{[ promptMessage ]}

Bookmark it

{[ promptMessage ]}


Here our mobvabon to design and engineer this

Info iconThis preview shows page 1. Sign up to view the full content.

View Full Document Right Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: like their MapReduce engine). IntegraBng Inherently ­Distributed Systems (1) 5   Consider now the case (with respect to the diagram to the right) in which:         A This o]en arises: it is the case in which there are pre ­exis2ng resources (both processes and data) that one may wish to integrate. Here, our moBvaBon to design and engineer this applicaBon as a distributed system is effec2veness, not efficiency. G O D ALGO ≠ A ≠L≠G ≠O (i.e., the code in each machine differs); DATA ≠ D ∪ A’ ∪ T ∪ A’’ (i.e., the inputs to each piece of code are not parBBons of a single data set). L A’ T A’’ ALGO DATA   In other words, we don’t have the choice of running the code in a single machine:     we may not own the resources (both processes and data) required, or they may be sca`ered within the enterprise. IntegraBng Inherently...
View Full Document

{[ snackBarMessage ]}