Technically Speaking

The Official Bigstep Blog

 

Is Splice Machine a Viable Option for Your Hadoop SQL Database?

In the grand land of databases, you have the traditional RDBMS (here's lookin' at you, SQL) and an impressive lineup of the sexy, modern NoSQLs (say hey to MongoDB, Cassandra, Redis, HBase, and the gang). The trouble with relational databases is and always has been scalability. Darn things just don't like to grow, and today's data sets do enjoy GROWING. But RDBMS retrieve data like nobody's business. Conversely, NoSQL databases are ACID-less.

In the grand land of databases, you have the traditional RDBMS (here’s lookin’ at you, SQL) and an impressive lineup of the sexy, modern NoSQLs (say hey to MongoDB, Cassandra, Redis, HBase, and the gang). The trouble with relational databases is and always has been scalability. Darn things just don’t like to grow, and today’s data sets do enjoy GROWING. But RDBMS retrieve data like nobody’s business. Conversely, NoSQL databases are ACID-less.

Dealing with those massively growing data sets usually comes with Hadoop. Hadoop also scales extraordinarily well, but it’s a doozie of a mind trip for those used to relational databases.

Splice Machine: The Best of Several Worlds

SQL-esque usability—Check! Hadoop-style scalability—Check! Spark-ish speed—Check! Splice Machine offers the best of a number of different worlds.

Splice Machine is built to solve all the problems. Though it is a relational database, it’s also immensely more scalable than RDBMS like SQL. It’s also ACID-compliant, and is built on top of HBase, a NoSQL database that utilizes Hadoop’s HDFS as storage. This is a beautiful blending of the relational DB, ACID, horizontal scalability, and HDFS, all rolled into one.

Splice Machine version one wasn’t shabby at all. But now it’s gone open source, and will soon be into version two, a double-whammy of solid improvements. The open source version is pretty much the same thing as the paid version, except for the obligatory stripping of a few enterprise-grade features. Oh, and it also integrated Apache Spark, making it one of the most well-rounded databases around.

That’s one part Hadoop scalability, one cup RDBMS ease of use, a tablespoon of open source freedom, one dash Spark speediness, and BAM! Splice Machine no longer looks like the redheaded stepchild of databases. Eat your heart out, Mongo.

Open Source Versus Paid Splice Machine

Going open source allows software to get all of the rigorous real-word testing plus benefit from a wide range of innovative developers’ minds. But to the businesses using these products, it basically just frees them from proprietary lock-in.

Splice Machine is also integrated with Kafka for the slick streaming stuff. However, one of the enterprise features that got ripped for the open source option was encryption. The freebie version is also missing column-level access control and the backup and restore functions. But, hey, if these are essential, you can always get your feet wet with Splice Machine open source and dive into the deep end of the pool by purchasing the enterprise-class features when you’re ready. Kerberos support also comes with the enterprise version.

Open source is a popular option for several reasons. First, it’s a good feeder model to get people onboarded to the paid version. Second, those free developers tend to add some nifty features, and the product gets proven in real-world testing grounds, which is about the only way to get conservative business types to invest in it. But the elephant in the open source room is that businesses are trying to get away from vendor lock-in. They get access to a wide community of developers and support experts, without being held hostage by proprietary software licenses.

So, is Splice Machine for you? For all the bells and whistles, the real selling point is its SQL-like functioning combined with NoSQL/Hadoop-style scalability. For many shops, this is the Holy Grail of the database world. Integration with Kafka, Kerberos, etc. is just embellishment.

If you’re building a big data store, you’d be making a mistake not to consider a data lake. And you’re in luck! Bigstep is offering the first Data Lake as a Service ever. For a limited time only, you can discover the first Bare-Metal Data Lake as a Service in the world. Get 1TB free for life - limited to 100 applicants. Start here.

Got a question? Need advice? We're just one click away.
Sharing is caring:TwitterFacebookLinkedinPinterestEmail

Readers also enjoyed:

Low latency and big data

Low latency was once only really spoken about within financial trading firms. But in 2013 low latency networks are desired and used by businesses of all…

Feed Your Hadoop with Bare Metal

Once limited to internet giants like Google, Hadoop is moving into the business mainstream, allowing businesses to ingest and analyze massive quantities…

Leave a Reply

Your email address will not be published.

* Required fields to post your comments.
Please review our Privacy Notice in order to understand how we process your personal data and what are your rights in this respect.