def laod_data(dim):
# Execute workload a for every dim
print("\n\n==> Loading data: " + str(dim) + " record")
command = commons.BIN_YCSB + "/ycsb.sh load jdbc -s -P " + commons.HOME_YCSB + "/workloads/workloada -P " + commons.SCRIPTS + "/postgres-prop -p recordcount=" + str(
dim
) + " -p db.url=" + postgres_connection_string + " > " + commons.getOutputFilename(
output_dir, "Load", None, dim, None, None)
commons.executeBashCommand(command)
def run_workload(workload, dim, t, c, op):
command = commons.BIN_YCSB + "/ycsb.sh run jdbc -s -P " + commons.HOME_YCSB + "/workloads/" + workload + " -P " + commons.SCRIPTS + "/postgres-prop -p recordcount=" + str(
dim
) + " -threads " + str(c) + " -target " + str(
t
) + " -p db.url=" + postgres_connection_string + " -p operationcount=" + str(
op) + " > " + commons.getOutputFilename(output_dir, "Run", workload,
dim, t, c)
commons.executeBashCommand(command)
def run_workload(workload, dim, t, c, op, usage) :
if usage=="local-docker":
command = commons.BIN_YCSB + "/ycsb.sh run voltdb -P " + commons.HOME_YCSB + "/workloads/" + workload + " -P " + commons.SCRIPTS + "/voltdb-prop > " + commons.getOutputFilename(output_dir, "Run", workload, dim, t, c)
elif usage=="local":
command = commons.BIN_YCSB + "/ycsb.sh run voltdb -P " + commons.HOME_YCSB + "/workloads/" + workload + " > " + commons.getOutputFilename(output_dir, "Run", workload, dim, t, c)
commons.executeBashCommand(command)
def laod_data(dim, usage) :
# Execute workload a for every dim
print("\n\n==> Loading data: " + str(dim) + " record")
if usage=="local-docker":
command = commons.BIN_YCSB + "/ycsb.sh load voltdb -P " + commons.HOME_YCSB + "/workloads/workloada -P " + commons.SCRIPTS + "/voltdb-prop > " + commons.getOutputFilename(output_dir, "Load", None, dim, None, None)
elif usage=="local":
command = commons.BIN_YCSB + "/ycsb.sh run voltdb -P " + commons.HOME_YCSB + "/workloads/workloada > " + commons.getOutputFilename(output_dir, "Run", None, dim, None, None)
commons.executeBashCommand(command)
def generateAndShowThroughputLatencyPlotsComparison(input_dim, throughputs,
clients, sources,
workload):
for c in clients:
for dim in input_dim:
# Generate plot Throughput / latency for every dim value
reads = []
updates = []
scans = []
rmw = []
for source in sources:
read = Serie(source)
update = Serie(source)
scan = Serie(source)
readModifyWrite = Serie(source)
cur_path = os.path.dirname(__file__)
cur_path = os.path.join(cur_path, source)
for t in throughputs:
# new_path = os.path.join(cur_path, source ,"output-Run-dim" + str(dim) +".txt")
# new_path = os.path.relpath("\\output\\test-mongo-scenarioA\\outputRun-dim" + str(dim) +".txt", cur_path)
filename = commons.getOutputFilename(
cur_path, "Run", workload, dim, t, c)
new_path = os.path.join(cur_path, filename)
f = open(new_path, "r")
for line in f:
if "[READ], AverageLatency(us)," in line:
parts = line.split("[READ], AverageLatency(us), ")
num = float(parts[1])
read.addPoint(t, num)
elif "[UPDATE], AverageLatency(us)," in line:
parts = line.split("[UPDATE], AverageLatency(us),")
num = float(parts[1])
update.addPoint(t, num)
elif "[SCAN], AverageLatency(us)," in line:
parts = line.split("[SCAN], AverageLatency(us),")
num = float(parts[1])
scan.addPoint(t, num)
elif "[READ-MODIFY-WRITE], AverageLatency(us)," in line:
parts = line.split(
"[READ-MODIFY-WRITE], AverageLatency(us),")
num = float(parts[1])
readModifyWrite.addPoint(t, num)
f.close()
reads.append(read)
updates.append(update)
scans.append(scan)
rmw.append(readModifyWrite)
marker = itertools.cycle((',', '+', '.', 'o', '*', 's', 'x'))
for serie in reads:
if ("postgres" in serie.name):
name_db = "Postgres"
