RethinkDB describes first-generation NoSQL technologies with two categories, developer-oriented and operations-oriented. Here’s what they have to say:
Developer-oriented products include MongoDB and CouchDB. They typically pay close attention to ease of use, have rich document structure, and offer flexible querying capabilities. However, when compared with their operations-oriented counterparts, they are more difficult to deploy to sharded environments and to scale to significant load.
Operations-oriented systems include Cassandra and Riak. These products are designed for highly available deployments and high scale. Unlike developer-oriented products, in their current form operations-oriented products typically have less powerful querying capabilities, and tend to focus less on ease of use.
Although RethinkDB only mentions NoSQL technologies, this distinction applies to databases in general. Using a database to develop software is a completely different skillset from ensuring that database never fails. Although concepts like DevOps try to confuse the issue, the reality is that executing queries against PostgreSQL bears little resemblance to the work of scaling PostgreSQL across hundreds of servers and multiple datacenters. Most migrations away from a database occur not because it lacked features, but because operating it required so much expertise. But don’t take my word for it: operating databases is hard.
Technologies like RethinkDB try to solve this by being both easy to use and easy to operate, but such tech is in its infancy, and structurally, the skills of using and operating a database will always and inevitably remain distinct. Consider how the skills to use a computer are so wide-spread that even infants get it, but only specialists build and repair such machines.
Other database systems, like Orchestrate, take the difficulty of operations off the table entirely, because our team operates it. Much like how Heroku or Nodejitsu make deploying web applications effortless because their teams understand deployment, we lend our operational expertise to everyone who uses Orchestrate, so you can focus on building better software.
Everyone promises their technology is resilient and reliable, but bad stuff happens. Hardware fails. Earthquakes hit datacenters. Crucial packets disappear. Servers misbehave. Users put unexpected data in unexpected ways. Every system fails. When it falls apart at 3am and you need to debug live, production systems on two winks of shut-eye, will you have the expertise to fix it?
Even as most of the foundational literature underpinning distributed databases dates to the 1970s, the entire industry continues to struggle with making datastores human-friendly.
We built Orchestrate to turn our operations expertise into mature database infrastructure for developers everywhere, and we continue to improve it every day. We combine best-in-class tech to provide expert reliability and familiar features. We stand on the shoulders of giants like HBase, ElasticSearch, Hadoop, and Zookeeper. Our database nerd compatriots and predecessors have built technological wonders; rather than reinvent any, we use everything for what it’s built to do.
Everyone in the industry manifest a different approach, so sign up and see what you think of ours. Happy coding!