Big Data // Software Platforms
News
3/25/2014
03:16 PM
Connect Directly
Google+
LinkedIn
Twitter
RSS
E-Mail
100%
0%

SAP Hana Spawns Virtual Data Warehouse

SAP's new Hana-powered In-Memory Data Fabric queries data without copying and stokes federated-access competition with IBM and Teradata.

Gartner has been talking up the "logical data warehouse." Forrester describes it as an "information fabric." SAP announced Tuesday that it's delivering federated access to information for analytics with its Hana-powered In-Memory Data Fabric, the big new feature of SAP Business Warehouse (BW) version 7.4.

The In-Memory Data Fabric means that you don't need to use time-consuming batch ETL processes to copy data into SAP BW. Queries run virtually against all data connected via the IMDF. There are several advantages to accessing data where it lies, according to SAP.

"First, the approach decreases the amount of storage required because you only store the data once," said Neil McGovern, SAP's senior director, product & innovation marketing, in an interview with InformationWeek. "The other advantage is that you access the latest version of the data, not a copy from the night before or last weekend." Infrastructure cost savings and timely insight are two hot buttons where data warehouse deployments are concerned.

[Want more on in-memory advantages? Read In-Memory Databases: Do You Need The Speed?]

The IMDF was developed internally and harnesses stream-processing and data-replication technologies SAP picked up through its 2010 acquisition of Sybase. SAP competitors, including IBM (with DB2 Information Integrator) and Teradata (with Unified Data Architecture), have already introduced their own federated-data-access features. Nonetheless, SAP says Hana gives it performance advantages in supporting federated querying.

"The technology has been in the market for some time, but delivering query response times within service-level agreements has always been challenging," said Ken Tsai, SAP's VP and head of SAP Hana and data management marketing. "Using Hana's smart data-access technology, we've enabled query plans to execute across multiple sources in a timely way."

There will be cases where Hana's in-memory advantages can't overcome the bandwidth constraints of conventional systems that organizations might want to access through IMDF. What's more, companies may be concerned about performance hits on mission-critical systems should they be exposed directly to virtual-data warehouse queries. In these situations, IMDF can tap an operational data store or materialized views to query recent (though not real-time) data. Operational data stores and materialized views are old data warehousing tools -- and not exactly consistent with the Hana ethos of not copying data. But sometimes you just can't get around the realities of data architectures.

SAP: Guinness Book Of World Records validates that SAP Hana-powered virtual data warehouse scaled to 12 petabytes and sustained sub-second query speeds.

SAP: Guinness Book Of World Records validates that SAP Hana-powered virtual data warehouse scaled to 12 petabytes and sustained sub-second query speeds.

Other upgrades introduced in SAP BW 7.4 aim to simplify the development of new queries and analytic applications. According to SAP customer Molson Coors, new data-modeling aids, including a CompositeProvider and an Open ODS View, have cut application development time in half.

"The SAP BW design layer has been simplified," said Pawel Mierski, Molson Coors' senior BI development lead, in an interview with InformationWeek. "Now instead of having to create five or six different objects, there's one object in which you can encapsulate everything you want to model in BW."

Fast, easy modeling is getting to be table stakes in data warehousing. The scramble for federated data-access supremacy looks like a next-stage battle to become the de facto (virtualized) enterprise data warehouse. SAP cited a Guinness Book of World Records-backed claim that it has created a 12-petabyte virtualized data warehouse using Hana and IMDF that was able to sustain query speeds nearly on par with a single-node, non-virtualized database.

We're guessing data-management types would be more interested in authoritative TCP performance benchmarks. But for now it's a talking point on a new data warehousing competitive front. We hear Teradata may up the ante on data warehouse virtualization as soon as next week.

Interop Las Vegas, March 31 to April 4, brings together thousands of technology professionals to discover the most current and cutting–edge technology innovations and strategies to drive their organizations' success, including BYOD security, the latest cloud and virtualization technologies, SDN, the Internet of things, and more. Find out more about Interop and register now.

Doug Henschen is Executive Editor of InformationWeek, where he covers the intersection of enterprise applications with information management, business intelligence, big data and analytics. He previously served as editor in chief of Intelligent Enterprise, editor in chief of ... View Full Bio

Comment  | 
Print  | 
More Insights
Comments
Newest First  |  Oldest First  |  Threaded View
Lorna Garey
50%
50%
Lorna Garey,
User Rank: Author
3/26/2014 | 10:47:24 AM
Re: Avoiding pain and suffering
Doug, is this concept of a "logical data warehouse" or "information fabric" relatable to object-oriented storage? The two seem to have similar characteristics, and at some point, it seems like the "save everything" trend in enterprises will run smack into the desire to make every scrap of data available for analysis, so that the virtual data warehouse and the archive or tiered storage system become interdependent.
D. Henschen
50%
50%
D. Henschen,
User Rank: Author
3/25/2014 | 4:08:12 PM
Avoiding pain and suffering
The logical data warehouse/information fabric concept is compelling because you avoid the ongoing pain of batch uploading and transforming data. Of course you still have a rigid database schema and you have to set up SAP's In-Memory Data Fabric to transform data in flight. And, as mentioned, you might have challenges tapping any old source if querying might create a performance hit. Long story short, we're still in the world of data warehousing here.

In my book anything in the neighborhood of petabyte league would be a candidate for a big data platform. I'm guessing a better play for virtual data warehouse would be a more focused use case involving smaller quantities of structured data -- and where near-real-time access to the latest data really matters.
In A Fever For Big Data
In A Fever For Big Data
Healthcare orgs are relentlessly accumulating data, and a growing array of tools are becoming available to manage it.
Register for InformationWeek Newsletters
White Papers
Current Issue
InformationWeek Must Reads Oct. 21, 2014
InformationWeek's new Must Reads is a compendium of our best recent coverage of digital strategy. Learn why you should learn to embrace DevOps, how to avoid roadblocks for digital projects, what the five steps to API management are, and more.
Video
Slideshows
Twitter Feed
InformationWeek Radio
Archived InformationWeek Radio
A roundup of the top stories and trends on InformationWeek.com
Sponsored Live Streaming Video
Everything You've Been Told About Mobility Is Wrong
Attend this video symposium with Sean Wisdom, Global Director of Mobility Solutions, and learn about how you can harness powerful new products to mobilize your business potential.