def generate_mirroring_port_utilization_compact_bar_plot(results_repository):
width = 0.25
ind = np.arange(11)
fig, ax = plt.subplots()
solution_labels = SOLUTION_LABELS
colors = cfg.BAR_PLOT_COLORS
hatch = cfg.BAR_PLOT_TEXTURES
bar_locations = [w for w in np.arange((width/2), len(solution_labels)*width, width)]
# mean_utils :: solution_type -> switch_id -> util_list
trial_name = "sub-trial-4"
mean_utils = defaultdict(lambda: defaultdict(list))
for run_name in ["run-%d" % run_idx for run_idx in range(3)]:
for solution_name in solution_labels:
topo, flows, switches, solutions, link_utilization_data = read_results(
results_repository, run_name, solution_name, trial_name)
link_ids = [(s, d) for s, t in link_utilization_data[0].items() for d in t.keys()]
collector_switch_dpid = topo_mapper.get_collector_switch_dpid()
id_to_dpid = topo_mapper.get_and_validate_onos_topo(topo)
dpid_to_id = {v: k for k, v in id_to_dpid.items()}
for s, d in [(s, d) for s, d in link_ids if d == collector_switch_dpid]:
link_utils_over_time = []
for time_idx, net_snapshot in enumerate(link_utilization_data):
try:
link_utils_over_time.append(net_snapshot[s][d])
except KeyError:
print("net_snapshot at time %d did not contain link %s -> %s" %
(time_idx, s, d))
source_switch_id = dpid_to_id[s]
mean_utils[solution_name][source_switch_id].append(
util.bytes_per_second_to_mbps(mean(link_utils_over_time[1:])))
for bar_idx, solution_name_to_switch_id in enumerate(mean_utils.items()):
solution_name, switch_id_to_util_list = solution_name_to_switch_id
data_tuples = sorted([(switch_id, mean(util_list))
for switch_id, util_list in switch_id_to_util_list.items()],
key=lambda kvp: kvp[0])
std_dev_tuples = sorted([(switch_id, np.std(util_list))
for switch_id, util_list in switch_id_to_util_list.items()],
key=lambda kvp: kvp[0])
ys = [d_i[1] for d_i in data_tuples]
y_err = [s_i[1] for s_i in std_dev_tuples]
ax.bar(ind+bar_locations[bar_idx], ys, width, color=colors[bar_idx], hatch=hatch[bar_idx],
label=solution_labels[solution_name], align="center",
ecolor="black", yerr=y_err)
plt.rc('text', usetex=True)
plt.rc('font', **cfg.FONT)
plt.xlabel("Switch ID", **cfg.AXIS_LABELS)
plt.ylabel("Switch load (Mbps)", **cfg.AXIS_LABELS)
plt.xticks(ind+(width*len(solution_labels))/2, (ind+1))
plt.grid()
plt.xlim(0, max(ind) + (width*len(solution_labels)))
# plt.legend(ncol=len(solution_labels), **cfg.LEGEND)
helpers.save_figure("sfm-plotthree.pdf", len(solution_labels))
def dropout_vs_no_dropout_plot():
with_dropout = log_file_dir / "with-dropout.log"
no_dropout = log_file_dir / "no-dropout.log"
more_dropout = log_file_dir / "dropout-0.75.log"
_, dropout = parse_log(str(with_dropout))
_, no_dropout = parse_log(str(no_dropout))
_, more_dropout = parse_log(str(more_dropout))
xs = [d_i["NumIters"] for d_i in dropout]
dropout_ys = [100 * (1 - d_i["accuracy"]) for d_i in dropout]
no_dropout_ys = [100 * (1 - d_i["accuracy"]) for d_i in no_dropout]
more_dropout_ys = [100 * (1 - d_i["accuracy"]) for d_i in more_dropout]
helpers.plot_a_line(xs,
no_dropout_ys,
label=r"$\mathbb{P}\{\text{dropout}\} = 0.0$",
idx=3,
plot_markers=False)
helpers.plot_a_line(xs,
dropout_ys,
label=r"$\mathbb{P}\{\text{dropout}\} = 0.5$",
idx=2,
plot_markers=False)
helpers.plot_a_line(xs,
more_dropout_ys,
label=r"$\mathbb{P}\{\text{dropout}\} = 0.75$",
idx=4,
plot_markers=False)
helpers.ylim((5, 30))
helpers.xlabel("Training Iteration")
helpers.ylabel(r"Validation Error (\%)")
helpers.save_figure(str(figure_output_dir / "dropout-comparison.pdf"),
num_cols=2)
def generate_mean_link_utilization_over_time_plot(parameter_name, parameter_value, trials):
"""
Generate a graph that shows the mean utilization across all the links over time
for each trial in the trial provider
"""
path_capacity = 50.0
for trial_idx, the_trial in enumerate(trials):
print(f"generate_mean_utilization_over_time_plot: {trial_idx}, {the_trial.name}")
link_utilization_over_time = the_trial.get_parameter("link-utilization-over-time")
data_for_links = {link_tuple: util_list
for link_tuple, util_list
in link_tuple_to_util_list(link_utilization_over_time).items()
if link_tuple[0] == "of:0000000000000001"}
ys = {link_tuple: [min(path_capacity, util_val) / path_capacity for util_val in util_val_list]
