IBM, with more than 400,000 employees, knows as well as any large enterprise that supporting multiple business...
intelligence (BI) deployments can put a drag on IT resources and create data silos.
So Big Blue has come up with an alternative, today launching what it says is the world's largest internal cloud for BI and advanced analytics. This will make it easier for IT to manage and provision the hardware – servers and storage devices – that supports them and will give workers access to more data, IBM said.
IBM is also offering the internal analytic cloud service to its large enterprise customers under the moniker Smart Analytics Cloud.
The private cloud, which is anchored by IBM's System z10 mainframe and Storage DS8000, will bring together more than 100 of IBM's relational and transactional data sources, which between them house more than a petabyte of data. Sales and support staff can tap into and analyze the wealth of data with Cognos 8 front-end BI tools, according to the company.
The idea, explained Harriet Fryman, head of IBM's analytics and performance management products division, is to give front-line workers the most current and comprehensive customer and sales data available while also lightening the burden on IT and creating greater IT efficiencies.
"The intent is to provide over 200,000 IBMers with the ability to view information from around the world and use that to make smarter decisions at the point of impact with the customer," Fryman said. "The virtualization that's allowed in a private cloud allows us to be more efficient with the hardware technology, and even the operations and support of the user community becomes more streamlined."
Previously, IBM, like most large organizations, had numerous line-of-business (LOB) and departmental BI deployments throughout the organization. Each had its own hardware requirements and data sources. IT managed each separately, Fryman said, making it difficult to respond quickly to upticks in server demand. Users were also confined to the data sources they could access.
The internal cloud centralizes IBM's information base and gives users views beyond the scope of LOB and departmental deployments, she said. "It makes sure that when we want to have global initiatives, everybody's singing from the same song sheet."
The private cloud will allow IT to manage IBM's multiple data sources from one centralized location, lowering costs and the amount of work required from IT. Like public cloud offerings, the internal IBM cloud will allow IT to easily provision hardware for BI and analytic applications throughout the company as and when needed.
IBM is using its Infosphere data warehouse and integration products to combine and house the data, which users can now report on and analyze with Web-based BI tools. The private cloud project, known internally at IBM as Blue Insight, has been in the works since 2007, according to Fryman.
The new offering, similar to one released earlier this year by data warehousing vendor Greenplum called Enterprise Data Cloud, aims to give customers the elasticity benefits of public cloud computing -- but inside the firewall.
"This is a step in journey that will culminate with IBM and other vendors giving completely cloud based deployments as an option to customers if they want to go there," said James Kobelius, an analyst with Cambridge, Mass.-based Forrester Research. Cloud-based BI and data warehousing, both public and private, is still an 'immature' approach, however.
The cloud-based model for analytics won't truly be considered mature until other vendors, in particular Oracle and Microsoft, embrace it, and until 'killer apps' emerge that justify the approach, Kobelius said.
An example of a killer application would be deep content analytics in the cloud, Kobelius said, which would allow customers to quickly create, support then dismantle petabyte-scale data marts for analyzing structured data and unstructured content for marketing campaigns that last for only a limited time. Such short-lived, large data warehouse deployments in a traditional environment are difficult to justify from a cost perspective, he said.
IBM will also provide customers with consulting services to design and implement internal analytic clouds, Fryman said. IBM envisions the Smart Analytics Cloud appealing mainly to large enterprises that have numerous applications written in different languages and data sources "strewn throughout the world."
Pricing information was not immediately available.