Featured Applications

Streaming analytics™ with Netezza provides a distinct advantage in a broad range of applications: for spatial analysis, text mining, risk-profiling, real-time pricing, network monitoring, fraud detection and many others. But in addition, the ability to run a complex analysis against a huge live database, without the delays and costs of moving data to separate hardware, opens up new types of analyses previously out of reach. Here are just a few examples of how the NPS appliance is being used for streaming analytics:

“Fingerprinting” with Hashing Algorithms: The Message-Digest algorithm 5 (MD5) is a standard cryptographic hash function with a 128-bit hash value. It is commonly used to store passwords and ensure that files transferred are intact. It is also used in chain of custody document fingerprinting. By performing the hash directly “on stream,” the NPS system runs hash algorithms on millions or even billions of records in seconds. This is typically hundreds of times faster than today’s method, which requires moving data from the warehouse to a supercomputing grid, performing the hash on the data and then loading the results back to the warehouse.

“Fuzzy Text” Search Analytics: Fuzzy text search analysis uses algorithms that provide a “best guess” of most likely results. One example is the Levenshtein edit distance algorithm, which calculates how many text edits would be required to manipulate, for example, “Madison Avenue” into “Main Street.” This type of algorithm is used by national security applications for complex analysis of names in port of entry data, as well as other text searching scenarios which require analysis of billions of text records. These types of capabilities open the Netezza appliance to analysis that was not only performance-constrained in the past but simply impossible through a SQL interface.

Predictive Model Scoring: Many companies use predictive modeling to finely segment their customers and make real-time decisions about promotions, pricing, fraud and other applications. This typically involves a time-consuming process: after transaction data is loaded into the data warehouse, the company performs large extracts to a separate cluster server system. The data must be denormalized and then fed back into a predictive modeling application for the actual scoring. Eventually, a score for each customer is loaded back into the data warehouse. The total round-trip can take hours, dominated by the latency of large data transfers. This entire process can now be done within the Netezza appliance, in a fraction of the previous time, enabling real-time offers and promotions to the right customers.

Geospatial Analytics: Geospatial analysis performs operations such as combining multiple maps or map layers according to predefined rules, or identifying regions within a specified distance of one or more features, such as roads or rivers. Geospatial analysis is used for solving problems like: “Find all properties within 10 miles of Hurricane Katrina’s eye path” or “Find all properties in Massachusetts that physically straddle county boundaries.” A Netezza partner is building an entire geospatial library on the NPS system, with analytics embedded within SQL functions. Users will be able to run geospatial applications on huge, comprehensive data sets, taking advantage of Netezza performance to make rapid and informed decisions.