Types of Inter-Operation Parallelism

Types of Inter-Operation Parallelism

Inter-operation parallelism is two types: 

1) pipelined Parallelism: It is an important source of economy of computation for database query processing. In pipelining output tuples of one operation A are consumed by a second operation B. Here the operation B starts even before the first operation has produced the entire set of tuples in its output. 

In pipelined parallelism, the output tuples of one operation, A, acts as input to a second operation, B, even before the first operation has produced the entire set of tuples in its output. 

Thus it is possible to run operations A and B simultaneously on different CPU's, so that the tuples produced by A is taken by B as input in parallel. 

Advantages of Pipelined Parallelism 

i) Pipelined operations are carried out in a sequence without writing any of the intermediate results to disk. 

ii) Pipelined parallelism is useful with a small number of processors. 

Disadvantages of Pipelined Parallelism 

i) It does not scaleup well. 

ii) Pipeline chain generally does not attain sufficient length to provide high degree of parallelism. 

iii) Pipelining cannot use with some relational operation which produce cannot output until all inputs have been accessed. For example, set difference operation. 

2) Independent Parallelism: Operations in a query which does not depend on one another can be executed in parallel. This form of parallelism is called independent parallelism. 
This type parallelism also does not provide high degree of parallelism. It is useful with less degree of parallelism. 

Advantage of Independent Parallelism 

Independent parallelism is useful with a lower degree of parallelism. 

Disadvantages of Independent Parallelism 

i) Like pipelined parallelism, independent parallelism does not provide a higher degree of parallelism. 
ii)It is less useful in a highly parallel system.