Robin Schumacher

Oracle and the NoSQL Effect

By Robin SchumacherJune 19, 2015

Oracle’s recent fiscal Q4 miss, which boiled down to an eight percent license revenue shortfall was, “the largest we’ve seen in memory and is surprising,” said Citigroup’s Walter Pritchard. Oracle’s announcement rounded out what’s been a disappointing year for the company where its total revenue for the past twelve months was flat from the previous year, at $38.23 billion – an even worse performance than in 2014 when its revenue grew only three percent.

While I certainly don’t think Oracle is doomed, I’d like to provide some perspective on what I believe to be the effect that open source and NoSQL are having on Oracle, and more importantly, what’s causing it. I’ll also say upfront that I do think that what they’re experiencing now is just a fraction of what’s to come.

The Switch From Master-Slave to Slave-Master

Let me tell you a quick story, which describes why I think Oracle is struggling right now.

I loved being a DBA (especially an Oracle DBA), and I’ll admit I felt some power in being one of the “data gods” in the various companies where I worked. In countless meetings that I had with application teams throughout my DBA career, I dictated to them what they could and couldn’t do with the apps they were building. If their app needed to do something the database didn’t support, well, sorry about your luck.

But when web apps began to come on strong, something changed. Suddenly, the apps teams didn’t take ‘no’ for an answer where database requirements were concerned. The fact is, they couldn’t because web apps ushered in a new model for development that demanded the database bend to its rules.

The database suddenly found itself living out a lyric from an old song by the Police: “you’ll find your servant is your master” and became somewhat wrapped around each web application’s finger. Or to put it in database terms, the database went to being a slave vs. the master.

When I ran product management for MySQL, we were scolded and given 39 lashes by database purists who accused us of breaking the rules of standard RDBMS practice. However, the fact is we had to in order to accommodate what web apps needed from their database. The end result was that MySQL became a huge success because of what it delivered to web v1.

Then things got even harder. The web continued to evolve, with mobile and IoT coming on the scene, which produced even tougher database requirements; prerequisites that exceeded what an RDBMS (even MySQL) could do. Things like:

“I need more than just failover. I need constant uptime with no chance of any outage.”

“I need to distribute enormous amounts of data anywhere in the world.”

“I need to be able to write (not just read) insanely fast amounts of data coming in from everywhere in the world.”

“I need to be able to double or quadruple my capacity at a moment’s notice.”

“I need to maintain constant database performance no matter the user or data load.”

“I need to handle all kinds of data efficiently without storage overload.”

“I need to do all the above and more without breaking my budget.”

And that brings us to where we are today…and why Oracle is struggling.

You Can’t Get There From Here

At a recent investor conference I was asked about what the near-term future holds for RDBMS’s and NoSQL. My answer was the following:

“Let me be clear. Relational databases are fundamentally ill suited to handle today’s distributed systems. Their master-slave architectures, methods for writing and reading data, and data distribution mechanisms simply cannot meet the key requirements of modern web, mobile and IoT applications. I tell you that not as an employee of a NoSQL company, but as a guy who has worked with RDBMS’s for over twenty-five years. In short, you simply can’t get there from here where relational technology is concerned, and that’s why NoSQL must be used for the applications we’re talking about.”

Again, I am not saying Oracle is doomed, but I am saying they are hurting because today’s web, mobile, and IoT applications have requirements that supersede an RDBMS’ capabilities, and nothing the relational vendors do to their engines will change that. Let me show you what I mean.

Here is a graphical depiction from Oracle’s latest documentation for their global data services:

Oracle Global Data Services

First, get a handle on all the various components needed (e.g., Oracle RAC, Dataguard, Goldengate, etc.). Then keep in mind it still only provides failover and not continuous availability, is very difficult to expand, will struggle with ingesting huge amounts of fast data, does not have a flexible data model, is unable to write data wherever needed, and will cost a princely sum of money. Lastly, remember what you’re looking at is just for two locations.

Now here’s how a NoSQL database like Cassandra works across multiple geographies and clouds:

Cassandra ring multi-dc and cloud

Cassandra’s masterless architecture, fast log-based write and read storage framework, multi-data center and cloud replication functionality that delivers data wherever you need along with constant uptime, “write anywhere” capabilities, lack of special hardware or add-on software needs, and flexible data model are able to elegantly and cost-effectively meet the requirements web, mobile and IoT apps have for their database. The best that Oracle has simply cannot.

This is why enterprises like Netflix and Intuit have switched their web/mobile applications from Oracle to DataStax. This is why one of our startup customers, jKool, uses Cassandra instead of Exadata (which they watched fail) and said about Oracle, “this is not the future; this is the past.”

This is why tech writers like Matt Asay and Matt Weinberger point to open source NoSQL databases as a real concern for Oracle. Perhaps this is also why Cowen & Company’s Mid-Year 2015 IT Spending Survey showed DataStax and other NoSQL vendors now taking up a decent portion of the pie for expected database and analytics purchases:

Cowen survey 2015 database

Cowen survey 2015 analytics

Oracle’s problems can’t be shrugged off by comparing the current revenues between them and a company like DataStax. The crux of the matter is squarely focused on understanding the absolute database must-have’s for web, mobile and IoT applications, and how open source NoSQL solutions like DataStax were built specifically to handle them, while Oracle can’t and won’t be able to in the future.

This is why I also think we are seeing just the tip of the iceberg now. With NoSQL becoming a requirement for doing business in web/mobile/IoT, things should only continue to grow, just as the best-selling book Crossing the Chasm states: “When markets go mass, platforms have the advantage.  Because they can participate openly in many value chains at the same time, they can be taken up in multiple segments simultaneously.  And when they become identified as simply a requirement for doing business . . . then they deploy like lightning” (my emphasis). 

For more information on how Oracle compares to a NoSQL solution like Cassandra and DataStax Enterprise, visit our RDBMS to NoSQL web page and give our Oracle comparison paper a read when you can.

 




SHARE THIS PAGE
SUBSCRIBE

Comments

Your email address will not be published. Required fields are marked *

Tel. +1 (650) 389-6000 sales@datastax.com Offices France GermanyJapan

DataStax Enterprise is powered by the best distribution of Apache Cassandra™.

© 2017 DataStax, All Rights Reserved. DataStax, Titan, and TitanDB are registered trademark of DataStax, Inc. and its subsidiaries in the United States and/or other countries.
Apache Cassandra, Apache, Tomcat, Lucene, Solr, Hadoop, Spark, TinkerPop, and Cassandra are trademarks of the Apache Software Foundation or its subsidiaries in Canada, the United States and/or other countries.