Is cloud computing finally opening up the doors to analytics? Everyone wants advanced data analytics, but the bar to achieving these capabilities has been high -- requiring high-level skills, as well as highly scalable data management and storage.
Cloud offers a way into the world of analytics, especially for small to midsize organizations that may may never have been able to afford it. For large enterprises, it provides ways to quickly spin up or test new applications and data sets without crowding out existing infrastructure. A recent survey by O'Reilly Media and sponsored by Teradata finds that 40 percent of respondents who identify themselves as big data practitioners currently use cloud services for analytics.
Most BI and analytics vendors offer some variations of cloud-based versions of their products, either through SaaS, or web-based dashboards. Cloud services vendors such as Google, Amazon Web Services and Salesforce.com are also getting into the act, offering increasingly robust analytics. However, cloud analytics presents a mixed bag for enterprises. "Cloud-based big data analytics squarely address ongoing issues of scale, speed, and cost," writes Mike Barlow, the report's author. "On the other hand, they also create new issues around privacy, latency, and veracity."
Issues cited with moving or deploying analytics in the cloud include data and privacy requirements (71 percent); having everything needed on premises (41 percent): as the primary reason; and issues with connectivity and bandwidth (31 percent). The data security aspect presents a huge obstacle, as it security policies won't allow for data movement outside the corporate firewall.
These concerns are counterbalanced by a number advantages seen with cloud. The reasons for choosing cloud over on-premises analytics include flexibility to scale up or down (cited by 80 percent); faster deployments (62 percent); followed by reduced capital expenditures (39 percent); and the need to access to advanced technologies (39 percent).
Predictive analytics and reporting top the list as applications most likely to be used through the cloud -- the choices of 74 percent of managers. Other leading applications include text analytics (61 percent); machine learning (57 percent); and testing (46 percent).