Snowflake's Data Warehouse Runs on Relational Database in the Cloud

Relational databases were not designed from the ground up for the cloud or, for that matter, to handle big data. Nobody knows that better than Bob Muglia, former head of the Server and Tools division of Microsoft (MSFT).

As the CEO of startup Snowflake Computing, Muglia has taken the wraps off a Snowflake Elastic Data Warehouse service in the cloud, built on top of a relational database that Snowflake developed specifically for that purpose.

In addition, Snowflake Computing announced it has picked up an additional $26 million in funding, part of which will be used to create a customer and partner ecosystem around the new cloud service.

After a brief detour at Juniper Networks, Muglia is back in the software infrastructure business. He said the Snowflake database enables data warehouse applications to independently scale users, data and workloads without there being any underlying contention for resources.

On top of that Snowflake has developed a new storage system for semi-structured data based on JavaScript Object Notation (JSON) data inter-change format that layers in on top of the company’s relational database, which enables analysts to query structured and semi-structured data running in the cloud. The unique thing about Snowflake, said Muglia, is that for the first time the JSON format that is at the core of so many modern Web applications has been married to a modern relational database.

Given the inherent cost advantages of running databases at scale on Amazon Web Services (AWS), Muglia said that data created in the cloud most likely will be best served by a data warehouse that also runs in the cloud. Muglia doesn’t envision every data warehouse application moving into the cloud. But he also notes that it doesn’t make much sense to incur the expense of moving data created in the cloud to a data warehouse running on premise.

It’s a rare event when a company of any size brings a new relational database to market. The vast majority of the data warehouse services in the cloud today are based on so-called NoSQL databases. While those databases typically support SQL, they generally make use of some type of key/value store database architecture that doesn’t make it as simple to keep track of the relationships between all the data stored in that database.

Regardless of the approach taken, it’s clear that databases-as-a-service (DBaaS) are rapidly emerging as a new sector of the cloud that offers great promise for all types of big data analytics applications.