elif ("mongo" in serie.name):
name_db = "Mongodb"
else:
name_db = "Voltdb"
if (len(serie.x) > 0):
plt.plot(serie.x,
serie.y,
linewidth=1,
marker=next(marker),
markerfacecolor='blue',
markersize=5,
label="READ - " + name_db)
for serie in updates:
if ("postgres" in serie.name):
name_db = "Postgres"
elif ("mongo" in serie.name):
name_db = "Mongodb"
else:
name_db = "Voltdb"
if (len(serie.x) > 0):
plt.plot(serie.x,
serie.y,
linewidth=1,
marker=next(marker),
markerfacecolor='blue',
markersize=5,
label="UPDATE - " + name_db)
for serie in scans:
if ("postgres" in serie.name):
name_db = "Postgres"
elif ("mongo" in serie.name):
name_db = "Mongodb"
else:
name_db = "Voltdb"
if (len(serie.x) > 0):
plt.plot(serie.x,
serie.y,
linewidth=1,
marker=next(marker),
markerfacecolor='blue',
markersize=5,
label="SCAN - " + name_db)
for serie in rmw:
if ("postgres" in serie.name):
name_db = "Postgres"
elif ("mongo" in serie.name):
name_db = "Mongodb"
else:
name_db = "Voltdb"
if (len(serie.x) > 0):
plt.plot(serie.x,
serie.y,
linewidth=1,
marker=next(marker),
markerfacecolor='blue',
markersize=5,
label="READ-MODIFY-WRITE - " + name_db)
# naming the x axis
plt.xlabel('Throughput')
plt.xscale('log')
# naming the y axis
plt.ylabel('Average latency (us)')
# giving a title to my graph
plt.title(workload +
' Throughput / latency curve with Input dimension=' +
str(dim) + " and " + str(c) + " clients")
plt.legend()
# function to show the plot
plt.show()
def generateAndShowInputLatencyPlots(input_dim, throughputs, clients, source,
workload):
cur_path = os.path.dirname(__file__)
cur_path = os.path.join(cur_path, source)
for c in clients:
for t in throughputs:
# Generate plot Input dim / latency for every throughput value
read = Serie("read")
update = Serie("update")
scan = Serie("scan")
readModifyWrite = Serie("Read-Modify-Write")
for dim in input_dim:
# new_path = os.path.join(cur_path, source ,"output-Run-dim" + str(dim) +".txt")
# new_path = os.path.relpath("\\output\\test-mongo-scenarioA\\outputRun-dim" + str(dim) +".txt", cur_path)
filename = commons.getOutputFilename(cur_path, "Run", workload,
dim, t, c)
new_path = os.path.join(cur_path, filename)
f = open(new_path, "r")
for line in f:
if "[READ], AverageLatency(us)," in line:
parts = line.split("[READ], AverageLatency(us), ")
num = float(parts[1])
read.addPoint(dim, num)
elif "[UPDATE], AverageLatency(us)," in line:
parts = line.split("[UPDATE], AverageLatency(us),")
num = float(parts[1])
update.addPoint(dim, num)
elif "[SCAN], AverageLatency(us)," in line:
parts = line.split("[SCAN], AverageLatency(us),")
num = float(parts[1])
scan.addPoint(dim, num)
elif "[READ-MODIFY-WRITE], AverageLatency(us)," in line:
parts = line.split(
"[READ-MODIFY-WRITE], AverageLatency(us),")
num = float(parts[1])
readModifyWrite.addPoint(dim, num)
f.close()
if (len(read.x) > 0):
plt.plot(read.x,
read.y,
linewidth=1,
marker='o',
markerfacecolor='blue',
markersize=5,
label="READ")
if (len(update.x) > 0):
plt.plot(update.x,
update.y,
linewidth=1,
marker='o',
markerfacecolor='blue',
markersize=5,
label="UPDATE")
if (len(scan.x) > 0):
plt.plot(scan.x,
scan.y,
linewidth=1,
marker='o',
markerfacecolor='blue',
markersize=5,
label="SCAN")
if (len(readModifyWrite.x) > 0):
plt.plot(readModifyWrite.x,
readModifyWrite.y,
linewidth=1,
marker='o',
markerfacecolor='blue',
markersize=5,
label="READ-MODIFY-WRITE")
# naming the x axis
plt.xlabel('Input dimension')
plt.xscale('log')
# naming the y axis
plt.ylabel('Average latency (us)')
# giving a title to my graph
plt.title(workload +
' Input dimension / latency with Throughput=' + str(t) +
" and " + str(c) + " clients")
plt.legend()
# function to show the plot
plt.show()