Cloud Analytics – Tapping the Value of Big Data - Page 2
The Next Step in the Evolution: Cloud Analytics
Of course, these first-generation cloud-storage implementations are focused solely on storing and managing raw data – without any concern for its structure, format, periodicity, or other characteristics. But, as Big Data becomes a greater concern for the modern corporate data center, there are other associated challenges. Perhaps the most compelling challenge lies in the desire to do more than simply keep the data. Instead, companies want to derive value from that mountain. They want more than “bulk data” – they want Big Data, and they want it to add value.
Of course, for years, business intelligence and online analytical processing disciplines have sifted through and mined corporate data to help companies detect hidden patterns, uncover extremely important insights, and gain competitive advantages. While few could question the value of business intelligence (BI) and related analytics, today’s data volumes are making BI a much more challenging proposition. Although statistical algorithms and query tools have improved and server processing power has continued to expand, the volume of data has become too great for those infrastructures. Even extremely sophisticated data centers are vulnerable to the “light-dimming query” – the processing-intensive analytical request that can take days to complete.
And cloud storage alone is not the answer. In fact, simple cloud-storage strategies could actually undercut BI efforts that rely on Big Data. A BI application in the corporate data center would need to access all of that Big Data stored elsewhere in the cloud, creating unacceptable bandwidth-hogging requests for terabytes of data that would have to cross an internetwork to be analyzed. In this model, the processing delay wouldn’t be CPU constraints – it would be caused by the delay in simply moving data from the cloud to the BI app in the data center.
But what if the cloud-storage facility also had the BI analytics applications and infrastructure as well? Like Willie Sutton’s famous answer about why he robbed banks (“Because that’s where the money is”), analysts and business users could run their BI analyses in the cloud – because that’s where the data is. In fact, in many ways, analytics might be the ideal cloud application – the convergence of cloud-storage and cloud-compute resources.
Such a paradigm places greater pressure on cloud vendors. Previously, they could simply provide bulk storage services - irrespective of the data and without regard to its usage. But with the emergence of cloud analytics, cloud-storage vendors now shift from “infrastructure-as-a-service” to a broader paradigm: providing analytic applications and managing the Big Data. Essentially, vendors must now manage very large data warehouses for their clients.
If cloud vendors can successfully make this transition, they will dramatically elevate the cloud-storage value proposition from mere storage to meaningful insights and competitive advantage. Companies may be wondering if it’s worth it to capture and store Big Data (for reasons other than regulatory compliance). By acting as the “tipping point” for moving data to the cloud, the emerging discipline of cloud analytics can go a long way toward transforming an expense line-item into an area where companies extract defined, monetizable value from their data, creating an important ROI-based avenue for the enterprise.
