In this blog series, we’re showcasing the top 10 reasons customers are turning to MapR in order to create new insights and optimize their data-driven strategies. Here’s reason #7: MapR provides the top-ranked NoSQL key-value database for current offering.
In his October 7 blog post, my colleague Jim Scott described why the MapR Data Platform, which exposes a full read/write file system that adds important benefits to Hadoop, is one reason why customers choose MapR. He touched on the real-time aspects of MapR, and how our integrated in-Hadoop NoSQL database, MapR-DB, delivers high performance and consistent low latency. Not surprisingly, MapR-DB is a key reason why customers choose MapR.
Forrester Research ranked MapR-DB as the strongest "Current Offering" when compared against 14 other leading NoSQL big data technologies.
Download the full report: The Forrester Wave™: Big Data NoSQL, Q3 2016Download Now
Performance and Consistent Low Latency
You might already be aware of the record-setting speed (MinuteSort and TeraSort) that the MapR Distribution can give you. Since MapR-DB runs on that same architectural platform, it also gets the same fast I/O benefits for high throughput. Thanks to the speed advantage, MapR customer Atzmon Hen-tov of Pontis notes, “MapR-DB requires about half the machines compared to other platforms (Apache HBase and other K/V stores). This dramatically reduces the cost of a new system.” Consistent low latency is another critical requirement for our customers, and you get that with MapR-DB. If you’re processing terabytes of data and you need to make sure your system runs optimally, MapR-DB avoids the delays associated with housekeeping tasks like compactions and garbage collection seen in other technologies.
Apache HBase API
MapR-DB has its roots in Google Bigtable, just like Apache HBaseTM, so you’ll know it as a flexible wide-column database that scales out well on commodity hardware. MapR-DB uses the HBase API, so you can take advantage of the growing talent pool of HBase application developers. In fact, you can also take advantage of your existing HBase applications, as they can run on MapR-DB with only a configuration change.
MapR-DB is one of the newer NoSQL databases on the market, but you’d think it was an industry veteran if you look at its portfolio of successful production customer deployments. While other technologies eventually added enterprise-grade features like high availability, data protection, and disaster recovery, MapR-DB had them from the start. That’s because MapR-DB was always focused on helping customers meet their stringent deployment requirements. An independent research firm recently recognized MapR-DB as the top-ranked NoSQL key-value database for current offering, which means MapR-DB can do a ton of stuff well, and our customers know it.
Why is MapR-DB able to do so many things well? It again goes to the architectural innovation of the MapR Data Platform, which allows us to outrun general purpose file systems, and gives us tremendous room to innovate. By delivering an optimized platform designed for intensive workloads, both on a batch and real-time level, MapR gives you a technology that will continue to advance for a long time to come.
MapR-DB and Hadoop
So now that we have the database and NoSQL specifics out of the way, let’s take a look at the big picture. The big-picture advantage of MapR-DB is its tight integration with Hadoop in the MapR Distribution. Think of MapR-DB as the real-time, operational data manager of a multi-function Hadoop deployment. Any deployment that requires real-time updates of live data along with large-scale analytics on that live data can benefit from this integration. Example solutions include operational reporting, network security threat detection, fraud prevention, real-time personalization, predictive maintenance, logistics optimization, and so on.
And even if your needs don’t pertain to a real-time operational analytics platform, you can still benefit from the Hadoop/NoSQL integration. The consolidation of Hadoop and NoSQL workloads means that when you run analytics on your database data, you don’t have to copy it over to a separate Hadoop cluster. Anyone who’s had to move big data between clusters knows that this is not something you want to do on a continual basis. Having all data in the same cluster avoids the overhead of data movement, but you also benefit from the consolidation of various functions around high availability, disaster recovery, and data governance, which are seamlessly applied across both files and tables in the MapR Data Platform. What this ultimately means is easier manageability and lower risk for your big data deployment.
Let me wrap up by saying that we’ve done quite a bit with Hadoop and NoSQL, as have our customers, but there’s still a lot more to come. Please stay tuned, and if you want to get started with MapR-DB, download the MapR Sandbox for Hadoop, and check out our upcoming webinar with IDC on October 14.
And get the complete top 10 list here.