Distributed Systems & SQL

MapReduce: The Ultimate Programming Paradigm for Parallel Processing MapReduce was one of the major initiatives taken in the field of parallel processing or distributed computing in 2004 by researchers of none other than the tech giant google which is dealing with a tremendous amount of data at that time. Since then, it is the survival of MapReduce theory for such a long time showcases the hard work and brainstorming of the inventors. Despite new frameworks getting developed, MapReduce has sustained itself throughout the course of time. In fact, many of the parallel computing engines followed the path carved by MapReduce theory. After 18 years down the line, enthusiasts generally get confused about where to start their journey in the era where so many low-code & no-code frameworks are evolving. But IMHO, any beginner enthusiasts should look to first get their hands dirty on the MapReduce concepts and then move to the trending distributed computation architectures. Once you get the ...

Yes, all the folks who are working with big data must have heard this statement, but very few of them are aware of the real reason why parquet has the upper hand when it comes to storing nested data. It is the underlying architecture and hierarchical data model of parqeut which is making things easy for the nested data. This architecture of parqeut follows the Dremel paper published by Google. Here is the blog post which justifies the famous statement: https://blog.twitter.com/engineering/en_us/a/2013/dremel-made-simple-with-parquet