Print Save PDF

About 5 minutes

There are good reasons why developers are turning to NoSQL databases in amazing numbers. Companies work with much more data than ever. It is also coming from a wide variety of sources. The rigid structure of relational database management systems (RDBMS) prevents the flexibility, scalability, and performance that organizations need to work with this data.

There are still some enterprise features of an RDBMS that we still want for some of our applications. Secondary indexes, for example, are utilized to more efficiently access records. RDBMS are also known for strong data consistency where some NoSQL databases are not.

relational_databases_mongodb

Enter MongoDB

These are some of the reasons that developers are turning to MongoDB. You get the performance, flexibility, and scalability that drew you to NoSQL in the first place. Plus, you get enterprise-class features that you were getting from your RDBMS.

Most developers start out with open source versions of NoSQL databases to speed implementation and reduce costs. As your applications become more mission-critical or gain wider adoption, you should turn to an enterprise-class distribution of the software.

MongoDB users turn to MongoDB Enterprise Advanced for greater reliability and management. It also provides:

  • Ops Manager for monitoring, automation, and backup.
  • In-memory computing for unmatched speed.
  • MongoDB Connector for business intelligence integration.
  • MongoDB Compass for schema visualization.
  • WiredTiger for improved performance and data management.
  • 24x7x365 support.

As with any technology, there are some those who will push the system to its limits. Developers are always trying to analyze more data and add more functionality to our applications, all without sacrificing performance or reliability. So what do you do if need more performance? 

Enter IBM Power Systems Running Linux

It seems that many developers do not pay a lot of attention to the infrastructure that their applications run on. They really should. The pioneering computer scientist, Alan Kay, has been quoted as saying, “People who are serious about software should make their own hardware.” That is unrealistic for most developers, but you still can gain some advantages by running your applications on hardware that was designed for your software.

IBM has completely changed their philosophy from a hardware company to a software company. They have sold off their commodity hardware to Lenovo, but they maintain their Power Systems line of servers because they know how important the hardware is to their developers.

IBM is committed to the open source community. In 2013, they invested $1 billion into the Linux ecosystem. They also open sourced their RISC-based Power processors through the OpenPOWER Foundation. IBM partnered with Red Hat, SUSE, and Ubuntu. And all of your favorite development tools are available on IBM Power Systems running Linux.

You might be thinking, “That’s great, but why should I care?” 

Unmatched Performance and Reliability

IBM’s latest POWER8 processor is revolutionary. It can process 4x threads per core compared to a comparable X86 commodity server. At the same time, it provides 4x more cache with 1.7x the clock speed. For data-heavy applications, these resource performance improvements can allow you to add functionality such as real-time analytics without sacrificing performance.

Linux_on_Power_Blog_Chart.png

Source: Clabby Analytics | Infrastructure Matters: POWER8 vs. Xeon x86


The real exciting breakthrough is in IBM’s CAPI I/O. There is a separate GPU that handles all of the read/write commands so the processor does not have to.

This creates incredible performance gains. If you are taking advantage of MongoDB’s in-memory computing feature, then the benefits grow exponentially. You can turn 40TBs of flash into extended RAM. This means you can run work with datasets up to 56TBs in memory.

That is all great. But what does this mean? Of course your applications can over 4x as fast. It also means you can do more with less. Your organization can greatly reduce the amount of servers required to run your application.

To learn more, get your copy of The NoSQL Developers’ Guide to Infrastructure and Maximizing Data Performance

nosql_guide_cta

 

Written by Steve Erickson