Print Save PDF

About 4 minutes

If you attended MongoDB World 2016, congratulations. We hope that you picked up a ton of useful pieces of information that you were able to take back to your everyday work lives. We also know that there is way too much content at these events to consume and absorb it all.


A Session You Most Likely Missed

Monday, June 27th was a pre-conference event. It included several technical workshops. This was also when the MongoDB Partner Summit was held. Unless you are with a business partner that sells or supports MongoDB, it is unlikely that you attended this event.

There was one session in particular that I found very interesting that I was lucky enough to see. It was a panel discussion that included key executives with IBM, Avnet (one of IBM’s distributor partners), and MongoDB. The exciting part was delivered by a large MongoDB client. The panel was hosted by Terri Virnig, VP and BLE of Power Ecosystem and Strategy at IBM.

Why Is Retail Embracing MongoDB?

A technology professional from a major retail chain that operates over 300 stores in almost 30 states took the stage with IBM and MongoDB. This organization competes with JCPenney, Sears, and Nordstrom. They also have a successful eCommerce business.

 

The organization was in the architectural design or proof of technology phase for a project. They were re-writing an application that they call Product Attribution. The app allows their buyers to enter and maintain product information about all of the items that they sell at the stores.

 

Thousands of Products that Are Very Different from Each Other

Each item that the company sells has very unique characteristics or attributes, and it becomes difficult to handle that information in an application. For example, they could have a men’s dress shirt that has a neck size, sleeve length, and a specific cut—plus other identifying traits. In addition, they needed to track women’s high-heeled shoes and their corresponding heel heights, colors, sizes, and so on.

 

The sheer multitude of unique items posed a major challenge with their relational database.

Flexible Data Schema

What the retail chain discovered with MongoDB is that the flexible schema or schema-less nature of the data engine was a perfect fit for this type of data. The company was able to have a document that uniquely represents each item. And with the flexible schema they could have as few or as many attributes as they needed for a given item.          

 

The other item that was important when selecting MongoDB was that it was JSON at its core. With JSON rapidly growing, it just made sense from a development and productivity perspective as opposed to using XML or some other application. It doesn’t hurt that they were already using MongoDB for 12 other applications in their infrastructure.

 

Why MongoDB on IBM Power?

The retail organization has a strategy of using virtualized environments, as they wanted to avoid the dreaded server-sprawl. They prefer to vertically scale instead of horizontally scale. They heard that the little endian Linux distributions are now available on IBM Power Systems and PowerKVM. So they began searching for test cases of existing or new applications running on Power.

 

They also had past experience with POWER8 and had confidence in its ability to scale and be reliable. When they learned that MongoDB was going to be optimized for the POWER8 environment, it just made sense. It would allow them to vertically scale and continue to grow their MongoDB usage without having to scale-out.

 

Wired-Tiger storage engine was also a key piece, with its feature of locking at the document level. Any issues with concurrent users in the past has been alleviated. The performance results of their POC has helped them make the decision to move forward with MongoDB on Power.

 

Want to See More?

The panel at MongoDB World 2016 went on to discuss some major use cases for MongoDB on IBM Power Systems that are seeing incredible performance gains. The team of experts also outlined the benefits of RapidBuild, which is a pre-configured IBM Power System running Linux and MongoDB. This makes procurement easier, quickens deployment quicker, and provides support from a single source.

 

To make sure you receive the latest announcements and updates from IBM, IBMs partners like MongoDB, and the OpenPOWER community, do not forget to subscribe to the IBM BP Network blog.

 

Written by Steve Erickson