Information Technology Reference
In-Depth Information
1. Only message places can be merged, and thus we should pay
special attention when we define P
C
.
2. When we merge two message places from two services, we
require that not only their color sets be identical, but also their
polarity be contrary.
Given these requirements that are specific to service composition,
we define the composition of two SWF-nets.
Definition 4.3. (Composition of SWF-nets)
Given two SWF-nets
N
i
¼
ð
P
Ii
[
P
Mi
;
T
i
;
F
i
; S
i
;
C
i
Þ
, i
¼
1 and 2, N
¼
ð
P
I
[
P
M
;
T
;
F
; S;
C
Þ
is the composition of N
1
and N
2
if
1. P
I1
\
P
I2
¼
1
;
P
I
¼
P
I1
[
P
I2
;
2. P
M
¼
P
M1
[
P
M2
;
P
C
¼
P
M1
\
P
M2
6¼
1
;
3.
8
p
2
P
C
, p in N
1
and p in N
2
have contrary polarity, that is,
C
1
p
ðÞC
2
p
ðÞ¼
1.
C
1
and
C
2
are the polarity functions of N
1
and N
2
, respectively, and
4. N
¼
:
N
1
P
C
N
2
The composition of two SWF-nets N
1
and N
2
via common
message places P
C
is denoted N
N
1
P
C
N
2
. We claim that operator
P
C
can be extended to other colored Petri nets as long as their places
are classified into two disjoint categories, that is, internal places (P
I
)
and message places (P
M
). Later we use this operator when we define
composition via mediation.
Figure 4.5 presents the method to transform BPEL composition to
SWF-net composition. In Figure 4.5, the composition of
¼
<
invoke
>
and
<
receive
>
/
<
reply
>
is modeled with SWF-net composition.
4.2.3 Mediator and Mediation-Aided Service
Composition
When one service provides functionality that another service
requires but two services were not programmed to collaborate in
advance, one may often find that it is impossible to directly compose
them. In our formalism, there does not exist a set P
C
such that
N
¼
N
1
P
C
N
2
.
