Information Technology Reference
In-Depth Information
The
Query Optimizer examines all algebraic expressions that are
equivalent to the given query and chooses the one that is esti-
mated to be the cheapest.
The
Code Interpreter transforms the access plan generated by the
optimizer into calls to the query processor.
The
Query Processor actually executes the query.
The
Runtime Optimizer modii es and adapts the query plans for
the Query Processor during the query evaluation.
It is obvious that under a streams environment, optimization tech-
niques are deployed before and after the execution of the query evalua-
tion. The situation is quite similar in terms of the data-intensive scientii c
workl ow. The optimization can be done both on the workl ow schema
level and on workl ow execution instance. In terms of specii c optimiza-
tion techniques, there are many categorization methods. Here we will
introduce two of them. Other methods can be referred to some other
references such as [39].
9.4.1.1
Static Optimization and Runtime Optimization
When a scientii c workl ow schema is given, we can evaluate and adjust
the execution plan, to make an optimized strategy to execute it. That is a
static optimization. During the execution of a workl ow, with the instance
data coming, the WFMS can know more and more extra information to
adapt its executing plan to gain a more optimized strategy. This is a
runtime optimization.
9.4.1.2
Semantic Optimization
Semantic optimization is a comparatively recent approach for the trans-
formation of given queries into equivalent alternative queries using
matching rules in order to select an optimal execution plan based on
the cost of executing alternatives [36,38]. To understand the semantic
optimization, we have to introduce two important concepts of semanti-
cally equivalent and semantic transformation.
Two plans are semantically equivalent if they return the same answer or
fuli ll the same task. A semantic transformation transforms a given
schema into a semantically equivalent one. Semantic optimization is the
process of determining the set of semantic transformations that results in
a semantically equivalent schema with a lower execution cost. The discus-
sion of optimization is indispensable in terms of any processing algorithm
and its corresponding execution plan.
XML stream-specii c optimization techniques have been well devel-
oped [36,38]. For example, the schema-based optimization (SQO) works
on one abstraction level, which uses schema knowledge to rewrite a query
 
Search WWH ::




Custom Search