for link_tuple, util_val_list in data_for_links.items()}
# The next line assumes that the same number of network snapshots were captured
# for each of the links, I think this will always happen but this will throw
# if that is not the case.
throughputs_over_time = [np.mean([util_list[time_idx] for util_list in ys.values()])
for time_idx in range(len(next(iter(data_for_links.values()))))]
xs = [idx for idx in range(len(next(iter(data_for_links.values()))))]
helpers.plot_a_scatter(xs, throughputs_over_time, idx=trial_idx, label=the_trial.name)
helpers.xlabel(helpers.axis_label_font("Time"))
helpers.ylabel(helpers.axis_label_font("Mean link utilization"))
helpers.save_figure(f"mean-utilization-over-time-{parameter_name}-{parameter_value}.pdf", num_cols=3)
def generate_measured_link_utilization_cdf(trial_provider):
trial = trial_provider.get_first_trial_that_matches(
lambda t_i: t_i.name == "greedy-path-hopping")
link_utilization = trial.get_parameter("measured-link-utilization")
link_utilization_data = sorted(list(link_utilization.values()))
helpers.plot_a_cdf(link_utilization_data)
helpers.save_figure("test-cdf.pdf")
def augment_vs_no_augment():
with_augment = log_file_dir / "training-with-flip-dataset.log"
no_augment = log_file_dir / "with-dropout.log"
augment_train, augment = parse_log(str(with_augment))
no_augment_train, no_augment = parse_log(str(no_augment))
xs = [d_i["NumIters"] for d_i in augment]
augment_ys = [100 * (1 - d_i["accuracy"]) for d_i in augment]
no_augment_ys = [100 * (1 - d_i["accuracy"]) for d_i in no_augment]
helpers.plot_a_line(xs,
augment_ys,
label="Dataset augmentations",
idx=4,
plot_markers=False)
helpers.plot_a_line(xs,
no_augment_ys,
label="Original dataset",
idx=5,
plot_markers=False)
plt.ylim(7.5, 20.0)
# augment_ys = [1 - d_i["loss"] for d_i in augment_train]
# no_augment_ys = [1 - d_i["loss"] for d_i in no_augment_train]
# xs = [d_i["NumIters"] for d_i in augment_train]
# helpers.plot_a_line(xs, augment_ys, label="Dataset augmentations", idx=6, plot_markers=False)
# helpers.plot_a_line(xs, no_augment_ys, label="Original dataset", idx=7, plot_markers=False)
helpers.xlabel("Training Iteration")
helpers.ylabel(r"Validation Error (\%)")
helpers.save_figure(str(figure_output_dir /
"dataset-augmentation-comparison.pdf"),
num_cols=2)
def generate_mean_throughput_over_time_plot(parameter_name, parameter_value, trials):
"""
Generate a graph that shows the mean throughput across all the links over time for
each trial in trial provider.
"""
path_capacity = 50.0
for trial_idx, the_trial in enumerate(trials):
print(f"generate_mean_throughput_over_time: {trial_idx}, {the_trial.name}")
# number_of_paths = the_trial.get_parameter("number-of-paths")
link_utilization_over_time = the_trial.get_parameter("link-utilization-over-time")
data_for_links = {link_tuple: util_list
for link_tuple, util_list
in link_tuple_to_util_list(link_utilization_over_time).items()
if link_tuple[0] == "of:0000000000000001"}
ys = {link_tuple: [min(path_capacity, util_val) for util_val in util_val_list]
for link_tuple, util_val_list in data_for_links.items()}
throughputs_over_time = []
for time_idx in range(len(next(iter(data_for_links.values())))):
total_throughput = sum(util_list[time_idx] for util_list in ys.values())
throughputs_over_time.append(total_throughput)
xs = [idx for idx in range(len(next(iter(data_for_links.values()))))]
helpers.plot_a_scatter(xs, throughputs_over_time, idx=trial_idx, label=the_trial.name)
helpers.xlabel(helpers.axis_label_font("Time"))
helpers.ylabel(helpers.axis_label_font("Mean throughput (Mi-bps)"))
helpers.save_figure(f"throughput-over-time-{parameter_name}-{parameter_value}.pdf", num_cols=3)
def generate_learning_rate_plots():
# - step: return base_lr * gamma ^ (floor(iter / step))
# - exp: return base_lr * gamma ^ iter
# - inv: return base_lr * (1 + gamma * iter) ^ (- power)
base_lr = 0.01
gamma_step = 0.9999
gamma_inv = 0.0001
step = 1
power = 0.75
xs = [x for x in range(1, 50001, 10)]
inv_learning_rate = [
base_lr * (1 + gamma_inv * iteration)**(-power) for iteration in xs
]
step_learning_rate = [
base_lr * (gamma_step**(floor(iteration / step))) for iteration in xs
]
helpers.plot_a_line(xs,
step_learning_rate,
label="Step",
idx=6,
plot_markers=False)
helpers.plot_a_line(xs,
inv_learning_rate,
label="Inverse",
idx=7,
plot_markers=False)
helpers.xlabel("Training Iteration")
helpers.ylabel("Learning Rate")
helpers.save_figure(figure_output_dir / "learning-rate.pdf", num_cols=2)
def generate_link_utilization_cdf(results_repository, provider_name):
"""
Generate a CDF of mean link utilization
"""
trial_provider = results_repository.read_trial_provider(provider_name)
for idx, trial in enumerate(trial_provider):
# print(trial)
utilization_results = trial.get_parameter("utilization-results")
links = get_link_set(utilization_results)
print("Number of links based on utilization results: %d" % len(links))
byte_counts_per_time_period = compute_byte_counts_per_time_period(utilization_results)
pp.pprint([len(b_i) * len(next(iter(b_i.values()))) for b_i in byte_counts_per_time_period])
network_utilization_per_time_period = compute_network_utilization_per_time_period(
byte_counts_per_time_period)
pp.pprint([len(m_i) * len(next(iter(m_i.values()))) for m_i in network_utilization_per_time_period])
mean_network_utilization = compute_mean_network_utilization(
network_utilization_per_time_period)
print(len(mean_network_utilization))
link_utilizations = sorted(mean_network_utilization.values())
plot_a_cdf(link_utilizations)
plt.xlabel("Link throughput (Mbps)")
plt.ylabel("P{x < X}")
plt.legend(**cfg.LEGEND)
helpers.save_figure("%s-cdf.pdf" % provider_name, no_legend=True)
def generate_link_utilization_box_plot(results_repository, provider_name):
"""
Generate a box and whisker plot showing the utilization of all the links in the
network for a set of trials.
