Relational Databases: Gradually, then Suddenly?

In The Sun Also Rises, the Ernest Hemingway character is asked “How did you go bankrupt?”. His response ‘Two ways. Gradually, then suddenly”

I thought of that relative to relational databases during a meeting with Hasso Plattner of SAP in Orlando recently. There was a palpable sense of vindication that his view on columnar databases like HANA has been tested and validated by thousands of customers

“Yeah, that's mainstream. I feel good about that, ten years ago, I had an idea which really worked: streamline ERP applications to basically record only transactions.

And, for everything else there are algorithms to exploit the data….Facebook doesn't know what aggregates are. Why do we have aggregates? They have no aggregates.

Today, it's all on demand. People are not permanently rolling up and looking at rolled up data. And, if they do that, then it's fast. If they want to change it, it's instant.

That is really gratifying”

Then I read this Werner Vogels (CTO of Amazon) post

“We found that about 70 percent of our operations were key-value lookups, where only a primary key was used and a single row would be returned. With no need for referential integrity and transactions, we realized these access patterns could be better served by a different type of database. Further, with the growth and scale of Amazon.com, boundless horizontal scale needed to be a key design point--scaling up simply wasn't an option. This, ultimately led to DynamoDB, a nonrelational database service built to scale out beyond the limits of relational databases.”

He then proceeds to talk about a wide array of databases Amazon now offers, including a relational one.

I still believe relational databases are declining “Gradually”, not “Suddenly” about to hit the wall. New applications are increasingly running on non-relational options. Converting legacy applications will take a long time. Retraining DBAs, adapting systems management tools to support new databases for stringent SLAs will take time. It’s unclear if the Teradata lawsuit will slow down HANA adoption. (we met Hasso before that announcement).

But I am now prepared to say that the “Gradually” decline is starting to accelerate.