Database Reference
In-Depth Information
the resource performance to query complexity. The analysed phase will sort out,
according to the requirements of the IoT user/admin or Agent designer -
“
how the
system is doing
.
The analysis phase has been separated in two different segments. After analysing
the IoT system health, the agent will be able to determine what changes are required to
put the IoT in a best condition. This can be maximizing resource, or can be minimizing
some other parameters that are consuming more resource than necessary. In case of
dealing with other elements of the IoT services, the agent needs to communicate with
other agents of the system. Agents will communicate with other agents, they will have
a common goal, they will exchange information, and decide based on exchanged
information the actions to take. The approach has been focused on the performance
maximization parameters of the agent based on local Analysis. Note that the com-
munication with other agents have not been implemented because the design of
autonomic agent is still under-way and the communication between different agents
and learning from each other and deciding on a common goal with the help of complex
algorithm is out of focus of this research at this current stage. The planning phase
should be able to determine what are the changes that might have an effect on the
performance improvement of the IoT system. During the planning phase, the auto-
nomic agent will inform the hosting infrastructure, or the owner of the system to
change the environment according to its analysis done in the previous cycle.
The agent will inform the platform or the owner to increase the memory. This
communication can be reverted as well. It is also possible that the owner con
”
, and put this status in its
“
record book
”
gure the
total infrastructure in such a way that the IoT system has to tune it up according to
the limited resources available. For example, in case the owner does not want to
increase the underlying platform, or set an upper limit of upgrade, and the IoT system
has already reached that limit of resource usage. At this moment, the only option to
make the system perform better is to add more indexes, or to reduce the maximum
buffer size. In this scenario, the IoT Autonomic Agent will do the communication with
the IoT service itself and in the execution phase, it will do the tuning accordingly and
make sure that the system is performing the best with the resource available. The
following logic has been implemented to set the status of the autonomic agent. This
algorithm checks if the pair of allocated_resource and system_parameters supposed to
perform in the best form. The agent checks the condition and set the system_status by
following the algorithm as follow:
while true
do
check allocated_resource
check system_parameters
map allocated_resource -> system_parameters
set system_status
The later part of the logic will check the system_status for the IoT platform. If the
system_status is Ok, then the IoT system has already been tuned up for the best
performance with available resource. If not, then the PLAN will happen.
Search WWH ::
Custom Search