Geek asks and answers

What types / classes of algorithms can be redone in the MapReduce paradigm?

A few quick questions:

  • What types / classes of algorithms can be recalculated in the MapReduce paradigm? (for example, the k-tool has an MR implementation)

  • Are there any expressions that cannot be expressed this way?

  • What characteristics of the algorithm make them less attractive / difficult to repeat in the MR paradigm

Thanks in advance for your help.

Max.

+3
source share
2 answers

I am working on the same issues to collect large data algorithms that come from the MPI world. Here is my welcome.

-, -, /. , , , / , , . MR.

MR, MR , -. , , , , , , , k- - , , .

:

  • ( , ).

- , , , SSD- , SSD- PCIe. SAS SATA, , . SSD , . , , .

MapReduce /: + + , . + + , , , . , MR : , , . Nishant - , , , .

, MR PDE , . , FutureGrid: Domains of Algorithmic Organization

+2

Map , " ", . , .

, , , MapReduce .

+2

More articles:


All Articles