Vorlesung 11 Flashcards
What does “System model” describe?
Describes how the system works
What does “Architecture” describe?
Describes how the system has to be implemented
What is a system?
A system is a combination of components that act together to perform a function not possible with any of the individual parts.
What is a model?
A model is a formal description of the system, which covers selected information.
Perception vs. reality
What we perceive and how we interpret it depends on the frame through which we view the world around us.
What are the characteristics of system models?
!Homogeneous or heterogeneous
!Concurrency/parallelism
!Hierarchy
!Communication
!Synchronization models
_Q_Q_Q_Q_Q hier war noch was hintendran zu den einzelnen Charakteristika, aber kA ob wichtig Q:Q_Q_Q_Q_Q
What are some exemplary system models and what attributes do they have?
Attributes: Determinism / Concurrent / Time
Finite State Machines (FSM) (yes, no, no)
Petri Net (no, yes, no)
Kahn Process Networks (KPN) (yes, yes, no)
Data Flow (yes, yes, no)
What are Finite State Machines (FSM)?
A FSM is a model for deterministic automata, which consists of states and transitions. The current state is changed as a reaction to input signals. Thus, the state holds information about the past. FSMs are ideal for modelling control flow dominant applications such as controllers.
What are examples for FSMs?
Mealy automata, Moore automata
What’s the difference between Mealy and Moore automata
Mealy generates an output signal with every transition, Moore with every state they reach.
Mealy’s output only depends on inputs. Moore’s output only depends on states.
What are modelled system characteristics of FSMs?
not concurrent, deterministic, no explicit notion of time
control-flow-oriented
input-output-relation
Pros / cons about FSMs
pro: comprehensive
con: data-flow / calculations are not modelled
What are Petri Nets?
Petri Nets are a very general model for concurrent asynchronous systems. It consists of places holding tokens and transitions which are connected to each other. In modelling, places represent conditions and transitions represent events.
When do transitions fire?
Transitions may fire, if all input places hold a token. Firing a transition removes one token from each input place of the transition and adds one token on each output place.
What are modelled system characteristics of Petri Nets?
asynchronous
concurrent, non-deterministic, no notion of time
distributed
pros / cons of Petri Nets
+ concurrent
+ control-flow
+ highly flexible
- high complexity, can be transformed into FSM —> state explosion
- dataflow not modelled
- lack of composition and scalability
What is a Kahn Process Network (KPN)?
deterministic model for distributed systems
directed graph
nodes = processes
arcs = FIFO queues
_____________________________________________
KPN is a deterministic model for distributed systems. A KPN is a directed graph where nodes represent processes and arcs represent FIFO queues connecting processes.
Processes are deterministic and can perform non-blocking write and blokcing (empty input queue) read operations
Modelled system characteristics of KPNs
distributed
deterministic
concurrent
pros / cons of KPNs
+ synchronisation through blocking reads
+ deterministic behaviour independent of execution order
+ control flow can be hidden in node
- FIFOs in reality have bounded size
- predictability of termination and boundedness of queues are not given
What is special about Data Flow Models
Each node must produce and consume a fixed amount of tokens per execution on each of its input/output arcs/fifos
pros / cons of Data Flow Models
PROS
comprehensive
hierarchy can be modelled using extended models
CONS
Control flow is not modelled
less expressive than KPN
modelled system characteristics Data Flow Models
deterministic
dataflow
inherent concurrency
continuous time model - modelled system characteristics
concurrency, timing, deterministic
usually used together with a discrete event model
pros/cons of continuous time models
+ synchronisation possible
+ hierarchy and scalability
- restricted to a limited set of applications
- simulation requires discretization
- low performance of simulation
example of continuous time model tool
simulink
