treeBlindLP module¶

class panco.blind.treeBlindLP.TreeBlindLP(network: Network, foi: int, filename='tree_blind.lp')¶

Bases: object

Class for the analysis for tree networks with blind multiplexing using linear programming techniques.

Parameters:

network (Network) – the network to analyze (must be a well-numbered tree)
foi (int) – the flow of interest for computing the performance bound
filename (str) – the name of the file where the linear program is written

property backlog: float¶

Computes the worst-case backlog of the flow of interest

Returns:: the backlog of the flow of interest
Return type:: float

>>> flows = [Flow([TokenBucket(1, 1)], [3, 4]), Flow([TokenBucket(2, 2)], [0, 3]),
...          Flow([TokenBucket(3, 3)], [2, 4]), Flow([TokenBucket(4, 4)], [1, 3, 4])]
>>> servers = [Server([RateLatency(10, 1), RateLatency(5, 0)], []), Server([RateLatency(20, 2)], []),
...            Server([RateLatency(30, 3)], []), Server([RateLatency(40, 3)], [TokenBucket(1, 40)]),
...            Server([RateLatency(50, 5)], [])]
>>> arrival_shaping = [(3, [0], [TokenBucket(0, 10)]), (1, [3], [TokenBucket(0, 20)]),
...                    (2, [2], [TokenBucket(0, 30)]), (0, [1], [TokenBucket(0, 40)])]
>>> tree = Network(servers, flows, arrival_shaping)
>>> TreeBlindLP(tree, 0, 'test_tree_blind_backlog.lp').backlog
11.03419973

backlog_objective(file)¶

Writes the backlog objective of the flow of interest.

Parameters:: file – file where constraints are written
Returns:: nothing

backlog_set_objective(set_flows: List[int], file)¶

Writes the backlog objective of a set of flows of interest flow of interest.

Parameters:

set_flows – set of flows of interest
file – the file where constraints are written

Returns:

nothing

backlog_set_of_flows(set_flows: List[int], max_burst: float) → float¶

Computes the worst-case backlog of a set of flows of interest.

Warning: do not use if you do not master [Bou19], implemented for the iterative method.

Parameters:

set_flows (List[int]) – set of flows of interest
max_burst (float) – maximum burst during the current iteration

Returns:

the maximum backlog of the set of flows for the next iteration

Return type:

float

burst_constraints(file)¶

Write the burst constraints (in a tree, bursts are equal to the sigma parameter of the flows)

Parameters:: file – file where to write the constraints
Returns:: nothing

property delay: float¶

Computes the worst-case delay of the flow of interest

WARNING: Flows must have a single token-bucket. otherwise, put them in the arrival shaping.

Returns:: the delay of the flow of interest
Return type:: float

>>> flows = [Flow([TokenBucket(1, 1)], [3, 4]), Flow([TokenBucket(2, 2)], [0, 3]),
...          Flow([TokenBucket(3, 3)], [2, 4]), Flow([TokenBucket(4, 4)], [1, 3, 4])]
>>> servers = [Server([RateLatency(10, 1), RateLatency(5, 0)], []), Server([RateLatency(20, 2)], []),
...            Server([RateLatency(30, 3)], []), Server([RateLatency(40, 3)], [TokenBucket(1, 40)]),
...            Server([RateLatency(50, 5)], [])]
>>> arrival_shaping = [(3, [0], [TokenBucket(0, 10)]), (1, [3], [TokenBucket(0, 20)]),
...                    (2, [2], [TokenBucket(0, 30)]), (0, [1], [TokenBucket(0, 40)])]
>>> tree = Network(servers, flows, arrival_shaping)
>>> TreeBlindLP(tree, 0, 'test_tree_blind.lp').delay
10.06361149

delay_objective(file)¶

Writes the delay objective and some additional constraints for the flow of interest

Parameters:: file – file where constraints are written
Returns:: nothing