tsworkload_smw2lt

# Copyright (c) 2020 YCSB contributors. All rights reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License"); you
# may not use this file except in compliance with the License. You
# may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
# implied. See the License for the specific language governing
# permissions and limitations under the License. See accompanying
# LICENSE file.

# Yahoo! Cloud System Benchmark
#
# Workload Defintion: Smart Metering Workload 2 (SMW2LT)
#
# This workload was created as part of a bachelor thesis
# at the Technical University of Berlin, by Michael Smart
# https://github.com/smartygus
#
# Data to load as basis:
#
# There are 50000 time series, each with one year of
# data at half-hourly (1800 seconds) interval.
# We are using a fixed starting time stamp representing
# 1/1/2019 at midnight.
# This amounts to a total of:
# 50000 (time series) * 2 (point per hour) * 24 (hours in a day) * 365 (days in 2019) = 876 Million records
#
# Each time series has three tags:
#
# - se_segment: representing some socio-economic category defined by the provider
#   - cardinality = 5 (ie. there are 5 different categories)
# - pricing_strategy: some customers have different pricing strategies
#   - cardinality = 2 (ie. there are 2 different strategies, eg. fixed, and dynamic)
# - household_id = an id representing the household. The ID is unique within the context of the other two tags
#   - cardinality = 5000
#
# Run phase of the workload:
#
# There is just a single SCAN query that makes up the run
# phase of this workload.
# The idea behind this workload is from the perspective of
# an energy provider that would like to know what the average
# hourly usage across its whole customer base is for a given
# day, grouped by pricing strategy. This might for example
# lead to some insights about how a time-of-day-based
# dynamic pricing strategy might influnce energy consumption.
workload=site.ycsb.workloads.TimeSeriesWorkload

recordcount=876000000
operationcount=1000
insertcount=876000000
insertstart=1546300800
# This workload is "offline", so during the run phase
# there are no INSERTs, just SCANs.
# Therefore we do not need to specify
# an inserttransactionstart timestamp

# We only have a single metric, so we are
# dividing up the write load into groups of
# individual time series instead of groups
# of metrics
threadedwritedistribution=timeseries

fieldlength=3
# single metric: 'kwh'
fieldcount=1
# three tags, "real world" names: se_segment, pricing_strategy, household_id
tagcount=3
tagcardinality=5,2,5000

sparsity=0.0
delayedseries=0.0
delayedintervals=0

timestampunits=SECONDS
# one measurement every half hour
timestampinterval=1800
randomtimeseriesorder=false

# Average Hourly Consumption across
# customer base for a given day,
# grouped by pricing_strategy
# No specified start timestamp, a random
# start time will be chosen.
# Query timespan is 24 hours
scanproportion=1.0
scantags=-1,-1,-1
scanquerytimespan=86400
scanqueryrandomtimespan=false
scangroupbyfunction=AVERAGE
scangroupbytagkeys=0,1,0
scandownsamplingfunction=SUM
scandownsamplinginterval=3600

readproportion=0.00
insertproportion=0.00
updateproportion=0.00