Loading presentation...

Present Remotely

Send the link below via email or IM


Present to your audience

Start remote presentation

  • Invited audience members will follow you as you navigate and present
  • People invited to a presentation do not need a Prezi account
  • This link expires 10 minutes after you close the presentation
  • A maximum of 30 users can follow your presentation
  • Learn more about this feature in our knowledge base article

Do you really want to delete this prezi?

Neither you, nor the coeditors you shared it with will be able to recover it again.


Copy of Embedded System Security

No description

Imen Khemaissia

on 5 November 2014

Comments (0)

Please log in to add your comment.

Report abuse

Transcript of Copy of Embedded System Security

Real-Time Reconfigurations of
Embedded Systems under low-power
and memory Constraints

khemaissia Imen
Future work
New reconfigurable Middleware for
feasible adaptive RT-LINUX
Reconfigurable bus CAN
New protocol for feasible re-
configurable networked controllers
Current works
Big question
general computing
add OS tasks

Remove OS tasks

Update tasks
Activate a processor

Deactivate processor

Real Embedded Characteristic
An embedded system is tightly coupled of hardware (HW) + software (SW) to perform a dedicated system
Contribution 1
How to get a feasible system after a reconfiguration scenario ?
The system should dynamically change its software in terms of new and added tasks
as well as hardware architecture without stopping its execution and without any
real-time perturbation.

This is a new challenge in industry for the
next generations of embedded systems
We are asked to propose software and hardware solutions in order to:

Minimize the power consumption

Optimize the memory capacity

Respect real-time constraints

After applying a reconfiguration scenario, processor utilization may increase and local memory processors will be full.
Architectural details

Communication protocol
Topology : Ring
Goal: Obtain a coherent system after applying
a reconfiguration scenario.
In this level we should validate the architecture that the microcontroller
desire to do it (tasks to be added/removed tasks)
Proposed solutions
Period modification of the message

WCTT modification

Message removal

Adjust the can speed
Tasks can be periodic and aperiodic
These tasks may:
share resources
Be with precedence constraint
Writing a paper for each contribution
Simulations and analysis for each
developed tool

Working on other contribution like:
the reconfigurable routing in
Reconfigurable middleware in embedded platforms
New automatic agent-based techniques for reconfigurable MP-SoC

New reconfigurable Middleware for feasible adaptive RT-LINUX

Reconfigurable CAN in real-time embedded platforms

New protocol for feasible reconfigurable
networked controllers

In this presentation we have focused on different contributions.
We have presented the different proposed solutions for each contribution.
A simulations will be done as soon as possible to highlight the effectiveness of these solutions.
Dealing with the real-implementation on smartphone will be a future work
Multi-agent based architecture

An agent master: check the evolution of the system’s environment before applying software-hardware reconfigurations.

An agent slave: It is defined for each microcontroller to evaluate its energy consumption and its memory capacity.
Solution 1: Parameter modifications of the processor that theirs utilizations greater than 1.

Solution 2: Re-allocation of the tasks according to the

Solution 3: Removal of some tasks with lower priority.

Solution 4: Activate or deactivate of processors

Solution 5: Remove tasks from local memory and add
them to shared memory.
The different functions used in the
communication protocol are:
Level 1: architecture
If the microcontroller has the authorization to add/remove some tasks, then it pass to the second level

After having authorization to apply a new architecture with a given composition, we pass to the third level. Tasks of a new reconfiguration may have the need for access to a shared data . So we have asked to manage access to these data. Each microcontroller has a matrix of access to data according to its current configuration priority.
A table is defined for each task:
Micro_id : microcontroller identity
Id_tache : Id task
Pre : precedent task
Post :next task
Response : microcontroller answer

Every microcontrooler authorizes the new composition by putting 0/1 in the field "answer" based on priority matrix

System architecture
Every STM have several tasks
The different tasks exchange message on
- When we add new tasks in controllers, new messages are added automatically on CAN, then some deadlines may be violated.
- A multi-agent based architecture is
proposed to check the correct transmission
of messages.
- If some deadlines are violated, theses agent will propose some solutions.
Architecture system
Middleware location
Proposed soltuion
Paramater modifcation of independent tasks
lambda_C of aperiodic tasks
Paramater modification of tasks
that share resources
Processor speed optimisation
(m,k) firm application
A reconfiguration request is charcterized by the following paramaters:
Micro_id : microcontroller(request sender) identity
config_id : reconfiguration to be applied identity
Response : the answer of the other microcontrollers

If all the resonses equal to 1, then the reconfiguration is applied
A simulator for RTLINUX

A tool for FPGA

A reconfiguration tool for bus CAN

Developement of an administrative
application for the network

Agent Master:
It is defined for the whole system
It will control the evolution of the system’s
environment before applying software-hardware reconfigurations.

Agent Slave:
Is defined for each processor
controls a processor to check if after a
reconfiguration scenario will not increase a lot the
energy consumption and the memory or violate a particular deadline.
Multi-agent-based architecture
2 agents are defined for each microcontroller
Agent RA: Request agent
Agent CA: Controller agent
proposed solution
proposed solutions by the agents
Communication between agent
Reconfigurations level
We have proposed:

EDF: to schedule the independent periodic tasks
FIFO: for the aperiodic tasks
IPCP: "Immediate priority ceilnig protocol" for tasks that share resources

MP-SoC is composed of :
Several processors
Local memory for each processor
Shared memory
Rt-linux is not designed to be

This layer will manage the addition/removal/update of the periodic and aperiodic tasks
It will be in interaction with the kernel Linux
This layer:
Full transcript