Information Technology Reference
In-Depth Information
The actions specified for a given system may depend on the claims that need to
be evaluated. Other actions, such as confirmation of the reservation between
Hotel
and
Agency
, were not considered here to keep the simplicity of the example.
3.2 Logs and Distribution
An event (
EV ENT
) is described as a tuple consisting of the type of event
(
OP
), the source and destination agents, and the action to be executed. A log
file (
LOG FILE
) is defined as a pair consisting of the set of agents that have
their events logged and the sequence of events, recorded in the order of execution.
A log architecture (
LOG ARCH
) represents a set of logs produced during
the execution of the system. The constant
Dist
describes how we have chosen
to regroup and distribute logs. Each element
X
Dist
represents that a single
log file records the events performed by the agents in
X
. Notice that in this
definition the events of an agent might be recorded in more than one log file.
Logs are modeled using the machine
LogModel
.
MACHINE
LogModel
INCLUDES
SystemInfo
SETS
OP
={
Send
,
Rec
}
CONSTANTS
EV ENT
,
LOG FILE
,
LOG ARCH
,
Dist
PROPERTIES
EV ENT
∈
=
OP
×
AGEN T
×
AGEN T
×
ACT ION
∧
/* Log files */
LOG FILE
)
∧
/* Definition of the chosen log distribution */
Dist
=
F
(
AGEN T
)
×
iseq
(
EV ENT
=
{···}∧
/* Log architectures defined as a set of log files */
LOG ARCH
⊆F
(
AGENT
)
∧
Dist
=
{
logs
|
logs
∈F
(
LOG FILE
)
∧∀
log.
(
log
∈
logs
⇒
agents
(
log
)
∈
Dist
)
}
END
In the rest of the paper we use the functions
agents
and
content
that maps
every log file into its agents and content respectively. The function
events
maps
every log file into the set of events (rather than the sequence) of the log's content.
We abuse our notation and assume that
agents
and
events
can also be applied
to log architectures: applying
agents
(or
events
)toalogarchitecture
log arch
is equivalent to applying
agents
(or
events
) to each log file that belongs to
log arch
and make the union of the results.
Example 4.
(log distribution) Let us consider the following two examples of dis-
tributions:
1.
Dist
=
{{
Client
}
,
{
Agency
}
,
{
Hotel
}
,
{
Bank
}}
}}
In the first example we describe a distribution where each agent is logged inde-
pendently. In the second example we assume that there is a single log file that
records the events performed by
Client
and
Agency
.
2.
Dist
=
{{
Client
,
Agency
}
,
{
Hotel
}
,
{
Bank
Search WWH ::
Custom Search