German philosopher, Arthur Schopenhauer once said, “All truth passes through three stages. First, it is ridiculed. Second, it is violently opposed. Third, it is accepted as being self-evident.”
When it comes to NoSQL, it appears that Oracle is trying to cover all of the bases. In a recent eWEEK article outlining a perplexing argument against NoSQL, Oracle ridiculed NoSQL, violently opposed it, and then said that they have NoSQL, too. I’d like to address several key points they tried to make in the piece and provide some real world counterpoints and perspective.
First, they claim that NoSQL is nothing new – it didn’t work in the past and it doesn’t work now. Claiming that modern NoSQL is the same as pre-relational technology betrays a deep misunderstanding of what modern NoSQL is about. Not only is NoSQL modern in comparison to pre-relational systems, it is modern in comparison to 35-year old relational technology as well. For example, NoSQL brings architectural innovations such as scale-out clusters that completely change the economics, acquisition processes, and agility of operating a database. Even more importantly, NoSQL databases are schema agnostic which help solve the very problems that rigid relational databases created. Add in MarkLogic capabilities such as semantics, geospatial, bitemporal, and tiered storage and it’s difficult to take the idea that NoSQL is nothing new seriously.
Second, Oracle claims that NoSQL is not suitable for complex, mission-critical projects. Granted, NoSQL is a very broad field with dozens of entrants. And while it is certainly true that many of these vendor technologies are not designed to support crucial enterprise applications, MarkLogic was specifically developed to do just that. Don’t take my word for it, just look at some of the most complex IT projects in existence to find MarkLogic at the core – replacing Oracle. Let’s start with Healthcare.gov. To launch this nationwide health insurance marketplace, the Centers for Medicare and Medicaid Services (CMS) had to bring together data from 50 states, numerous federal agencies, and insurance companies. They worked on this for several years before realizing that they wouldn’t be successful using relational technology. So, CMS migrated to MarkLogic and we were able to provide them with all of the flexibility they needed along with enterprise grade security and reliability.
MarkLogic is also used by a growing number of the world’s top investment banks for some of the most mission-critical jobs you can imagine. As just one example, take JPMC where MarkLogic is at the core of the derivative trading system that handles trillions of dollars of trades. Hardly a simple or inconsequential task.
Given that over 50% of our business comes from completing projects that failed on Oracle, the notion that NoSQL is unsuitable for complex, mission-critical projects is completely without merit.
Finally, Oracle states that NoSQL is not very productive. And they go on to say that many problems can be solved with a few lines of SQL while NoSQL requires pages of code. Once again it is important to look at what’s happening in real world situations. Every major IT organization has many different systems with different schemas and data models. They all need to break the silos and integrate their data to achieve business value or satisfy regulatory requirements. When using relational technology there are months, or sometimes years, of data modelling work that has to happen before anything else can happen. That’s why many projects are abandoned.
By contrast, organizations using MarkLogic pull data from all of these different sources and start deriving value right away. Our customers take advantage of rich query capabilities that go beyond what is available in SQL to integrate structured queries, semantics, bitemporal, and search of unstructured data. In fact, organizations are much more productive with our NoSQL product than they are with relational.
By the way, at MarkLogic we understand that SQL tools and relational data are ubiquitous. That’s why we include SQL query capabilities in our product. We want to give organizations a bridge that makes their use of NoSQL as productive as possible.
Oracle goes on to say, “You can’t do things like reporting and analytics with these (NoSQL) databases.” In reality, with MarkLogic’s built-in SQL capability alongside it’s NoSQL capabilities, users can leverage their existing BI tools to perform discovery-based analytics that empower them to explore their data much more effectively than using traditional approaches. You need look no further than the US Intelligence Community for some of the most demanding requirements for real-time, discovery-based analytics operating on MarkLogic NoSQL.
For many enterprises, relational is the only database technology they’ve ever used. It’s what they’re used to and is therefore their choice by default. Not surprisingly, Oracle likes this situation. The hitch is that you can’t solve a problem using the same thinking that created it. Enterprises need to integrate all of their data from across silos for operational and analytical use. MarkLogic Enterprise NoSQL offers an advantageous new approach for achieving this.
I’ll close with more words of wisdom from the astute Schopenhauer, “Talent hits a target no one else can hit; Genius hits a target no one else can see.”
We at MarkLogic understand that proof of value lies only with customer results. If Oracle has honest questions about the value of Enterprise NoSQL, we invite them to learn from industry leaders who’ve used both MarkLogic and relational technology in their operations. No genius required – the results are clear enough for anyone to see.