Innovator Chat: The Power of Predictive Analytics in the Cloud
Predictive analytics and cloud computing. Individually, these ground-breaking technological innovations are transforming the way businesses operate. Combined, their potential only multiplies -- especially in our era of big data.
How can businesses maximize the value of predictive analytics in the cloud? Brenda Dietrich, Vice President and Chief Technology Officer for Business Analytics at IBM, tackles our questions on the powerful convergence of these two innovative technologies.
Q: Why is unstructured data growing so rapidly? Will the growth continue?
A: Data is growth comes from multiple sources: From the use of computing within an enterprise for business functions such as commerce, customer service, and resource management; From the use of communication devices, including computers, tablets, and mobile phone by individuals for both work and personal use; And the instrumentation of physical infrastructure, such as electrical grids, highways, and buildings for more efficient monitoring and management. We expect this to continue as both enterprises and individuals find value in access to information and communication.
Q: How does moving data into an agile, virtualized infrastructure increase flexibility?
A: New forms of data generation, particularly in social media, and new forms of data usage by individuals are emerging. An agile, virtualized infrastructure can grow both data storage and data access capacity as usage patterns change. Computing resources can be redeployed as needed to better serve the users.
Q: Is it possible to integrate cloud-based analytics systems with on-premise systems?
A: Yes, IBM's Cast Iron Cloud Integration software enables seamless connection of cloud-based and on-premise applications. This enables organizations to connect their hybrid world of private clouds, public clouds and on-premise applications in days based on a simple "configuration, not coding" approach -- which can reduce integration costs by as much as 80%. Users get real-time visibility of data in the cloud and are able to maximize their productivity.
Q: How long does it generally take to deploy a business intelligence solution in the cloud? What about an advanced analytics solutions?
A: There is considerable variability, depending on the nature of the data being used and the client's use of the data. However, the ease of deploying analytics in the cloud is generally considered moderate, with the potential gains on the high side of the spectrum.
Q: How do public and private solutions differ with respect to deployment times?
A: Assuming the private cloud has already been established, deployment of a specific solution should be similar.
Q: Distributed servers used for analytics applications often leave substantial system power unused. How does IBM software tackle this challenge?
A: IBM has specifically developed bundled warehousing and analytics solutions to tackle this issue. The IBM Smart Analytics System 7700 and 7710 are complete business-ready analytics solutions based on IBM Power Systems. They are designed as workload-optimized solutions for high-performance analytics to help accelerate delivery of insights for faster, smarter action.
Q: What's the basic process for deploying a predictive model on IBM's SmartCloud?
A: Predictive analytics software, such as SPSS, is part of the IBM Software portfolio and can be deployed like any other IBM cloud-based application. In addition, some IBM partners, such as Zementis, have made their products available to run on SmartCloud.
Q: What tends to be the biggest inhibitor to adopting a cloud-based analytics solution?
A: Some large enterprises have invested in BI applications at the business unit level, in a separate and autonomous fashion, sometimes even using different products and platforms in different departments. This siloed and often duplicative approach creates the great potential for savings through moving to cloud -- but it also creates the implementation complexity that may require outside expertise to overcome.
Q: What are the key features of a great CIO dashboard?
A: A great CIO dashboard should enable the CIO to understand the current and likely future states of the computing infrastructure at a glance. It should of course enable the CIO to identify the "hot spots" that require immediate attention, but it should also alter her to trends and shifts in the use of HW, SW and data resources, an identify opportunities for more effective use of these computing resources.
Q: What did IBM learn about cloud-based analytics from the launch of its own private cloud?
A: IBM's internal analytics cloud put an end to siloed business intelligence and the six-figure funding required for new BI projects. Organizations across IBM can access a centralized analytics cloud for tools and intelligence aggregated from hundreds of information warehouses. Additionally, IBM has invested in a team of analytics experts that guides the organizations through the analytics process and facilitates learning and reuse across organizations. The associated savings are expected to reach tens of millions over five years.