Ranking method for services composition

Document Sample
Ranking method for services composition Powered By Docstoc
					Ranking services for composition

      Hong Qing Yu (Harry)
     Service composition
    “Composition of Web services has received
    much interest to support business-to-business
    or enterprise application integration.” [1]
   Static

   Dynamic

Issues for composition
   Global services registration
   Service search/discovery
   Understanding composition requirements
   Service selection
   Workflow generation
   Service invoking
Ranking problem for selection
   If there are more than two services satisfying
    functional requirements,
   Which one is best to use?
    Cheapest one
    Fastest one
    Best performance
   Other non-functional properties.
   Logic Scoring preference is a technique can
    help us.
Logic scoring preference
    Traditional Scoring Techniques are simple
E=W1E1+W2E2+…+WnEn, 0 ≤ E ≤ 1.
    There is a problem [4]
It is regardless of the level of importance, the
contribution of component Ei to the global score
is limited to Wi
    LSP (Logic Scoring preference)
Logic scoring preference
   Differences are r & W

E=(W1Er1+W2Er2+…+WnErn)1/r, 0 ≤ E ≤ 1,
W1+W2+…+Wn=1, Wi>0, i=1,2,…,n.

   r is a real number selected to achieve the
    desired logical properties of the aggregation
Logic scoring preference

                     [4] [5]
Ranking by composition context
            : is an European project
   The meaning of context in the project

   Context affects service selection

   We need a simpler way to define r
 Designing evaluation rules
E=(W1Er1+W2Er2+…+WnErn)1/r, 0 ≤ E ≤ 1,
W1+W2+…+Wn=1, Wi>0, i=1,2,…,n.

1.   Filtering rules
2.   Evaluation function
3.   r selection
Filtering rules
Cost<$35      Speed>30/s Quality>85

           Irreplaceable preference criteria

           Replaceable preference criteria

If the service’s properties do not achieve
the irreplaceable preference, then it will
be filtered out.
    Evaluation function
    Exact match Es=1 (if the criteria is matched) or 0
     (if is not matched)
    Set overlap Es=(e1+e2+…+ei) /i (with Ei being a
     score for each criteria)
    Level match if i is the number of levels and ic is
     current service level value, then we define: Es=ic/i
Evaluation function
   Specific value if vx is the maximum value of all
    relevant services in one criteria, vn is the minimum
    value and vi is the current service value, then we
r selection
   E=W1E1+W2E2+...WnEn
   Can we compute the weight for choosing the
    r instead of using the way introduced in [5].
   On the one hand, Filter makes all aspects
    criteria is replaceable, which means that we
    need conjunction.
   On the other hand, if the weight of each
    criterion are so difference, we also need
r selection rules
   We are in a very balanced position, and we
    can narrow our r selection tables

   To simplify defining the r value, we just select
    1.5, 1, 0.5.
   If (highest weight – lowest weight)>average weight, then r=1.5
   If (highest weight – lowest weight)<average weight, then r=0.5
   If (highest weight – lowest weight)=average weight, then r=1
Worked Example
    Criterion requirement:
1.   More people’s weight=0.6
2.   Quality’s weight=0.3
3.   Cost’s weight=-0.1
    The result
Etalkfly=11.5·0.6 +(1/3)1.5·0.3+0=0.658
Ehotmail =11.5·0.6+11.5·0.3+(0.6)1.5·0.1=0.946
5.   “Continuous Preference Logic for System Evaluation”,
     Jozo J. Dujmovic, USA


mikesanye mikesanye