"""
trial_provider = results_repository.read_trial_provider(provider_name)
link_utilization_results_for_trials = []
x_axis_labels = []
for idx, trial in enumerate(trial_provider):
utilization_results = trial.get_parameter("utilization-results")
byte_counts_per_time_period = compute_byte_counts_per_time_period(utilization_results)
network_utilization_per_time_period = compute_network_utilization_per_time_period(
byte_counts_per_time_period)
mean_network_utilization = compute_mean_network_utilization(
network_utilization_per_time_period)
link_utilization_results_for_trials.append(list(mean_network_utilization.values()))
x_axis_labels.append("K=%d" % trial.get_parameter("K"))
bp = plt.boxplot(link_utilization_results_for_trials, labels=x_axis_labels,
whiskerprops={"linestyle": "--"},
flierprops={"marker":"x", "markerfacecolor": "red", "markeredgecolor": "red"})
for element in ["boxes"]:
plt.setp(bp[element], color="blue")
plt.setp(bp["medians"], color="red")
plt.ylabel("Link Throughput (Mbps)")
helpers.save_figure("%s-box-plot.pdf" % provider_name, 3, no_legend=True)
def generate_loss_rate_cdf(results_repository):
trial_name = "test-trial-830656277"
seed_numbers = [trial_name]
for idx, trial_type in enumerate(TRIAL_TYPES):
loss_rates_for_seed_number = []
for seed_number in seed_numbers:
loss_rate_per_flow = generate_loss_rates(results_repository,
trial_type, seed_number)
xs = [0.0]
ys = [0.0]
for ctr, loss_rate in enumerate(sorted(loss_rate_per_flow)):
xs.append(loss_rate)
ys.append((ctr + 1) / len(loss_rate_per_flow))
plt.plot(xs,
ys,
linestyle="-",
marker=helpers.marker_style(idx),
label=LABEL_NAMES[idx],
color=helpers.line_color(idx))
plt.rc("font", **cfg.FONT)
plt.rc("legend", edgecolor="black")
# xtick_locations = [200, 400, 600, 800, 1000]
plt.xlim(0.0)
plt.ylim(0.0, 1.0)
xtick_locations, xtick_labels = plt.xticks()
ytick_locations, ytick_labels = plt.yticks()
plt.xticks(xtick_locations[1:],
[helpers.tick_font(x_i, "%.2f") for x_i in xtick_locations][1:])
plt.yticks(ytick_locations[1:],
[helpers.tick_font(y_i, "%.1f") for y_i in ytick_locations][1:])
plt.ylabel("CDF", **cfg.AXIS_LABELS)
helpers.save_figure("loss-rate-cdf.pdf", 1)
def generate_heterogeneous_links_instantaneous_rate_plot():
link_rate_count = 100
labels = {
(100, 0): "\\LARGE{$\\mu=100$, $\\text{CoV}=0.0$}",
(100, 50): "\\LARGE{$\\mu=100$, $\\text{CoV}=0.5$}",
(100, 100): "\\LARGE{$\\mu=100$, $\\text{CoV}=1.0$}"
}
colors = {(100, 0): "red", (100, 50): "lime", (100, 100): "blue"}
plt.plot(range(link_rate_count), [100 for _ in range(link_rate_count)],
label=labels[100, 0],
color=colors[100, 0])
for mu, sigma in [(100, 50), (100, 100)]:
link_rates = heterogeneous_links.get_heterogeneous_link_rates(
mu, sigma**2, link_rate_count)
xs = range(len(link_rates))
ys = link_rates
plt.plot(xs, ys, label=labels[mu, sigma], color=colors[mu, sigma])
plt.rc('font', **cfg.FONT)
plt.xlim(0, link_rate_count - 1)
plt.xlabel("Time", **cfg.AXIS_LABELS)
plt.ylabel("Transmission Rate (Mbps)", **cfg.AXIS_LABELS)
xtick_locations, xtick_labels = plt.xticks()
ytick_locations, ytick_lables = plt.yticks()
plt.ylim(0, ytick_locations[-1])
plt.xticks(xtick_locations[1:], ["" for x_i in xtick_locations])
plt.yticks(ytick_locations[1:],
[helpers.tick_font(y_i, "%.0f") for y_i in ytick_locations][1:])
helpers.save_figure("actual-model-tx-rates.pdf", 1, no_legend=False)
def generate_max_mirror_port_utilization_bar_plot(results_repository):
# utilization_data, _ = compute_theoretical_and_actual_mean_utilization(results_repository)
# actual_error = compute_theoretical_and_actual_error(results_repository)
run_count = 3
solution_names = ["approx", "optimal"]
trial_count = 5
utilization_data, _ = compute_theoretical_and_actual_utilization_by_run(
results_repository, run_count, solution_names, trial_count)
# mean_util_lists :: solution_name -> flow_count -> [util_vals]
mean_util_lists = defaultdict(lambda: defaultdict(list))
for run_name, solution_name_to_flow_count in utilization_data.items():
for solution_name, flow_count_to_util_list in solution_name_to_flow_count.items():
for flow_count, util_list in flow_count_to_util_list.items():
mean_util_value = util.bytes_per_second_to_mbps(mean(util_list[1:]))
mean_util_lists[solution_name][flow_count].append(mean_util_value)
# std_deviations :: solution_name -> flow_count -> std_dev
std_deviations = defaultdict(dict)
# mean_utils :: solution_name -> flow_count -> mean_util
mean_utils = defaultdict(dict)
for solution_name, flow_count_to_util_list in mean_util_lists.items():
for flow_count, util_list in flow_count_to_util_list.items():
std_deviations[solution_name][flow_count] = np.std(util_list)
mean_utils[solution_name][flow_count] = mean(util_list)
