As Published In March/April 2012

FEATURE

A Better Breed

Monsanto develops better seeds for better crops with Oracle Exadata and the Sun ZFS Storage Appliance.

High-capacity and high-performance engineered information systems enable organizations to handle more data and process it faster than ever before. What engineered systems contribute to a company and its customers varies by industry—from spotting and taking advantage of sales trends in retail and bringing competitive products to market more quickly in financial services to cutting batch-processing time in chemicals and increasing on-time delivery in consumer goods. For one agricultural company, engineered systems are helping to feed the planet.

Monsanto Company is a leading global provider of agricultural products that improve farm productivity and food quality. The company’s charter is to improve farmers’ lives by helping them produce more from their land while conserving natural resources. Through innovative breeding and biotechnology techniques, Monsanto produces seeds that—among other things—help farmers to reduce agricultural inputs such as pesticides and improve crop yields.

“September to November is a critical time frame for us,” says Wendy Poulsen, R&D IT strategy and operations lead at Monsanto. “We gather a tremendous amount of information from the field that must be analyzed and turned back around to breeders very quickly so they can determine which seeds to ship to winter nurseries. The speed by which that analysis can be done is extremely critical to us, but it has become increasingly challenging over time, as more information is available from which to make decisions.”

In addition to greater data processing speed, Monsanto’s R&D team needed more data storage capacity.

“When we plan for the harvest season, we look at what we expect to see and then we plan for two or three times that load on our information systems,” explains Todd Eyrick, business alignment and strategy lead in Monsanto’s Global Infrastructure organization. “There are multiple factors, such as weather, over which we have no control that can cause our database management systems to be under much greater load than we would have otherwise planned for.”

To satisfy its need for greater speed, capacity, and throughput, at the end of 2010 Monsanto purchased three full-rack Oracle Exadata Database Machines. These engineered systems are now divided into a production environment, a preproduction environment, and a development/testing environment.

In July 2011, Monsanto consolidated four major R&D databases and 75 percent of its data to the Oracle Exadata platform. During the next phase of the project, wrapping up in early 2012, Monsanto will migrate 18 smaller databases that represent the remaining 25 percent of the data.

Monsanto used Oracle Consulting to supplement its internal resources during the implementation, and the company is currently engaged with Oracle Consulting Managed Services to support the Oracle Exadata environment. “We faced a very tight timeline to get this new database environment up in time for harvest,” says Poulsen. “Oracle partnered with us and brought in the right people when issues arose, which reduced the implementation risks throughout the project. The responsiveness Oracle demonstrated showed they cared about our business and was one of the key reasons for our success.”

The Oracle engineered systems have now been running in production during one complete harvest cycle. “We definitely have experienced a very smooth, highly available system during the past several months,” says Eyrick.

Faster Growth Around the World

One of Monsanto’s goals when it adopted Oracle Exadata was to minimize the turnaround time for making the raw field data available to researchers anywhere in the world. According to Poulsen, some of Monsanto’s researchers used to wait as long as 36 hours for the results from field experiments. Now data that arrives by 6 p.m. at Monsanto’s U.S. data center is available for researchers in Europe at the start of their next workday. “We’ve cut the turnaround time in half,” she notes.

Postharvest data analysis helps researchers determine which products to advance. Thanks to the increased speed of SQL queries on the Oracle Exadata platform, researchers can now run multiple queries with different data sets, which allows them to conduct more-thorough analyses and make better decisions. Tests on Monsanto’s Oracle Exadata platform revealed significant performance gains for 90 to 95 percent of SQL queries.

In working with Monsanto to understand its business process and find opportunities to speed the breeding harvest cycle, Oracle recommended that Monsanto use Oracle Coherence to speed up data-processing performance by caching frequently used data. Oracle Coherence is an in-memory distributed data grid solution for clustered applications and application servers—and an important addition to Monsanto’s Oracle WebLogic Server environment.

“Oracle took the time to understand our business process so they could recommend how we can streamline the database activities,” says Eyrick. “They continually asked, 'What else can we bring to the table that will help Monsanto speed up this harvest process?’”

One key benefit of using the Oracle Exadata system is its inherent redundancy, which permits administrators to manage, patch, and upgrade Oracle Exadata components without incurring an outage. “As our business becomes more global, we can’t afford even brief outages,” says Poulsen. “The built-in redundancy within Oracle Exadata has insulated users, so if there is an issue, it has no impact on the harvest. Routine system administration is a lot easier as well, because the Oracle Exadata system integrates the storage platform with the database environment.”

Storing the Grain

While improving its information capacity and performance with Oracle Exadata, Monsanto has also been improving its disaster recovery and backup capabilities by implementing Oracle’s Sun ZFS Storage Appliance. Because of the large amount of information stored, the company realized that the overall volume of data would make it difficult to perform complete backup and recovery operations in a reasonable time frame. Eyrick and his team deployed the Sun ZFS Storage Appliance to create a more reliable hot-backup environment that permits system administrators to restore data from disk very rapidly.

Next Season

When asked for words of wisdom to share with other companies considering Oracle engineered system implementations, Eyrick doesn’t mince words.

“Having an integrated, single-vendor solution was very appealing to us,” says Eyrick. “The Oracle Exadata systems contain Oracle hardware, Oracle software, and Oracle database technology, all integrated into a cohesive system. Working with a single vendor has made it a lot easier to resolve issues. There’s no question about responsibility or ownership if something goes wrong.”

Beyond the technology and support, getting the right people involved in deploying Oracle engineered systems is also key.

“Having Oracle implement its engineered systems and support them through the first harvest has worked out exceedingly well,” Eyrick concludes. “Whether or not you plan to ultimately support this type of environment internally, it’s a great idea to engage Oracle to get started. Their consultants have a direct line to all the right technical resources, and they can turn things around very quickly if you run into problems.”