#memeconnect #emc #bd
One of the most cited benefits of big data applications is flexibility. It can often take months or years to change a data warehouse application, while changes to big data applications may only of few hours or days. Moreover, big data apps typically require a complete re-load of the data, operate on entire data sets, pull distributed information not only from inside the firewall but increasingly from the cloud, while data warehouse apps normally see incremental growth of data from internal sources.
Fortunately this highly dynamic nature of big data apps presents big savings opportunities. Like the data, the IT infrastructure can also be dynamic by leveraging cloud computing and cloud storage. In essence the infrastructure can be rented from a bevy of smaller players or big ones such Amazon, Google, and Yahoo. Indeed, many offer HADOOP-based cloud services along with their own added value. In addition, these providers often offer data buyers in search of some particular data set a way to visualize the data before purchasing it, thereby ensuring buyers don’t waste money on data they don’t want. An early example is Google Books, which provides sample pages, but cloud service providers data visualization services are far more sophisticated.
Increasingly, ‘rented’ cloud infrastructure will be the norm for big data analysis, reducing the reliance on in-house IT, providing faster time to solution and significant cost savings.
Action Item: If in doubt, outsource infrastructure to the cloud. For big data apps, IT organizations should take an experimental mindset, avoid investments in in-house infrastructure where possible, and rent from the cloud.
Footnotes:
