Three NoSQL Databases You’ve Never Heard Of
There’s a certain class of data problem that is elegantly addressed by NoSQL databases, which is why the market for NoSQL databases is growing faster than the overall market. The market is led by the Big Four, including Couchbase, Datastax, MarkLogic, and MongoDB, but there’s a long tail of other players in the NoSQL market, including some older products that are still going strong.
If you read this newsletter, you’re probably aware of the some up-and-coming NoSQL players like Aerospike, Basho, and Redis, as well as specialized products like the Neo4j‘s graph database, Splunk‘s log database, Elastic‘s search engine, and Sqrrl‘s security-focused NoSQL.
But there are many, many more NoSQL databases out in the wild doing some interesting things. Here are three of them you’ve probably never heard of.
Before there was Apache Cassandra—heck, even before Oracle was a gleam in Larry Ellison’s eye—there was MUMPS.
MUMPS, which stands for Massachusetts General Hospital Utility Multi-Programming System, was first released way back in 1966 to handle the inflow of massive amounts of transactional data into early healthcare applications.
The 51-year-old database sports features of NoSQL databases, as well as features of relational databases, namely ACID (Atomicity, Consistency, Isolation, Durability) compliance. At its core, MUMPS is a key-value store, and is able to store data in a schema-less fashion, which is handy for less-structured clinical data that a hospital may need to store. The database is inherently parallel, horizontally distributed, and runs on commodity hardware.
Today, the biggest backer of MUMPS is the Cambridge, Massachusetts-based company InterSystems. In 1997, InterSystems launched a product called Caché, which is essentially a commercialized version of MUMPS; the two products share the same multidimensional array structures for storing hierarchically structured data.
InterSystems’ biggest customer is Epic Systems, the electronic health record (HER) software giant that uses Caché as its internal database. Seeing as how about half of the hospitals in the United States rely on Epic’s EHR, and that MUMPS (via Caché ) is the database powering both transactional and analytic uses, it’s clear that MUMPS is one of the most successful and widely deployed NoSQL databases that few have heard of.
Despite the anonymity of MUMPS, InterSystems is getting some attention with Caché. Gartner featured Caché in the leader’s quadrant of its Magic Quadrant for operational database management systems in 2015, and the company is looking to ramp up its visibility in other industries besides healthcare.
Like MUMPS, c-treeACE (Advanced Core Engine) is an older multi-modal data store that straddles the NoSQL and relational database worlds.
The software, which is a key-value store that runs on a variety of symmetric multi-processor (SMP) servers (in other words, it’s not distributed), was developed by Columbia, Missouri-based FairCom in 1984 to provide a b-tree-like data store.
The database is distinguished in part by its dual access methods. Developers can either use a record-oriented, Indexed Sequential Access Method (ISAM) structure, which offers fast and responsive access to files and indexes. Or they can use SQL to access data or run queries, which many programmers are familiar with.
c-TreeACE, which FairCom calls a “No+SQL” database, offers full ACID guarantees, making it suitable for critical business applications. That’s something that not all NoSQL databases can offer. The flexibility to support unstructured data, however, puts the database firmly in the NoSQL camp.
Since FairCom was founded in 1984, it’s sold thousands of licenses for c-treeACE, which Gartner includes in its “niche players” category in the 2015 Magic Quadrant for operational databases.
The company claims that 43 percent of Fortune 100 companies are customers, including UPS, IBM, Thomson Reuters, and Visa. Many of these customers have embedded c-treeACE into other applications, which is one of the biggest use cases for the database.
FinchDB is a distributed NoSQL database that combines analytic and search functionality atop an in-memory document store.
FinchDB software is backed by Reston, Virginia-based Finch Computing, formerly Synthos Technologies, which is the products division of Qbase. According to Finch, the database’s most distinctive feature is that models are embedded in queries. “They change as the data does, and users can pass-in dynamic context on every query,” the company says.
FinchDB was recently named a “Cool Vendor” in in-memory computing by Gartner. The IT analyst group said some good things about FinchDB, including that “application architecture leaders looking for a platform to support real-time contextual big data analytics should look at products such as FinchDB…”
It’s been said that over 100 NoSQL databases exist in the world, so we’ve undoubtedly missed some that are worth mentioning. If you’re using a notable NoSQL database that you think is worth mentioning, then by all means drop us a line.