More big data is not in the cloud

Amazon's chief technology officer, Vonna Vogle, has opened a topic about using cloud computing to complete big data, what do you want him to do?

This view is compelling, including the need for large data analysis, especially for real-time analysis. Companies want to have that ability, for Vogel, which means they need the public cloud--especially the Amazon's public cloud. Vogel also said that we all hope that infrastructure like Hadoop will be able to hide behind an analysis layer like the Amazon redshift.

For now, Vogel is half right, and after 5 years his views will be completely correct.

In fact, the big data is really big, most companies already have some big data projects, and they all see the same benefits as Vogel, such as real-time analytics, including predictive analytics, that will help CIOs bring great value to the enterprise in it.

The public cloud platform does attract attention, such as automatic and self-service instant scalability features, and built-in large data services like Hadoop. But in fact, most of the projects now called Big Data are just running on the enterprise's own data centers. Therefore, it seems that it will take some time to develop ahead.

The reason is simple and understandable: using a local storage system means that integration with the operations database does not cause problems with network connectivity issues. In many cases, it feels good to store huge amounts of data in a supplier's data center, but you'll want to be able to load it properly if it's dependent on the vendor.

In addition, security and compliance issues remain a major concern for the public cloud, and it is easier to store data locally. In addition, the performance of the local system may be better because there is no latency problem with network transmission. Finally, software and hardware are now very cheap, and the return on the system to the cloud is not necessarily better than you think.

(Responsible editor: Fumingli)