pp.pprint(mean_utils)
width = 0.35
ind = np.arange(1, 6)
fig, ax = plt.subplots()
labels = SOLUTION_LABELS
half = len(labels) // 2
bar_locations = [w for w in np.arange((width/2), len(labels)*width, width)]
colors = cfg.BAR_PLOT_COLORS
hatch = cfg.BAR_PLOT_TEXTURES
for bar_idx, solution_name_to_flow_count in enumerate(mean_utils.items()):
solution_name, flow_count_to_mean_util = solution_name_to_flow_count
data_tuples = sorted([(flow_count, util_val)
for flow_count, util_val in flow_count_to_mean_util.items()],
key=lambda kvp: kvp[0])
yerr_tuples = sorted([(flow_count, std_dev)
for flow_count, std_dev in std_deviations[solution_name].items()],
key=lambda kvp: kvp[0])
xs = [d_i[0] for d_i in data_tuples]
ys = [d_i[1] for d_i in data_tuples]
yerr_values = [s_i[1] for s_i in yerr_tuples]
ax.bar(ind+bar_locations[bar_idx], ys, width, color=colors[bar_idx], hatch=hatch[bar_idx],
label=labels[solution_name], yerr=yerr_values, align="center",
ecolor="black")
plt.rc('text', usetex=True)
plt.rc('font', **cfg.FONT)
plt.grid(**cfg.GRID)
plt.xticks(ind+(width*len(labels))/2, ind*10)
plt.xlabel("Number of Flows", **cfg.AXIS_LABELS)
plt.ylabel("Maximum switch load (Mbps)", **cfg.AXIS_LABELS)
# plt.legend(ncol=2, **cfg.LEGEND)
helpers.save_figure("sfm-objective.pdf", 2)
def generate_data_recovery_vs_time_cdf(trial_provider):
for trial in trial_provider:
if trial.get_parameter("path-length") == 10:
random_path_attacker = trial.get_parameter(
"random-path-hopping-attacker-recovered-messages")
random_node_attacker = trial.get_parameter(
"random-node-hopping-attacker-recovered-messages")
ideal_path_attacker = trial.get_parameter(
"ideal-random-path-hopping-attacker-recovered-messages")
one_node_per_path_attacker_captured = trial.get_parameter(
"one-node-per-path-attacker")
fixed_attacker_captured = trial.get_parameter("fixed-attacker")
helpers.plot_a_cdf(sorted(random_path_attacker),
idx=0,
label="Random Path Attacker",
plot_markers=False)
helpers.plot_a_cdf(sorted(random_node_attacker),
idx=1,
label="Random Node Attacker",
plot_markers=False)
helpers.plot_a_cdf(sorted(one_node_per_path_attacker_captured),
idx=2,
label="One Node per Path Attacker",
plot_markers=False)
helpers.plot_a_cdf(sorted(fixed_attacker_captured),
idx=3,
label="Fixed Attacker",
plot_markers=False)
helpers.save_figure("data-recovery-cdf.pdf", num_cols=2)
def generate_theoretical_vs_actual_utilization_bar_plot(results_repository):
utilization_data, theoretical_data = compute_theoretical_and_actual_mean_utilization(results_repository)
actual_error = compute_theoretical_and_actual_error(results_repository)
std_deviations = defaultdict(list)
for provider_name, data_list in actual_error.items():
for flow_count, data in data_list:
std_deviations[provider_name].append(np.std(
[util.bytes_per_second_to_mbps(d_i) for d_i in data]))
width = 0.35
ind = np.arange(1, 6)
fig, ax = plt.subplots()
xs = [t[0] for t in utilization_data["approx"]]
ys = [util.bytes_per_second_to_mbps(t[1]) for t in utilization_data["approx"]]
ax.bar(ind-(width/2), ys, width, color="skyblue", hatch=".", tick_label=xs, label="Measured",
yerr=std_deviations["approx"], ecolor="black")
xs = [t[0] for t in theoretical_data["approx"]]
ys = [t[1] * 100 for t in theoretical_data["approx"]]
ax.bar(ind+(width/2), ys, width, color="green", hatch="\\", tick_label=xs, label="Expected")
plt.rc('text', usetex=True)
plt.rc('font', **cfg.FONT)
plt.xlabel("Number of Flows")
plt.ylabel("Maximum mirroring port rate ($\\frac{Mb}{s}$)")
plt.legend(loc="upper center", bbox_to_anchor=(0.5, cfg.LEGEND_HEIGHT), shadow=True, ncol=2)
helpers.save_figure("sfm-theorypractice.pdf")
def generate_per_class_accuracy_bar_plot():
conf_mat_1 = more_dropout
conf_mat_2 = undersampled
fig, (ax1, ax2) = plt.subplots(2)
bar_width = 0.1
bar_1_xs = np.arange(0.0, 0.3 * 6.5, 0.3)
bar_2_xs = [x_i + bar_width for x_i in bar_1_xs]
print(bar_1_xs, bar_2_xs)
bar_1_ys = [
100 *
(conf_mat_1[idx][idx] / sum(get_samples_that_are(conf_mat_1, idx + 1)))
for idx in range(7)
]
bar_2_ys = [
100 *
(conf_mat_2[idx][idx] / sum(get_samples_that_are(conf_mat_2, idx + 1)))
for idx in range(7)
]
plt.xticks([x_i + 0.5 * bar_width for x_i in bar_1_xs],
[1, 2, 3, 4, 5, 6, "Outlier"])
helpers.plot_a_bar(bar_1_xs,
bar_1_ys,
idx=0,
label="Augmented Dataset",
bar_width=bar_width,
axis_to_plot_on=ax2)
helpers.plot_a_bar(bar_2_xs,
bar_2_ys,
idx=1,
label="Undersampled Dataset",
bar_width=bar_width,
axis_to_plot_on=ax2)
helpers.xlabel("Class label")
helpers.ylabel(r"Validation Accuracy (\%)", formatter=lambda x: x, ax=ax2)
undersampled_dataset = json.loads(
path.Path("./segments-undersampled.json").read_text())
histogram = Counter([
d_i["data"]["segment_type"]["data"]
