Netezza Adds Geospatial Features
Netezza's new geospatial facility promises to let customers roll out location-aware applications at a "fraction of the cost" of general-purpose RDBMS
- By Stephen Swoyer
- September 17, 2008
IBM, Microsoft, and Oracle all have it. Now Netezza has it, too.
At its Netezza User Conference yesterday, Netezza trumpeted Netezza Spatial, a new in-database geospatial facility that it says makes it possible for companies to support location-aware applications at a "fraction of the cost" of general-purpose RDBMS systems -- such as those marketed by RDBMS giants IBM, Microsoft, and Oracle.
"Some of our competitors have spatial capabilities, but … those tend to be deployed as either operational systems or [in deployment are] completely separate from customers' data warehouse environments," comments Jon Shepherd, general manager of location-based solutions for Netezza.
Netezza's value-add, Shepherd reiterates, is -- as always -- price/performance.
"We saw an opportunity for Netezza Spatial to extend the data warehouse to include spatial capabilities to give them [i.e., location intelligence applications] the same performance advantage that we have with SQL -- [namely, an analytical capability] that's 20 times or 30 times faster, along with the simplified administration Netezza is also known for," he says. "There's a big difference between being location-aware, [where] you can pull a map up with your address on it, and doing spatial intelligence -- that is, taking all of your data to do analytics. It's a huge difference in scale."
Netezza Spatial is not a homegrown facility. Vice-president of product management Phil Francisco describes it as an "outgrowth" of the Netezza Developer Network (NDN) that his company announced last September.
In this case, Netezza contracted with partner company Intelligent Integration Systems Inc. (IISI), which developed Spatial. In this regard, Francisco says, Spatial is similar to about a half-dozen other applications -- including modules for Monte Carlo simulation and fuzzy name matching -- that exploit NDN APIs. Unlike those other applications, however, Netezza owns Spatial and plans to sell it as a for-purchase module. He believes the business case for a Netezza-branded Spatial capability is compelling thanks to the largely segregated (or siloed) way in which organizations today do geospatial analysis.
"The reason we did this with the spatial package is because it's an application set that cuts across many of the vertical markets that we sell into, so we viewed it as solving a significant problem in [customer] data centers -- that being the segregation of spatial analysis for their businesses from their standard business intelligence processing. That was the fundamental reason that we thought it was important to bring to market," Francisco told BI This Week.
Far from being a niche or arcane feature, geospatial analysis has the potential to enhance a number of business activities, says veteran data warehousing (DW) consultant Mike Schiff, a principal with BI and DW consultancy MAS Strategies. "You can do a lot with spatial data and geocoding. You can use it to figure out within X square miles what are the demographics, which can help you figure out where you want to put a new store, for example, or how you want to target a[n advertising] campaign] -- stuff like that," Schiff comments.
As Schiff sees it, however, Netezza's move -- considered in tandem with its showcasing of several different partner-developed applications at this week's Netezza User Conference -- begs another question. "What is Netezza's business? Are they now becoming an application engine? It sounds as if they're trying to break out of being stereotyped as an appliance or basically as an engine for analytics. The competition continually hits them as being first generation, or proprietary, so I think they're trying to show there's a lot more than you can do with [their NPS systems]," Schiff observes.
On the other hand, Schiff suggests, Netezza's NDN program has gestated very slowly: "Some of the applications they're touting [at this year's show] were in the press release when they first announced the partner program a year earlier, and you'd think those would have been in the works when they announced NDN last year. Why don't they have anything else new [to announce]?"
Netezza's Francisco, for his part, insists that NDN is "growing" and cites contributions from half a dozen different partners -- with more on the way, he says.
"These are applications that [partners] are selling to our customers. We are not involved in the sales process." In some cases, he continues, Netezza customers are developing NDN-based applications. "We've given them an API to develop to, we've given them an IDE, and they've brought forward their applications and algorithms," he indicates.
NDN provides API-level access to the vanilla Netezza warehouse, along with an IDE to help accelerate coding. Its goal, according to Francisco, is to "let [customers] do more inside the warehouse than just SQL." Netezza Spatial exploits NDN APIs and was built using the NDN IDE.
Could Netezza users expect other specialty facilities -- perhaps even a Netezza-native MapReduce capability? Shepherd says Netezza Spatial is a very special case.
"This [geospatial capability] is something that addresses a clear need in this marketplace. SQL Server now has spatial [capabilities], and certainly Oracle has had spatial capabilities for a few years now, but there's a difference between being able to be location-aware and being able to do really high-performance analytics on spatially-encoded data, and that's what we think is going to be game-changing about what we're bringing to the marketplace. It's an order of magnitude faster than what people are used to using, without all of the indexing and portioning that's required in other spatial packages," he says.
What about in-Netezza MapReduce, ala Greenplum Inc. and Aster Data Inc.? It could happen, Shepherd says. "We're looking at it. Within the NDN, we've had a discussion about just how easy it is to implement MapReduce, using the [APIs and IDE] we've provided," he points out, demurring on the question of a timeline.