Optimize with a SATA RAID Storage Solution
Range of capacities as low as $1250 per TB. Ideal if you currently rely on servers/disks/JBODs
While you were sleeping, 10Gen -- the company behind the MongoDB document database -- changed its name to MongoDB, Inc., raised $231 million, and became the first billion-dollar open source startup. That's right, an open source NoSQL database startup has a $1.2 billion valuation.
Founded in 2007, MongoDB got to this point in only six years, a feat that took Red Hat (founded 1993) nearly two decades. By contrast, the last open source startup I was involved in raised less than $20 million and sold for less than $400 million. At the time, this was considered pretty good.
To what does MongoDB owe its success? Oracle! That's right, Oracle is the best thing that ever happened to MongoDB. Oracle has the lion's share of the $30 billion database market. In 2011, the last year for which I can find database revenue numbers, RDBMS licensing revenue hit $16.75 billion. According to the analyst firm Gartner, Oracle owns 48.3 percent of the RDBMS market. It's the biggest, fattest gorilla imaginable.
Big red legacy
Oracle has many great advantages, beginning with an entrenched installed base. Many internal IT applications are written in Oracle's stored procedure language PLSQL.
But Oracle is not fundamentally different from the database I learned to love and hate in the mid-'90s on HP/UX PA-RISC boxes. In fact, it hasn't fundamentally changed since the '80s. Legacy is great and terrible at the same time. Oracle requires a lot of hardware and a good amount of support staff to keep running. It also does not affordably scale to the tens or hundreds of terabytes required by some -- or the millions of users required by others.
To scale in that way, Oracle would need underlying software architecture changes. Oracle is trying to address this by bolting other technologies onto its RDBMS, but as I've said before, this is a little like stapling a goose to a Mack truck and calling it an airplane. Moreover, that sort of scaling requires a fundamentally different license model. This is difficult to do without cannibalizing your existing market.
Meanwhile, Oracle has been a great consolidator of the database market. It gobbled up parts of MySQL (including its InnoDB storage engine), then Sun, which had in turn gobbled up MySQL. That was one move at the end of a long line of consolidations, which have been great for Oracle in its battles with familiar competitors like IBM and Microsoft. But that approach is also a great weakness when confronted with a new technology.
The disruptive force of NoSQL
The history of the tech industry is a history of disruption. Microsoft did it to Novell in the '90s, and Apple continues to disrupt Microsoft. You need an entrenched, immovable dinosaur that is well-adapted to its environment, but also unable to adapt its technology or cost structure quickly enough to new market conditions. It's like the Aztecs: very adaptable to their environment but helpless versus Spanish guns and smallpox. The successful insurgent uses the entrenched player's weaknesses against it while living naturally in a new technical environment. The insurgent is an agent of change, rather than a force to keep the technical and business climate in check.