refactor(tools): nest cloudlab/local config under a providers section in config.yaml (#469)

Author: milindsrivastava1997

asap-tools/experiments/CONFIG_PARAMETERS_REFERENCE.md

@@ -30,20 +30,20 @@ These parameters must be provided for all experiment scripts:
 - **Validation**: Cannot be `???` or empty
 - **Used by**: All experiment scripts
 
-#### `cloudlab.num_nodes` (int, required)
+#### `providers.cloudlab.num_nodes` (int, required)
 - **Description**: Number of CloudLab nodes to allocate for the experiment
 - **Range**: 1-50 (recommended)
 - **Example**: `4`
 - **Validation**: Must be positive integer
 - **Used by**: All experiment scripts
 
-#### `cloudlab.username` (string, required)
+#### `providers.cloudlab.username` (string, required)
 - **Description**: Your CloudLab username for SSH access
 - **Example**: `"myuser"`
 - **Validation**: Must be valid CloudLab username
 - **Used by**: All experiment scripts
 
-#### `cloudlab.hostname_suffix` (string, required)
+#### `providers.cloudlab.hostname_suffix` (string, required)
 - **Description**: CloudLab experiment hostname suffix
 - **Example**: `"myexp.cloudlab.us"`
 - **Validation**: Must be valid hostname format
@@ -612,7 +612,7 @@ These parameters come from the `experiment_type` config group and are prefixed w
 ```python
 # Recommended validation rules
 VALIDATION_RULES = {
-    "cloudlab.num_nodes": {"type": int, "min": 1, "max": 50},
+    "providers.cloudlab.num_nodes": {"type": int, "min": 1, "max": 50},
     "prometheus.scrape_interval": {"type": str, "pattern": r"^\d+[smh]$"},
     "experiment_params.exporters.exporter_list.fake_exporter.start_port": {
         "type": int, "min": 1024, "max": 65535
@@ -685,9 +685,9 @@ VALIDATION_RULES = {
 python experiment_run_e2e.py \
   experiment.name=my_test \
   experiment_type=cloud_demo \
-  cloudlab.num_nodes=4 \
-  cloudlab.username=myuser \
-  cloudlab.hostname_suffix=myexp.cloudlab.us \
+  providers.cloudlab.num_nodes=4 \
+  providers.cloudlab.username=myuser \
+  providers.cloudlab.hostname_suffix=myexp.cloudlab.us \
   prometheus.local_config_dir=/path/to/prometheus/config
 ```
 
@@ -719,7 +719,7 @@ python experiment_run_e2e.py \
 ```bash
 python experiment_run_e2e.py \
   experiment_type=cloud_demo \
-  cloudlab.num_nodes=2 \
+  providers.cloudlab.num_nodes=2 \
   logging.level=DEBUG \
   flow.steady_state_wait=60 \
   streaming.do_local_flink=true \
@@ -742,7 +742,7 @@ for nodes in 4 8 16; do
   python experiment_run_e2e.py \
     experiment.name=nodes_${nodes} \
     experiment_type=cloud_demo \
-    cloudlab.num_nodes=${nodes} \
+    providers.cloudlab.num_nodes=${nodes} \
     [required params...]
 done
 ```