for d_i in undersampled_dataset.values()
])
xs = [x_i for x_i in range(1, 8)]
total_samples = sum(histogram.values())
ys = [c[1] / total_samples for c in sorted(histogram.items())]
helpers.plot_a_bar(xs, ys, idx=1, axis_to_plot_on=ax1, label_data=False)
helpers.ylabel(r"$\mathbb{P}\{x = \mathcal{X}\}$",
formatter=lambda x: x,
ax=ax1)
ax1.set_yticks([0.1, 0.2, 0.3])
ax1.grid(**cfg.GRID)
ax1.xaxis.set_ticklabels([])
legend_params = deepcopy(cfg.LEGEND)
legend_params["bbox_to_anchor"] = (0.5, 0.975)
fig.legend(**legend_params, ncol=2)
helpers.save_figure(figure_output_dir / "per-class-error-bar-plot.pdf",
no_legend=True)
def generate_simulation_run_time_plot(results_repository):
def mean_confidence_interval(data, confidence=0.95):
a = 1.0 * np.array(data)
n = len(a)
m, se = np.mean(a), scipy.stats.sem(a)
h = se * scipy.stats.t.ppf((1 + confidence) / 2., n - 1)
# return m, m-h, m+h
return h
series_data = {}
for flow_count in [100, 500, 1000]:
data_for_node_counts = []
for node_count in [100, 200, 300, 400, 500, 600, 700, 800, 900, 1000]:
data_for_single_count = []
seed_numbers = [
742349042, 2787879137, 3102739139, 2303690384, 1672982999,
2501915964, 2853959085, 3163893110, 462786850, 4124106680
]
for seed_number in seed_numbers:
sim = read_results(results_repository, flow_count, node_count,
seed_number)
data_for_single_count.append(sim.solution_time)
pp.pprint(data_for_single_count)
ci = mean_confidence_interval(data_for_single_count)
data_for_node_counts.append(
(node_count, np.mean(data_for_single_count), ci))
series_data[flow_count] = data_for_node_counts
for idx, (flow_count, data_for_node_counts) in enumerate(
sorted(series_data.items(), key=lambda t: t[0])):
xs = [d_i[0] for d_i in data_for_node_counts]
ys = [d_i[1] for d_i in data_for_node_counts]
std_dev = [d_i[2] for d_i in data_for_node_counts]
plt.errorbar(xs,
ys,
label="\\LARGE{%d VLs}" % flow_count,
marker=cfg.MARKER_STYLE[idx],
linestyle=helpers.line_style(idx),
color=helpers.line_color(idx),
yerr=std_dev,
capsize=2)
plt.rc("text", usetex=True)
matplotlib.rcParams['text.latex.preamble'] = [
r'\usepackage{amsmath}', r'\usepackage{amssymb}'
]
plt.rc("font", **cfg.FONT)
plt.rc("legend", edgecolor="black")
xtick_locations = [200, 400, 600, 800, 1000]
ytick_locations = [2, 4, 6, 8, 10, 12, 14]
plt.xticks(xtick_locations,
[helpers.tick_font(x_i) for x_i in xtick_locations])
plt.yticks(ytick_locations,
[helpers.tick_font(y_i) for y_i in ytick_locations])
plt.xlabel("Number of Nodes", **cfg.AXIS_LABELS)
plt.ylabel("Running Time (s)", **cfg.AXIS_LABELS)
helpers.save_figure("sim-running-time.pdf", 1)
def generate_substrate_topology_graph():
# trial_provider = results_repository.read_trial_provider("multiflow-tests")
# topo = trial_provider.get_metadata("substrate-topology")
topo = nx.complete_graph(10)
the_layout = nx.circular_layout(topo)
# nx.draw(topo, node_color="skyblue", layout=the_layout)
nx.draw_circular(topo, node_color="skyblue")
helpers.save_figure("network-topology.pdf", no_legend=True)
def generate_learning_rate_comparison_plot():
_, validation_data_step = parse_log(
str(log_file_dir / "lr-step-training.log"))
_, validation_data_inv = parse_log(
str(log_file_dir / "training-with-flip-dataset.log"))
fig, (plot1, plot2) = plt.subplots(2)
xs = [d_i["NumIters"] for d_i in validation_data_step]
ys_step = [100 * (1 - d_i["accuracy"]) for d_i in validation_data_step]
ys_inv = [100 * (1 - d_i["accuracy"]) for d_i in validation_data_inv]
plot2.set_ylim(7.5, 20)
helpers.xlabel("Training Iterations", ax=plot2)
helpers.ylabel(r"Validation Error (\%)", ax=plot2, formatter=lambda x: x)
helpers.plot_a_line(xs,
ys_inv,
label="Step",
plot_markers=False,
idx=6,
axis_to_plot_on=plot2)
helpers.plot_a_line(xs,
ys_step,
label="Inverse",
plot_markers=False,
idx=7,
axis_to_plot_on=plot2)
base_lr = 0.01
gamma_step = 0.9999
gamma_inv = 0.0001
step = 1
power = 0.75
xs = [x for x in range(1, 50001, 10)]
inv_learning_rate = [
base_lr * (1 + gamma_inv * iteration)**(-power) for iteration in xs
]
step_learning_rate = [
base_lr * (gamma_step**(floor(iteration / step))) for iteration in xs
]
helpers.plot_a_line(xs,
step_learning_rate,
idx=6,
plot_markers=False,
axis_to_plot_on=plot1)
helpers.plot_a_line(xs,
inv_learning_rate,
idx=7,
plot_markers=False,
axis_to_plot_on=plot1)
plot1.xaxis.set_ticklabels([])
plot1.grid(**cfg.GRID)
helpers.ylabel("Learning Rate", ax=plot1, formatter=lambda x: x)
legend_params = deepcopy(cfg.LEGEND)
legend_params["bbox_to_anchor"] = (0.5, 0.975)
fig.legend(ncol=2, **legend_params)
helpers.save_figure(figure_output_dir / "learning-rate-comparison.pdf",
no_legend=True)
def generate_expected_link_utilization_cdf(results_repository, provider_name):
trial_provider = results_repository.read_trial_provider(provider_name)
