Database Wars 2.0
For old guys like me ( I started my career in tech in 1987) we remember well the “Database Wars” of the late 80’s and early 90’s. Oracle, Sybase, Informix and DB2 battled for database supremacy. Newsflash, Oracle won.
For a separate people-operations perspective, this background covers documenting serious workplace conduct.
In the Beginning: Hegemony
For the next 20 years, give or take, Oracle dominated this market. Sure there were others. Postgres showed up in the late 80’s, and companies like Progress Software had viable database solutions as well, but ultimately none came close the hegemony Oracle exerted over the market. Why and how this happened is an important lesson we can learn from as we enter “Database Wars 2.0”.
Then Revolt
About 5 years ago we saw the first serious threat to Oracle’s dominance emerge via the NoSQL ecosystem. While open source relational DB’s like MySQL made inroads against Oracle in the previous decade, none really threatened them, and in the end Oracle simply bought them out. Microsoft SQL Server was always popular, especially with companies that already used lots of MS technology, but could never overtake the giant. One major reason why MS and MySQL could not break away to win was the technology was simply not compelling enough. The differences between all the major RDBMS were relatively minor. For MySQL, other than being open source vs. Oracle proprietary approach, users did not gain much advantage with one over the other. And SQL Server could never quite convince people it had the scale to run the big apps like Oracle did so successfully.
Fast forward to today. The rise of NoSQL is dramatic, and completely powered by a technically different approach to data management than the traditional RDBMS. And not only is NoSQL different, within NoSQL there are multiple kinds of databases, each with a different technical approach targeting different use cases. The term polyglot is often used to describe the rise and use of multiple DB’s within an organization for specific use cases. Most agree this will be the persistent state for enterprises going forward. The days of “standardizing” on a single DB seem to be in the rearview mirror.
History Doesn’t Repeat, But It Often Rhymes
This describes the current database wars quite well. The current war won’t be exactly the same as the last, but make no mistake, each player has the same goal. Be the next Oracle! The original RDBMS were all proprietary, whereas NoSQL DB’s are universally open source. And the internet was not really a factor in the first wars. So many things have changed. What seems to have not changed is the eerie similarity in the early stages of the current war to the original DB wars.
Our NoSQL database (insert vendor name here) is the “FASTEST” most SCALABLE”!
These are the claims being made by many NoSQL companies today. It’s the exact same argument all the vendors made in the first database wars. Everyone wanted to talk about speed and scale, as if nothing else mattered. The reality of speed claims is they have always been use case dependant. While a given DB might be lightening fast for one use case, it’s horribly slow for another, so generic speed claims just don’t work, in the past or today. Another approach is they tried to differentiate on small esoteric differences between their DB and the competition.
This never works, just ask Informix.
The NoSQL Warriors
It’s clear the NoSQL wars will be different. The landscape is more varied than before with lots of players: Key-Value Store (Riak, Dynamo), Document-based (MongoDB, CouchDB) and Column-based (HBase, Cassandra), Graph-based (Neo4J). Then there’s Hadoop. While not technically a NoSQL database, Hadoop is the other side of the NoSQL coin and often holds the same or similar data as NoSQL DB’s. So this time around will be more complex with higher diversity than the first wars, but the ground rules will be similar. Making generic claims about being faster or more scalable won’t cut it in the long run. And trying to carve out highly specific use cases is a great way to build a niche business, but it won’t make you the next Oracle. No, in order to win the war this time it’s going to take a bigger vision and some serious execution.
Formidable Opponents… And The Winner Won’t Take All
The NoSQL database ecosystem is the first real threat to the RDBMS/Oracle data management market in 20 years. The new wave of warriors are companies like MongoDB, Datastax (Cassandra) and Elastic. Each represents a different technical approach, but offers the same basic benefits to the user/developer.
- Flexible data model - gone are the days of rigid schema’s
- Hyper scalability and easy distribution of workloads
- Rapid deployment and application development
Compared to their RDBMS cousins, NoSQL is making developers lives much easier in many ways. However not all the RDBMS players are quietly conceding the market. PostgreSQL has added strong support for JSON (a primary data format of NoSQL) and Oracle has their own entrant into the NoSQL database market.
David vs. Goliath > Goliath vs. Goliath
Regardless, MongoDB has gone from zero to the number 4 most-used DB worldwide according to DBEngines most recent survey in less than 5 years, so clearly its “game on” for Database Wars 2.0. But it’s worth remembering we are still very early, and while some players have opened up strong early leads it’s far too early to declare a winner.
Part 2 is coming out very soon. I’ll share my thoughts on why Oracle won the first database war – and what we can expect over the short term. Spoiler: It’s definitely not Oracle.
SlamData is Visual Analytics for NoSQL
SlamData is an open source solution that makes it easy for people to see and understand modern NoSQL data, without relocation or transformation.
- Learn More
Characteristics of NoSQL Analytics Systems
Get the whitepaper that lays out the past, present and future of NoSQL Analytics. Written by John A De Goes, CTO of SlamData
- Download. Now.