How to Run

1. Create the workloads directory using

   mkdir workloads/

2. Create build directory and switch directories

   mkdir build/
   cd build/

3. Compile using CMAKE

   cmake ..
   make

4. Input workload description through args.toml
5. Run the workload generator using the following format:

   ./sortedness_data_generator -P <total partitions> -F <relative workload description path>

For example, a sample ingestion workload to create 1M entries with K=L=10 (10% of 1M entries) will look like:

./sortedness_data_generator -F ../src/args.toml

Here, we used "1234" as a seed value, and alpha=1 while beta=1 for the sortedness distribution. We place this created workload in the "workloads/" directory, and use a payload size of 252 Bytes. Note, the data generator requires the name of the output file as input. By default, the benchmark uses the following format for nomenclature:

createdata_N<num_entries>_K<k%>_L<l%>_S<seed_val>_a<alpha_val>_b<beta_val>_P<payload_size>.txt

The following workload description is used to input the previous values.

title = "Single Workload Description"

[global]
domain = 10000000

[[partition]]
start_index = 0
number_of_entries = 1000000
K = 10
L = 10
seed = 1234
alpha = 1
beta = 1
payload = 252
window_size = 1
output_file = "../workloads/createdata_N1000000_K10_L10_S1234_a1_b1_P252.txt"
is_fixed = false
is_binary = false
reverse_order = true

This will by default generate a comma separated txt file with the key in the first column and a randomly generated payload (string) in the second column.

To run the data generator with mutiple workload descriptions, follow the format shown below in the sample completed workload description:

title = "Multiple Workload Descriptions"

[global]
domain = 50000

[[partition]]
start_index = 0
number_of_entries = 2000
K = 10
L = 10
seed = 1
alpha = 1
beta = 1
payload = 0
window_size = 1
output_file = "../workloads/createdata.txt"
is_fixed = false
is_binary = false
reverse_order = true

[[partition]]
start_index = 2000
number_of_entries = 3000
K = 10
L = 20
seed = 1
alpha = 1
beta = 1
payload = 0
window_size = 5
output_file = "../workloads/createdata.txt"
is_fixed = true
is_binary = false
reverse_order = false

[[partition]]
start_index = 5000
number_of_entries = 5000
K = 50
L = 10
seed = 1
alpha = 1
beta = 1
payload = 0
window_size = 10
output_file = "../workloads/createdata.txt"
is_fixed = false
is_binary = false
reverse_order = false

To run with these multiple partions, use the following script:

./sortedness_data_generator -P 3 -F ../src/args.toml