I recently attended Oracle OpenWorld 2016 in San Francisco and listened to Larry Ellison’s keynote. As reinforced in this statement, Ellison claimed that when compared side by side, Oracle is twenty years ahead of Amazon databases in the cloud.
Ellison also stated that Oracle Database as a Service (DBaaS) yields faster workload performance. He claimed that Oracle’s cloud database service is 24x to 35x to 105x to 1000x better than AWS, depending on the workload. He also asserted that the Oracle Cloud is optimized for running Oracle Databases, while Amazon Web Services (AWS) is not. Of course, Amazon has previously touted its Aurora RDS service as having 5x performance over standard SQL.
We can debate whether the measures that Ellison used were fair and accurate, but fundamentally, I think that he was asking the wrong question. The specific performance metrics of databases in the cloud is not the most important topic to assess, but rather if the cloud provider offers you the choices to find the best database technologies to suit your task and does so without locking you into one particular solution.
Let’s take a look at some of the claims and what they mean.
Are Amazon’s Database Capabilities 20 Years Behind Oracle?
While I agree that Amazon may be years behind in certain capabilities, there are also certain areas where they are undoubtedly ahead. There is so much innovation in the database space. However, just because something worked over the past twenty years doesn’t mean it's the best choice at present. It seems Ellison is returning to his roots of competing on performance here given his history of beating DB2, Sybase, and other relational databases using ads with the classic Oracle “big red bars” comparing system performance.
Though AWS may be trailing compared to Oracle in raw performance and some enterprise features, AWS is bringing more choice and flexibility to their users. For example, AWS not only offers several different relational databases as part of RDS but also NoSQL and data warehouse tools in the form of DynamoDB and RedShift.
Looking at the bigger picture, it’s about the whole toolchain and not just about the database; the more flexibility one has in selecting the tools in their toolchain, the better. It’s about choosing the right cloud for one’s workload. Amazon gives users a broader set of tools than Oracle including things such as Elastic Beanstalk, Elasticsearch, Kinesis, Data Pipeline, to their recent “serverless” offering, Lambda. Of course, the flip side of this is that the more Amazon specific tools you take advantage of, the more you become locked into their Amazon’s application program interfaces (APIs) and technology. This brings us to our next topic.
How Do You Define “Choice”?
Ellison claimed that one of the main benefits to Oracle’s database services is that their databases can run anywhere. With Amazon, you can only run Aurora, Redshift, and DynamoDB in AWS, giving you much less choice. Ellison points out that Oracle runs everywhere with a variety of choices: on the hardware of your choosing, specialized machines, and even on their public cloud. It’s true that Oracle has been investing in enterprise database capabilities for many years and that they are ahead in feature functionalities in the enterprise. Oracle RAC is still the preeminent database architecture supporting enterprise class applications. That said, this doesn’t mean that Oracle database is the best tool for every job. The proliferation of NoSQL, NewSQL, and open-source databases in recent years is evidence that developers are leveraging new data management architectures to optimize specific use cases more than ever before.
This then leaves us with a challenge. Ellison argues that he’s providing more choice by letting you work on any environment, public cloud, private cloud or traditional hardware in your data center, as long as you’re willing to use Oracle (a bit like the old line from Henry Ford that you can have any color Model T that you want as long as it’s black). Amazon, on the other hand, will provide access to a wider range of database types but they will only work on AWS.
What does this mean for the current database market? While it’s great to have choice, most enterprises are going to take advantage of the tried and true technologies, no matter the provider.
Are the Market Factors in Oracle’s Favor?
In his OpenWorld keynote, Ellison also introduced Oracle’s Exadata Express Cloud Service as the new entry-level version of the Oracle Database in the cloud, with pricing starting at $175 per month. While that’s great, I think there are other factors to consider besides the raw numbers. Looking at the rapidly changing database market, both Oracle and Amazon are taking advantage of enviable market positions. Oracle has an incredible market reach in terms of the number of apps supported, customers with enterprise site licenses, etc.
However, Amazon is coming from a different place; they are operating at a scale where market factors are extremely in their favor, making Oracle the underdog. I think the focus here should be with the consumers. Database as a Service is only interesting in terms of the cloud in which you are operating your database; consumers should, therefore, be looking to have more power over their cloud. Open APIs, standard based protocols, and open-source software allow for this flexibility and control. As an increasing number of consumers realize this, we’ll see the market dynamics shift.
Oracle and Amazon are two of the biggest names in the database market and will continue to be as we see the market grow and change. Consumers have many options when it comes to how they manage their databases in the cloud, and the most important thing is to get the most choice out of their database provider.