the_trial = trial_provider.get_first_trial_that_matches(
lambda t: t.get_parameter("trial-name") == "optimal-3")
expected_utilization_results = the_trial.get_parameter("link-utilization")
link_utilizations = sorted(expected_utilization_results.values())
plot_a_cdf(link_utilizations)
helpers.save_figure("expected-link-utilization-cdf.pdf", no_legend=True)
def generate_port_mirroring_port_utilization_cdf(results_repository):
labels = ["rnd", "det", "df", "greedy", "optimal"]
legend_labels = ["$\\epsilon$-LPR", "LPR", "DuFi", "Greedy", "Optimal"]
markers = ["^", "*", "^", "o", "x"]
colors = ["red", "green", "blue", "orange", "purple"]
trial_name = "sub-trial-4"
run_name = "run-0"
for solution_idx, solution_name in enumerate(labels):
topo, flows, switches, solutions, link_utilization_data, ports = read_results(
results_repository, run_name, solution_name, trial_name)
link_ids = [(s, d) for s, t in link_utilization_data[0].items()
for d in t.keys()]
mean_utils = []
labels = []
errors = []
collector_switch_dpid = topo_mapper.get_collector_switch_dpid()
id_to_dpid = topo_mapper.get_and_validate_onos_topo(topo)
dpid_to_id = {v: k for k, v in id_to_dpid.items()}
for s, d in [(s, d) for s, d in link_ids
if d == collector_switch_dpid]:
link_utils_over_time = []
for time_idx, net_snapshot in enumerate(link_utilization_data):
try:
link_utils_over_time.append(net_snapshot[s][d])
except KeyError:
print(
"net_snapshot at time %d did not contain link %s -> %s"
% (time_idx, s, d))
mean_utils.append(
util.bytes_per_second_to_mbps(mean(link_utils_over_time)))
errors.append(
util.bytes_per_second_to_mbps(np.std(link_utils_over_time)))
labels.append(dpid_to_id[s])
cdf_data = sorted(mean_utils)
xs = [0.0]
ys = [0.0]
for idx, d in enumerate(cdf_data):
xs.append(d)
ys.append((1 + idx) / len(cdf_data))
plt.plot(xs,
ys,
label=legend_labels[solution_idx],
marker=markers[solution_idx],
color=colors[solution_idx])
plt.legend(loc="upper center",
bbox_to_anchor=(0.5, cfg.LEGEND_HEIGHT),
shadow=True,
ncol=len(labels))
plt.rc('text', usetex=True)
plt.rc('font', **cfg.FONT)
plt.grid()
plt.xlabel("Mirroring Port Rate $\\frac{Mb}{s}$")
plt.ylabel("$\\mathbb{P}\\{x < \\mathcal{X}\\}$")
helpers.save_figure("pm-plot-three-cdf.pdf")
def generate_node_probability_histogram(trial_provider):
discrete_probability_distribution = trial_provider.get_metadata("node-probability-distribution")
bar_x_locations = np.arange(len(discrete_probability_distribution))
plt.bar(bar_x_locations, discrete_probability_distribution, hatch="/")
plt.xticks(bar_x_locations, ["" for _ in bar_x_locations])
plt.xlabel("Node ID")
plt.ylabel(r"$\mathbb{P}\{n \in \{s, t\}\}$")
node_selection_type = trial_provider.get_metadata("node-selection-type")
helpers.save_figure("node-probability-distribution-%s.pdf" % node_selection_type,
no_legend=True)
def generate_dataset_histogram(dataset, output_path):
histogram = Counter(
[d_i["data"]["segment_type"]["data"] for d_i in dataset.values()])
xs = [1, 2, 3, 4, 5, 6, 7]
ys = [c[1] / sum(histogram.values()) for c in sorted(histogram.items())]
plt.xticks(xs, [1, 2, 3, 4, 5, 6, "Outlier"])
helpers.ylabel(r"$\mathbb{P}\{x = \mathcal{X}\}$")
helpers.xlabel("Class Label")
helpers.plot_a_bar(xs, ys, idx=1)
helpers.save_figure(output_path, no_legend=True)
def generate_traffic_on_path_bar_plot(data_file):
def read_literal_from_file(data_file):
with data_file.open("r") as fd:
python_literal = eval(fd.read())
return python_literal
path_data = read_literal_from_file(data_file)
xs = []
ys = []
target_path_ratios = [0.5, 0.3, 0.2]
cs = []
bar_width = 0.2
pos = 0.1
for path_idx, (k, v) in enumerate(
sorted(path_data.items(), key=lambda t: t[1], reverse=True)):
y_val = v / sum(path_data.values())
cs.append("blue")
cs.append("green")
plt.bar(pos - 0.01, y_val, bar_width, color="gray", edgecolor="black")
pos += bar_width
plt.bar(pos + 0.01,
target_path_ratios[path_idx],
bar_width,
color="blue",
edgecolor="black")
pos += bar_width
pos += 0.1
print("VALUE: %f" % (abs(y_val - target_path_ratios[path_idx]) /
target_path_ratios[path_idx]))
# labels=["Path %d" % path_idx for path_idx in range(1, 4)]
# plt.bar(xs, ys, edgecolor="black", color=cs)
legend_labels = [
"\\Large{expected data volume}", "\\Large{actual data volume}"
]
legend_colors = ["blue", "gray"]
patches = [
matplotlib.patches.Patch(color=c, label=l, edgecolor="black")
for c, l in zip(legend_colors, legend_labels)
]
plt.legend(handles=patches, ncol=2, **cfg.LEGEND)
xtick_labels = ["Path %d" % path_id for path_id in range(1, 4)]
xtick_locations = [l_i for l_i in np.arange(0.2, 1.5, 0.5)]
plt.xticks(xtick_locations, xtick_labels)
ytick_locations, ytick_labels = plt.yticks()
