Northrop Grumman, CDFS II

Read full Case Study in PDF
Northrop Grumman, CDFS II

Versant’s high performance object ingestion and retrieval enabled the developers to successfully implement the large-scale simulations needed for CDFS II.


Developers at Northrop Grumman were challenged to develop a next generation simulation application for the U.S. Air Force that could deliver results 6 times faster with 4 times higher resolution.

Operations, intelligence, surveillance and reconnaissance platforms and sensors are impacted by weather conditions, therefore accurate spatial and temporal resolution is necessary for military operations. The Cloud Depiction and Forecast System – Version II (CDFS II) provides real-time global cloud analysis and forecasting. Operated by the Air Force Weather Agency (AFWA), CDFS II uses highly advanced algorithms and operational infrastructure to collect and analyze the sensor data from three classes of military and civilian meteorological satellites.

A total of nine satellites feed data to CDFS II. In addition, CDFS II uses a feed of weather forecasts from the Air Force Global Theater Weather Analysis and Prediction System. The cloud forecast is a simulation that requires a very complex database model. The new system must produce forecasts 12 hours into the future for the entire globe each hour and a sixty-hour forecast every six hours. CDFS II must perform roughly 6 times faster than the old system, and at the same time must provide a 4 times higher resolution.

The developers decided that the most effective way to program the complex CDFS algorithms was to use object-oriented programming techniques and Versant was their number one choice for the database. Object-oriented techniques resulted in faster development time, lower development cost, and greater ability to evolve the code to accommodate future requirements. Data enclosed in classes developed in C++ are saved and retrieved with high performance.


With the Versant Object Database, there is no additional code required for mapping the complex C++ objects to a relational database, to store and retrieve data in a special purpose database, or to write the data to the file system. The in-memory object model represents the Versant database model one-to-one without any intermediate data definition language or runtime translation layer.

As a result, CDFS II needed significantly less queries and correspondingly, the number of database indexes required were substantially reduced. By implementing the Versant Object Database, CDFS II was able to eliminate the over-abundance of index structures and the mapping layer which resulted in much higher data ingestion rates, allowing CDFS II to perform data ingestion and analysis an order of magnitude faster compared to the old CDFS system.


The previous system could only create a 48-hour forecast every six hours and a nine-hour forecast on request. With the implementation of the Versant Object Database, CDFS II can now create a 60 hour forecast every 6 hours and auto generates a 12 hour forecast every hour. The old system had a resolution of 29 miles, while CDFS II has a resolution of 14.5 miles, providing twice the precision and managing four times the data of the previous system. CDFS II is the only operational cloud forecast system in the world to date. In addition to consumption by the Air Force, the cloud forecasts are used by other military and civilian agencies such as the Navy Fleet Numerical Meteorology and Oceanography center and the National Oceanic and Atmospheric Administration (NOAA).

About Northrop Grumman

Northrop Grumman Corporation (NYSE: NOC) is a leading global security company whose 120,000 employees provide innovative systems, products, and solutions in aerospace, electronics, information systems, shipbuilding and technical services to government and commercial customers worldwide.