SNOWFLAKE - A CRITIQUE
Bill Inmon
Founder, Chairman, CEO, Best-Selling Author, University of Denver & Scalefree Advisory Board Member
SNOWFLAKE – A CRITIQUE
By W H Inmon
(The following is a personal critique. It is strictly my personal opinion. You are welcome to do your own research and come to your own conclusions.)
Snowflake advertises themselves as a data warehouse on the cloud. In my opinion they are not a data warehouse at all.
In order to be a data warehouse an organization must integrate data. If you don’t do integration of data then you are not a data warehouse. It is true that Snowflake ingests data. It is true that Snowflake places the data on the cloud. It is true that Snowflake can handle large amounts of data. These things are all true.
But when it comes to integrating data
In truth the Kimball model was for data marts, not for a data warehouse. A data mart and a data warehouse are fundamentally different things. I don’t think Kimball ever understood that distinction.
The truth of the matter is that without integration of data – transformation of data
领英推荐
In order to build a data warehouse you need the process of ETL – extract/transform/load
It would be nice if Snowflake would advertise themselves as the EL model. What is going to happen is that people are going to put up data into Snowflake and think they are getting a data warehouse. Then, after a while people will grow tired of the siloed, unintegrated data in the Snowflake environment and data warehouse will get the blame of failing. When in fact there never was a data warehouse. People thought there was a data warehouse with Snowflake but it was never there.
Snowflake has solved the problem of handling large amounts of data
You can put a big mess on the cloud. But the cloud does NOTHING to solve the problem of having a big mess.
What do you have when you put a big mess on the cloud? You have a big mess on the cloud.
It is still a big mess regardless of where it sits.
Bill Inmon lives in Denver with his wife and his dog Jeb. Jeb goes for his afternoon walk around the neighborhood unless it is snowing or icy. Jeb is a scotty. When people see Jeb walking they like to come up and pet him. Jeb likes that and wags his short tail at them.
Advisor, Co-Founder | Digital & Agile Transformation, AI, LLM, Data Driven Solutions, DLT, web3
2 年Let’s take these thoughts one level upwards and look at the success factors of any IT related activity: People - Processes - Tools although I truely love the ?Tools“ part, the ?People & Process“ (mindset, design paradigms / concepts/ capabilities, business / data understanding etc. etc.) dimensions are much more important. So Bill Inmon I would agree to you in this sense: don‘t buy a tool to solve your design problems (but don’t blame the tool if you do)
Data & AI Advisory Solution Architect
2 年Simon Ander
RWE Data Manager at Sanofi CHC
2 年Hi Bill, as someone who have studied statistics but ended up doing data engineer and learning it by doing in the context of tools like Snowflake, Airflow ... maybe I did not have access the the right bibliography and historical perspective of the concepts you mention and on which you base your critic. If you see this and have time to answer I would be very pleased to have some guidance on where to find learning materials to better understand your point of view. Best!
"In order to build a data warehouse you need the process of ETL – extract/transform/load. The Snowflake model is the cloud version of E and L without the T." I don't understand that at all. The vast majority of engineer time is spent transforming data in Snowflake. Why wouldn't it be? Bill, why do you say Snowflake doesn't support that? It doesn't make sense.
Senior Data Engineer | Blockchain Data Engineer | Web3 Analytics | Expert in ETL/ELT, SQL, Data Lakehouse, Apache Spark, Apache Hudi, Delta Lake, Apache Iceberg & AWS | Co-founder of Data Train Community
2 年Francisco Arpino Magioli