plt.yticks(ytick_locations,
[helpers.tick_font(y_i, "%.1f") for y_i in ytick_locations])
plt.ylabel("Normalized Data Volume", **cfg.AXIS_LABELS)
helpers.save_figure("per-path-data-volume.pdf", no_legend=True)
def generate_mirroring_port_utilization_bar_plot(results_repository):
topo, flows, switches, solutions, link_utilization_data = read_results(results_repository, "approx", "sub-trial-4")
link_ids = [(s, d) for s, t in link_utilization_data[0].items() for d in t.keys()]
mean_utils = []
labels = []
errors = []
collector_switch_dpid = topo_mapper.get_collector_switch_dpid()
id_to_dpid = topo_mapper.get_and_validate_onos_topo(topo)
dpid_to_id = {v: k for k, v in id_to_dpid.items()}
for s, d in [(s, d) for s, d in link_ids if d == collector_switch_dpid]:
link_utils_over_time = []
for time_idx, net_snapshot in enumerate(link_utilization_data):
try:
link_utils_over_time.append(net_snapshot[s][d])
except KeyError:
print("net_snapshot at time %d did not contain link %s -> %s" % (time_idx, s, d))
mean_utils.append(util.bytes_per_second_to_mbps(mean(link_utils_over_time)))
errors.append(util.bytes_per_second_to_mbps(np.std(link_utils_over_time)))
labels.append(dpid_to_id[s])
ind = np.arange(1, 12)
width = 0.35
fig, ax = plt.subplots()
ys = [mu for mu, l in sorted(zip(mean_utils, labels), key=lambda t: t[1])]
xs = labels
errors = [e for e, l in sorted(zip(errors, labels), key=lambda t: t[1])]
ax.set_xticks(ind+(width/2))
ax.set_xticklabels(range(1, 12))
ax.bar(ind-(width/2), ys, width, label="Measured", color="skyblue", hatch=".",
yerr=errors, ecolor="black")
mirroring_utils = defaultdict(float)
for flow in flows.values():
mirroring_switch_for_flow = solutions[flow.flow_id].mirror_switch_id
mirroring_utils[mirroring_switch_for_flow] += flow.traffic_rate
ys = []
xs = []
for switch_id in range(1, 12):
aggregate_rate = mirroring_utils[switch_id]
xs.append(switch_id)
ys.append(aggregate_rate * 100)
ax.bar(ind+(width/2), ys, width, color="green", hatch="\\", label="Expected")
plt.legend(loc="upper center", bbox_to_anchor=(0.5, cfg.LEGEND_HEIGHT), shadow=True, ncol=2)
plt.rc('text', usetex=True)
plt.rc('font', **cfg.FONT)
plt.xlabel("Switch ID")
plt.ylabel("Mean Mirroring Port Rate ($\\frac{Mb}{s}$)")
helpers.save_figure("plot-three.pdf")
plt.clf()
def generate_theoretical_vs_actual_utilization_bar_plot(results_repository):
utilization_data, theoretical_data = compute_theoretical_and_actual_mean_utilization(
results_repository)
actual_std_deviation = compute_actual_std_deviation(results_repository)
width = 0.35
labels = ["rnd", "det", "df", "greedy", "optimal"]
legend_labels = ["\\epsilon-LPR", "LPR", "DuFi", "Greedy", "Optimal"]
colors = ["orange", "green", "skyblue", "purple", "yellow"]
hatch = ["//", "\\", "//", "\\", "//"]
for solution_name in labels:
ind = np.arange(1, 6)
fig, ax = plt.subplots()
data_tuples = sorted(
[(k, v) for k, v in utilization_data[solution_name].items()],
key=lambda kvp: kvp[0])
yerr_values = actual_std_deviation[solution_name]
xs = [flow_count for flow_count, _ in data_tuples]
ys = [util.bytes_per_second_to_mbps(data) for _, data in data_tuples]
ax.bar(ind - (width / 2),
ys,
width,
color="green",
hatch="\\",
tick_label=xs,
label="Measured",
yerr=yerr_values,
ecolor="black")
theoretical_tuples = sorted(
[(k, v) for k, v in theoretical_data[solution_name].items()],
key=lambda kvp: kvp[0])
xs = [flow_count for flow_count, _ in theoretical_tuples]
ys = [
util_val * pm_cfg.rate_factor for _, util_val in theoretical_tuples
]
ax.bar(ind + (width / 2),
ys,
width,
color="skyblue",
hatch=".",
tick_label=xs,
label="Expected")
plt.rc('text', usetex=True)
plt.rc('font', **cfg.FONT)
plt.xlabel("Number of Flows")
plt.ylabel("Maximum mirroring port rate ($\\frac{Mb}{s}$)")
plt.grid()
plt.legend(ncol=len(labels), **cfg.LEGEND)
helpers.save_figure("plot-two-%s.pdf" % solution_name)
def generate_link_utilization_box_plot(results_repository):
list_of_mean_utils = []
for trial_type in TRIAL_TYPES:
mean_link_utils_per_seed_number = []
for idx, seed_number in enumerate(["test-trial-830656277"]):
utilization_results, the_vle_trial, end_host_results = read_results(
results_repository, trial_type, seed_number)
link_utilization_data = compute_network_util_over_time(
utilization_results)
link_ids = [(s, d) for s, t in link_utilization_data[0].items()
for d in t.keys()]
mean_utils = {}
for s, d in link_ids:
link_utils_over_time = []
for time_idx, net_snapshot in enumerate(link_utilization_data):
try:
link_utils_over_time.append(net_snapshot[s][d])
except KeyError:
print(
"net_snapshot at time %d did not contain link %s -> %s"
% (time_idx, s, d))
mean_utils[s, d] = ((mean(link_utils_over_time) * 8) / 10**6 /
1000)
list_of_mean_utils.append(mean_utils)
