The report evaluated 15 vendors based on 26 criteria, including data consistency, performance and scalability, multimodel, and data security. MongoDB scored five out of five in 21 of 26 criteria.
You have to put your info in to read whole report. Worth it IMO, but it’s probably inappropriate to post what’s on the other side. Mongo is not just a leader they are the leader.
Darth, I love your caption… thought it must be emphasized…
Mongo is not just a leader they are the leader.
and this is emerging, large white space…
How often can you find a combination where
you can be very confident of a market thats still emerging but WILL BECOME huge
you can be very confident that there is ONE WINNER TAKES MOST…
and you can be confident on WHO that winner is
and you can INVEST IN THAT WINNER in pure and direct way (not as a group in some other company)
May be over last two decades, you can count such opportunities on your fingers… Salesforce, VMWare (if you count that tracker stock as direct investment), Workday, Facebook, Netflix… may be a few more but gets difficult…
Apple, Amazon are more complex entities and transitioning from one to another… but may be you can add them…
I think ZS is another such play… ESTC may get there …
No TWLO, AYX, TTD not in that category… tough to call any of those in the market that will certainly be “Winner take most”…
But none of these are as clear as MDB… and it is fully supported by market sustaining its high price through December melt down and through all the FUDs…
And talk about selling MDB??? I cant think of that for next few years… I wont be able to sleep if I sell MDB… no way.
We’ve been here before:
First in the 70s and 80s with the IBM product IMS DB/TP. This product was the replacement for tape mounted serial batch systems. Even as tape data was copied to disc drives, the applications were coded in such a way as to be unable to utilize the random access capabilities of disc. Also lacking was a large scale teleprocessing environment necessary for remote access. There were competitors, lots of them, but IMS ruled the majority of the database world.
Then came Oracle with a relational model and UNIX to provide remote access. Oracle released the first relational DBMS in 1979, but they didn’t really hit their stride until the late 80s. Again, there were a slew of also ran competitors, but the majority of the market went to Oracle. It didn’t take long for Oracle to dominate the market. They were the first (even though the theoretical underpinnings were developed by E. F. Codd, an IBM scientist). IBM faced the dilemma of cannibalizing IMS sales so they dragged their feet. Eventually, when the entered the market with DB2 they were a day late and a dollar short and it only ran on IBM h/w while distributed UNIX was rapidly becoming the way of the computing world. (BTW, Codd quit working at IBM due to well publicized frustration with management).
For quite some time there has been, and I assume there will continue to be special use DBMSs. Geospatial data has at least a few DBMSs designed specifically to retain and manipulate data about objects located in three dimensional space (a number of relational DBMSs now have special extensions for this data type). Teradata has purpose built machines and a DBMS for data warehouse applications. I imagine game developers have purpose built databases. I’m sure there are other products that have been designed to address special data types and situations.
We are at this inflection point again. But, when it comes to a general purpose DBMS suitable for most business applications, the jury has already returned the verdict MongoDB is the winner. Will it too be disrupted? I’m sure it will, but I can’t predict with what. But, I’ll go out on a limb, barring a monumental management messup I’d wager that this investment is probably good for 10 years or more.
I’m sorry, Brittlerock, but I have to take exception to your history*. In particular, there were a number of flat file and ISAM approaches which preceded the move to relational. The pre-relational versions were limited in performance compared to the relational and had redundant data and other limitations, but were similar in that there were fixed record and field types. The relational databases reduced redundancy, improved performance, added multiple indexes on the same table, and many other features as time went on so that they were clearly better at solving the same types of use cases as their predecessors, only better.
What is going on with NoSQL and document DBs in particular is a recognition that not all data is nicely structured in the relational pattern and creating DB technology to deal with that data … in Mongo’s case, specifically data organized into documents. This is data that is not dealt with very well in the relational model, but which Mongo and other deal with very well. So, the NoSQL databases aren’t even trying to be a replacement for relational databases … except in those cases where a document application has been shoehorned into a relational model. NoSQL is addressing a new family of use cases which had been poorly addressed previously.
Not directly relevant to the DB replacement issue but I was a little amused by your characterization of the 70s and 80s as one of the things I did in 1969 and 1970 was serve as a principle architect for an on-line college selection system. This used a Burroughs B3500 computer with an application written in COBOL and bits of assembler to do logical add operations on “property vectors” some 3500 properties in length. The B3500 had a limited number of ports, so we used 4 mini computers with 64 ports each who would compose messages to send over a pair of ports to the B3500. We had to have the phone company come out to specially condition the lines to support 110 baud! A wee bit different than a tape based system or serial access to disk files.