vertical markets, people have found out that they want to have ways for more complex analytical algorithms, for example, self-learning algorithms. The easiest way to break it down is that people who do electronic trading on Wall Street, they mainly build stock market behaviors, and then when it doesn’t work according to their models, they make purchases or sales. They make a very complex analysis. There is a whole range of markets where complex analytics decide. Accordingly, such solutions as Paradigm4 will work very well, because they are based on fast self-learning algorithms.
Question: Do you have any other plans in your plans?
Answer: What I’m thinking now is a huge problem for most global enterprises — data integration . A typical large enterprise has several thousand production database systems, as a rule, written separately, in different groups, without thinking about the tasks of integrating all of their data. I think about this. This is what I am working on at the Massachusetts Institute of Technology. If I come up with something that is commercially viable, then I will think about starting the next startup. For me, this is entirely due to a good idea, because without a good idea one cannot move anywhere.
Michael Stonebraker, when creating a new company, does not strive for big profits, Michael Stonebraker , instead, he is thinking about developing an idea that can revolutionize an entire industry.
From Wikipedia: Stonebriker is a database expert and professor at the Massachusetts Institute of Technology . He is also an entrepreneur, co-founder of eight companies.
')
Some of these startups have already been acquired, including the very first, Ingres Corp. , and, for example, Vertica is currently acquired by Hewlett-Packard, the amount of the transaction has not been disclosed. One of several posts currently is Technology Director (CTO) at Paradigm4 Inc. , a secret startup developing analytics for massive datasets.
Stonebraker worked with a number of well-known venture capital firms, including Accel Partners, Bessemer Venture Partners, Highland Capital Partners, Kleiner Perkins Caufield & Byers, New Enterprise Associates and Sigma Partners .
Michael is one of the leading thinkers of the next wave of innovation in data warehousing (both in management and in analytics), which is called “ big data ”. He recently moderated a discussion on this issue at an event organized by the Massachusetts Council of Technology Leaders ( the Massachusetts Technology Leadership Council ).
We talked with Mike Stonebreaker about entrepreneurship and his projects. Here is the edited
Question: What makes you create startups?
Answer: Let me give you a historical background. The first company I founded was Ingres Corp. - there developed the eponymous popular relational DBMS. I was an assistant professor at the University of California at Berkeley and read an original article on Ted Codd 's relational databases in June 1970. Immediately, there was a gap between the followers of Ted Codd, who thought that relational databases were the best discovery since the invention of the wheel, and the traditional majority who said: “We cannot implement this effectively, because we do not understand modern languages. ” It became obvious that we needed a serious implementation of these ideas.
In 1972, we started working on Ingres in Berkeley as a research prototype and by 1975 a lot worked. Most research projects with the achievement of the performance of the prototype declare the success of the implementation and move on to the next idea. For some reason, we continued to get a really workable Ingres database. Around 1977 or 1978, Ingres worked very well, and we had about 100 external users. Arizona State University even agreed to use Ingres for their student accounting databases, but the whole project went to the dump when it became clear that by that time there was no COBOL programming language for Unix, and they used only COBOL, and this was the collapse of the project.
It became clear that if we want to do something different from the model one — get a serious introduction to real use of Ingres — we cannot do it with an academic prototype. Then we said, “Eureka!”, Let's create our company. This led to the creation of Ingres Corp. in 1980. Ingres worked really well.
After you have started (company), then you understand that if you have new ideas, if you want them to see the light, you either have to find a company to get the idea up or you have to start your own company. And my experience shows that it is many times easier, at least for me, to organize my own company. The general model is such that when I have new ideas and it seems to me that they have commercial prospects, then I would like to get a good working capacity inside the university to ensure the attractiveness of venture capitalists who are ready to invest.
Question: How do you decide which venture capital firms are right for you?
Answer: Venture capitalists need to get more than just money. First and foremost, what I want is for someone to work hard to make something successful. Secondly, I need someone who has really good contacts and connections. Most venture capitalists are really smart. And finally, I want to deal with someone I can get along with.
Question: How much do you think about the possible profit at the start of the company?
Answer: I never think. In the case of Ingres, the relational database management system, there were new ideas, and I was fascinated by them. They have experienced more than 25 years in the database market. It gives me great satisfaction. This leads to a paradigm shift. This is what fundamentally motivates me - change for the better.
Question: What advice would you give to entrepreneurs who want to start their own business?
Answer: What you need is a good idea and the backbone of the team, in order to move the idea forward, you need to have at least one sales or marketing director and super-competent programmers working on implementation. I find the venture capital community very competent in assessing these three characteristics. But if you do not have all the three components, I think it will be very, very difficult to get funding.
The biggest plus, as I see it, is that I now have experience, and it will be much easier for me to do the next project. But if you do not have a “credit history”, then the venture capital community will look askance at you. It's like in the book / movie "Trick-22" ( Catch 22 ).
Another thing I want to add is that the software of modern systems is becoming more and more complex. Federal government agencies, which usually provide money for research, are now virtually unwilling to invest in a commercial profile infrastructure. So if you want to get funds that really work, you have no choice but to look for it outside the traditional channels of research funding.
Q: Now you have CTO positions in two companies, VoltDB and Paradigm4 . Are these your main projects at the moment?
Answer: Yes. It all started with Vertica . It became clear that you need to make data warehouse systems work quickly, organize storage in columns, where you need to organize column-to-column data, it works tremendously faster than line-by-line storage, where data is stored line-by-line. This made me think about “vertical” tools for the market, since in the data storage market, such vertical tools will blow conventional tools to pieces. VoltDB is a tool for the vertical market, it focuses on the operational processing of transactions. Paradigm4 is a specially-created tool for the vertical market, which is focused on the segments of integrated analytics and science, and uses the characteristics of this market to get a realization that is completely different from both Vertica and VoltDB .
I am a passionate fan of specially-built tools, because “One size cannot be suitable for all cases”, and that if you decide to take advantage of the characteristics of specific vertical markets, you can get a solution at a speed of one to two orders of magnitude faster than using standard solutions. Current - I affectionately call them "elephants" - the main suppliers of relational databases, "cut one size fits all", providing a standard tool for all occasions.
Drastic changes occur, because if you asked a question a few years ago what people want from analytics, they said: “I want to know the average salary of employees grouped by departments”. But in many vertical markets, people have found out that they want to have ways for more complex analytical algorithms, for example, self-learning algorithms. The easiest way to break it down is that people who do electronic trading on Wall Street, they mainly build stock market behaviors, and then when it doesn’t work according to their models, they make purchases or sales. They make a very complex analysis. There is a whole range of markets where complex analytics decide. Accordingly, such solutions as Paradigm4 will work very well, because they are based on fast self-learning algorithms.
Question: Do you have any other plans in your plans?
Answer: What I’m thinking now is a huge problem for most global enterprises — data integration . A typical large enterprise has several thousand production database systems, as a rule, written separately, in different groups, without thinking about the tasks of integrating all of their data. I think about this. This is what I am working on at the Massachusetts Institute of Technology. If I come up with something that is commercially viable, then I will think about starting the next startup. For me, this is entirely due to a good idea, because without a good idea one cannot move anywhere.
Source: https://habr.com/ru/post/115598/
All Articles