pp.pprint([list(v.values()) for v in list_of_mean_utils])
bp = plt.boxplot([list(v.values()) for v in list_of_mean_utils],
labels=LABEL_NAMES,
whiskerprops={"linestyle": "--"},
flierprops={
"marker": "x",
"markerfacecolor": "red",
"markeredgecolor": "red"
})
for element in ["boxes"]:
plt.setp(bp[element], color="blue")
plt.setp(bp["medians"], color="red")
plt.rc("font", **cfg.FONT)
plt.rc("legend", edgecolor="black")
# xtick_locations = [200, 400, 600, 800, 1000]
# ytick_locations = [2, 4, 6, 8, 10, 12, 14]
ytick_locations, _ = plt.yticks()
plt.yticks(ytick_locations,
[helpers.tick_font(y_i, "%.1f") for y_i in ytick_locations])
plt.ylabel("Link Utilization", **cfg.AXIS_LABELS)
helpers.save_figure("link-utilization-box-plot.pdf", 3, no_legend=True)
def generate_theoretical_vs_actual_compact_bar_plot(results_repository):
utilization_data, theoretical_data = compute_theoretical_and_actual_mean_utilization(results_repository)
actual_error = compute_theoretical_and_actual_error(results_repository)
std_deviations = defaultdict(list)
for provider_name, data_list in actual_error.items():
for flow_count, data in data_list:
std_deviations[provider_name].append(np.std(
[util.bytes_per_second_to_mbps(d_i) for d_i in data]))
width = 0.35
ind = np.arange(5)
fig, ax = plt.subplots()
labels = ["optimal", "approx"]
half = len(labels)//2
legend_labels = ["Optimal", "BiSec"]
colors = cfg.BAR_PLOT_COLORS
hatch = [".", "\\"]
bar_locations = [w for w in np.arange((width/2), len(labels)*width, width)]
pp.pprint(utilization_data)
pp.pprint(theoretical_data)
for bar_idx, solution_name in enumerate(labels):
data_tuples = sorted([(k, v) for k, v in utilization_data[solution_name]],
key=lambda kvp: kvp[0])
theoretical_tuples = sorted([(k, v) for k, v in theoretical_data[solution_name]],
key=lambda kvp: kvp[0])
xs = [flow_count for flow_count, _ in data_tuples]
measured_ys = [util.bytes_per_second_to_mbps(data)
for _, data in data_tuples]
theoretical_ys = [util_val * pm_cfg.rate_factor
for _, util_val in theoretical_tuples]
theoretical_ys = [abs(theoretical_ys[idx] - measured_ys[idx])
for idx in range(len(theoretical_ys))]
ax.bar(ind+bar_locations[bar_idx], theoretical_ys, width, color=colors[bar_idx],
hatch=hatch[bar_idx], label=legend_labels[bar_idx],
align="center", ecolor="black")
plt.rc('text', usetex=True)
plt.rc('font', **cfg.FONT)
plt.xlabel("Number of Flows")
plt.ylabel("Maximum switch load (Mbps)")
plt.xticks(ind+(width*len(labels))/2, (ind+1)*10)
plt.grid()
plt.xlim(0, max(ind) + (width*len(labels)))
plt.legend(loc="upper center", bbox_to_anchor=(0.5, cfg.LEGEND_HEIGHT),
shadow=True, ncol=len(labels))
helpers.save_figure("sfm-theorypractice.pdf")
def generate_number_of_successful_requests_bar_plot(parameter_name, parameter_value, trials):
"""
Generate a bar plot showing the number of requests that completed successfully in each of
the trials.
"""
bar_width = 0.2
bar_x_locations = np.arange(0.0, (len(trials)+1)*2*bar_width, 2*bar_width)
bar_labels = [the_trial.name.replace("-approximate", "") for the_trial in trials]
for idx, the_trial in enumerate(trials):
print(f"generate_number_of_successful_requests_bar_plot: {idx}, {the_trial.name}")
y_value = the_trial.get_parameter("number-of-successful-flows")
helpers.plot_a_bar(bar_x_locations[idx], y_value, label=bar_labels[idx], bar_width=bar_width, idx=idx)
plt.xticks(bar_x_locations, bar_labels)
helpers.save_figure(f"successful-requests-bar-{parameter_name}-{parameter_value}.pdf", no_legend=True)
def generate_computed_link_utilization_box_plot(trial_provider):
grouped_by_name = collect_trials_based_on_name(trial_provider)
labels = [helpers.axis_label_font(multiflow_label(group[0].name))
for group in grouped_by_name]
# Make lists of link utilization data
link_utilization_data = [reduce(op.add,
[list(t_i.get_parameter("measured-link-utilization").values())
for t_i in group
if t_i.has_parameter("measured-link-utilization")], [])
for group in grouped_by_name]
plot_a_box_plot(link_utilization_data, labels)
plt.ylabel(helpers.axis_label_font("Link utilization"))
node_selection_type = trial_provider.get_metadata("node-selection-type")
helpers.save_figure("computed-link-utilization-%s-box.pdf" % node_selection_type,
no_legend=True)
def plot_share_delay_histogram(packets):
"""
Takes a list of PacketInfo tuples as input and plots the PDF of the intershare delay.
"""
seq_num_to_ts = defaultdict(list)
for p_i in packets:
seq_num_to_ts[p_i.seq_num].append(p_i.timestamp)
inter_share_delays = []
for seq_num, ts in seq_num_to_ts.items():
inter_share_delay = max(ts) - min(ts)
inter_share_delays.append(inter_share_delay)
plt.histogram(inter_share_delays, density=True)
helpers.save_figure("share-delay-pdf.pdf")