Big data, the sixth thing CIOs should know

In a series of articles on large data proposals, IDG Group PC reporter Joab Jackson wrote "Big data, CIO should know five things" I think is definitely the best one. His five-point advice is almost always mentioned in our conversations with our clients:

1, large data entry cost is relatively low. We now see the cost of getting started with the big data for http://www.aliyun.com/zixun/aggregation/14294.html >, at least according to Capex's point of view, which is quite low. such as Hadoop, Cassandra, MongoDB, MapReduce, and other open source tools, coupled with relatively inexpensive cloud computing costs, allow organizations to collect, store, and analyze their data collections without having to spend huge costs.

2. Useful data can come from anywhere. Once "discarded" data is now a way to classify large data, Gazzang CEO Larry Warnock compares large data to a huge fishing net that scatters data to the ocean. Customers often treat the customer's transaction history, geographic location and some personal identification information such as medical records and bank account numbers as large data. How to use these seemingly isolated pieces of data to improve your business, or to push a project, is what big data is about.

3, the need for new large data professionals. Will big data become the next growth industry? We hope so, and we are sure of it.

4, large data do not need to accumulate prior to the organization. We liken the big data to a "garbage dump". We have compared large data to useless data discarded on the ground, fishing nets and garbage dumps that salvage data fragments. If the big data is still a baby, then he is also a child in the treatment.

But one thing is right anyway. Big Data allows us to ingest whatever we want, and we worry about how this huge amount of data will be used in the future.

5, large data is not just hadoop. There are a number of popular open source tools that can help you analyze massive amounts of data today. Joab mentioned tools include Splunk, HPCC systems and marklogic. There are also users using MongoDB, Infochimps Ironfan, and chef for cloud infrastructure automation and so on. Soon, Gazzang will also bring a new large data monitoring and diagnostic tool--zops to the market.

Finally, I would like to add a 6th to the five points suggested by Joab.

6. Security must be considered before starting. We have often heard that in a large data environment, the enterprise will allow the data to be in an unprotected state, such as user name and password, credit card data, or medical data are at risk of exposure. Fortunately, so far the risk has not been hurt to anyone (as far as we know), but it may be a matter of time.

How to integrate security into existing large data clusters will be a challenge. Large data clusters may contain thousands of nodes. It takes time to understand how data is collected, what data is worth protecting, and so on.

Data encryption and key management can be the last line of defense against unauthorized access or attack. The cost is relatively inexpensive and does not significantly affect the performance or availability of large data. So our advice to our customers is that if you think you have some sensitive data in your corporate environment, you must first secure it. (Wave Compilation)

(the writer has worked for many years in famous enterprises such as Dell, AMD and BMC, and now Gazzang is responsible for product marketing, focusing on large data security.) ）

(Responsible editor: The good of the Legacy)