By Judith Hurwitz, Alan Nugent, Fern Halper, Marcia Kaufman
The word polyglot is lent and redefined for giant data as some applications which use several core database technologies, which is probably the most likely results of your implementation planning. The state meaning of polyglot is “someone who speaks or writes several languages.” It will be a challenge to select one persistence style regardless of how narrow your method of big data may be.
A polyglot persistence database can be used when it’s essential to solve an intricate problem by breaking this problem into segments and applying different database models. This will make it essential to aggregate the outcomes right into a hybrid data storage and analysis solution. Numerous factors affect this decision:
-
You’re already using polyglot persistence inside your existing workplace. In case your enterprise or organization is big, you’re most likely using multiple RDBMSs, data warehouses, data marts, flat files, cms servers, and so forth.
This hybrid atmosphere is typical, and you must know it to be able to make a good decisions about integration, analytics, timeliness of information, data visibility, and so forth. You must know all that because you have to work out how it will squeeze into your big data implementation.
-
Perfect of environments, in which you only have one persistence technology, is most likely not suitable for big data problem-solving. At the minimum, you will have to introduce another type of database along with other supporting technologies for the new implementation.
-
With respect to the variety and velocity of the big data gathering, you may want to consider different databases to aid one implementation. Opt for your needs for transactional integrity. Must you support Acidity compliance or will BASE compliance be adequate?
Suppose you need to identify all of the customers for the product who’ve purchased within the last 12 several weeks and also have commented on social websites regarding their experience — AND when they have been had any support cases, where they acquired the merchandise, the way it was delivered, the things they compensated, the way they compensated, when they have been visited the organization website, the number of occasions, the things they did, and so forth.
Then suppose that you would like to provide them a marketing discount for their smartphone when they’re entering your (or your partners’) stores.
This can be a big data challenge at its best. Multiple causes of data with completely different structures have to be collected and examined to be able to obtain the solutions to those questions. You will want see whether the shoppers entitled to the promotion and, instantly, push them a coupon providing them something interesting and new.
This kind of problem can’t be solved easily or cost-effectively with one sort of database technology. However some from the fundamental details are transactional and most likely within an RDBMS, another details are nonrelational and can require a minimum of two kinds of persistence engines (spatial and graph). You have polyglot persistence.
Resourse: http://dummies.com/programming/big-data/engineering/big-data-and-polyglot-persistence/