Cliff Saran is the managing editor (technology) on Computer Weekly magazine responsible for commissioning, writing and overseeing the magazine strategy concerning all matters relating to technology from up-and-coming research and development to systems management challenges and legacy support and maintenance.
Cliff has been writing about these subjects since the early 1990s. In his current role, he writes a regular blog called Cliff Saran’s IT FUD blog which aims to unravel the hype, weed out the fear uncertainty and doubt spun by the massive marketing machinery in the IT industry.
You can contact Cliff by emailing [email protected].
How to find Cliff Saran online
Cliff's photography website and Flickr photostream.
[email protected] 020 8652 8460
Enterprise data warehouse suppliers will integrate open source big data platform Hadoop, according to Forrester.
One of the challenges of social CRM is analysing vast data sets. The industry uses the term "big data", to describe these data sets, but tools for big data are still immature. One of the rising stars is the open source Hadoop platform, which provides the analytics behind sites such as Facebook.
Forrester predicts Hadoop will become a key component in the product strategies of leading enterprise data warehouse suppliers. EMC Greenplum and IBM already offer Hadoop products and other suppliers offer degrees of Hadoop integration in their product families, or have announced intentions to do so.
Forrester expects that by Q3 2012, companies such as Teradata, Oracle, SAP/Sybase, Microsoft and HP/Vertica will acquire Hadoop start-ups such as Cloudera, MapR Technologies, DataStax, HStreaming and Outerthought.
But Forrester also said Hadoop is an unfamiliar technology to many enterprise data analytics and IT professionals. In a report on using Hadoop in the enterprise, senior Forrester analyst James Kobielus warned that even top-notch advanced analytics and enterprise data warehouse veterans will find it a tough slog to get their heads around Hadoop specifications, tools and approaches.
In the report, Enterprise Hadoop Best Practices: Concrete Guidelines From Early Adopters In Online Services, he explains that training and professional certification services for Hadoop are in short supply, making projects not only difficult but also challenging and even riskier. Further, industry-consensus best practices are conspicuous by their absence.
The report highlights key areas of concern such as the core specification for Hadoop, which is still being developed by the Apache Hadoop community. In the report, Kobielus warns that federation, metadata, high availability and machine learning are missing from the specification
This means businesses are stuck with proprietary functions that suppliers selling Hadoop put into the open source software.
There is also a lack of a single, integrated, Hadoop software distribution. Many Hadoop implementations involve some custom coding. Given the complexity of Hadoop deployments, Kobielus warned that the messy complexities of custom Hadoop development and data modelling may present a formidable learning curve that delays full deployment.