asap-tools/experiments/HYDRA_CONFIG_USAGE.md

@@ -9,9 +9,9 @@ This guide explains how to use the Hydra framework integration for experiment sc
 python experiment_run_e2e.py \
   experiment.name=my_test \
   experiment_type=cloud_demo \
-  cloudlab.num_nodes=4 \
-  cloudlab.username=myuser \
-  cloudlab.hostname_suffix=myexp.cloudlab.us \
+  providers.cloudlab.num_nodes=4 \
+  providers.cloudlab.username=myuser \
+  providers.cloudlab.hostname_suffix=myexp.cloudlab.us \
   prometheus.local_config_dir=/path/to/prometheus/config
 ```
 
@@ -20,9 +20,9 @@ python experiment_run_e2e.py \
 ### Required Infrastructure Parameters
 All experiment scripts require these core parameters:
 - `experiment.name`: Human-readable experiment name
-- `cloudlab.num_nodes`: Number of CloudLab nodes
-- `cloudlab.username`: Your CloudLab username
-- `cloudlab.hostname_suffix`: CloudLab experiment hostname suffix
+- `providers.cloudlab.num_nodes`: Number of CloudLab nodes
+- `providers.cloudlab.username`: Your CloudLab username
+- `providers.cloudlab.hostname_suffix`: CloudLab experiment hostname suffix
 
 ### Script-Specific Required Parameters
 - **experiment_run_clickhouse.py**: `experiment_type=clickhouse`, `experiment_params.dataset.name`, `experiment_params.dataset.local_data_file`, `experiment_params.query_groups[0].sql_file`
@@ -107,9 +107,9 @@ python experiment_run_e2e.py \
 python experiment_run_e2e.py \
   experiment.name=dev_test \
   experiment_type=cloud_demo \
-  cloudlab.num_nodes=2 \
-  cloudlab.username=myuser \
-  cloudlab.hostname_suffix=dev.cloudlab.us \
+  providers.cloudlab.num_nodes=2 \
+  providers.cloudlab.username=myuser \
+  providers.cloudlab.hostname_suffix=dev.cloudlab.us \
   prometheus.local_config_dir=/path/to/prometheus/config \
   logging.level=DEBUG \
   flow.steady_state_wait=60
@@ -121,7 +121,7 @@ python experiment_run_e2e.py \
 python experiment_run_e2e.py \
   experiment.name=arroyo_perf \
   experiment_type=my_exp_config \
-  cloudlab.num_nodes=8 \
+  providers.cloudlab.num_nodes=8 \
   streaming.engine=arroyo \
   streaming.enable_object_reuse=true \
   profiling.arroyo=true \
@@ -136,9 +136,9 @@ python experiment_run_e2e.py experiment.name=delay_60 experiment_type=cloud_demo
 python experiment_run_e2e.py experiment.name=delay_120 experiment_type=cloud_demo experiment_params.query_groups.0.repetition_delay=120 [required params...]
 
 # Test different node counts
-python experiment_run_e2e.py experiment.name=nodes_4 experiment_type=cloud_demo cloudlab.num_nodes=4 [required params...]
-python experiment_run_e2e.py experiment.name=nodes_8 experiment_type=cloud_demo cloudlab.num_nodes=8 [required params...]
-python experiment_run_e2e.py experiment.name=nodes_16 experiment_type=cloud_demo cloudlab.num_nodes=16 [required params...]
+python experiment_run_e2e.py experiment.name=nodes_4 experiment_type=cloud_demo providers.cloudlab.num_nodes=4 [required params...]
+python experiment_run_e2e.py experiment.name=nodes_8 experiment_type=cloud_demo providers.cloudlab.num_nodes=8 [required params...]
+python experiment_run_e2e.py experiment.name=nodes_16 experiment_type=cloud_demo providers.cloudlab.num_nodes=16 [required params...]
 
 # Test different streaming engines
 python experiment_run_e2e.py experiment.name=flink_test experiment_type=cloud_demo streaming.engine=flink [required params...]
@@ -165,9 +165,9 @@ cat > config/experiment_type/clickhouse.yaml << 'EOF'
 #   python experiment_run_clickhouse.py \
 #     experiment_type=clickhouse \
 #     experiment.name=my_test \
-#     cloudlab.num_nodes=1 \
-#     cloudlab.username=myuser \
-#     cloudlab.hostname_suffix=myexp.cloudlab.us \
+#     providers.cloudlab.num_nodes=1 \
+#     providers.cloudlab.username=myuser \
+#     providers.cloudlab.hostname_suffix=myexp.cloudlab.us \
 #     experiment_params.dataset.name=clickbench \
 #     experiment_params.dataset.local_data_file=/path/to/hits.json \
 #     'experiment_params.query_groups[0].sql_file=/path/to/queries.sql'
@@ -203,9 +203,9 @@ EOF
 python experiment_run_clickhouse.py \
   experiment_type=clickhouse \
   experiment.name=my_clickbench_run \
-  cloudlab.num_nodes=1 \
-  cloudlab.username=myuser \
-  cloudlab.hostname_suffix=myexp.cloudlab.us \
+  providers.cloudlab.num_nodes=1 \
+  providers.cloudlab.username=myuser \
+  providers.cloudlab.hostname_suffix=myexp.cloudlab.us \
   experiment_params.dataset.name=clickbench \
   experiment_params.dataset.local_data_file=/path/to/hits.json \
   'experiment_params.query_groups[0].sql_file=/path/to/queries.sql'
@@ -231,9 +231,9 @@ experiment_params.dataset.max_rows=10000
 python experiment_run_e2e.py \
   experiment.name=full_test \
   experiment_type=cloud_demo \
-  cloudlab.num_nodes=4 \
-  cloudlab.username=myuser \
-  cloudlab.hostname_suffix=myexp.cloudlab.us \
+  providers.cloudlab.num_nodes=4 \
+  providers.cloudlab.username=myuser \
+  providers.cloudlab.hostname_suffix=myexp.cloudlab.us \
   prometheus.local_config_dir=/path/to/prometheus/config
 ```
 
@@ -242,9 +242,9 @@ python experiment_run_e2e.py \
 # Only exporters and Prometheus - no streaming
 python experiment_run_exporters_and_prometheus.py \
   experiment.name=monitoring_test \
-  cloudlab.num_nodes=4 \
-  cloudlab.username=myuser \
-  cloudlab.hostname_suffix=myexp.cloudlab.us
+  providers.cloudlab.num_nodes=4 \
+  providers.cloudlab.username=myuser \
+  providers.cloudlab.hostname_suffix=myexp.cloudlab.us
 ```
 
 ### experiment_run_empty_flink.py (Simplified Streaming)
@@ -253,9 +253,9 @@ python experiment_run_exporters_and_prometheus.py \
 python experiment_run_empty_flink.py \
   experiment.name=empty_flink_test \
   experiment.config_file=/path/to/config.yml \
-  cloudlab.num_nodes=4 \
-  cloudlab.username=myuser \
-  cloudlab.hostname_suffix=myexp.cloudlab.us
+  providers.cloudlab.num_nodes=4 \
+  providers.cloudlab.username=myuser \
+  providers.cloudlab.hostname_suffix=myexp.cloudlab.us
 ```
 
 ### experiment_run_e2e_no_queryengine.py (E2E without Query Engine)
@@ -264,9 +264,9 @@ python experiment_run_empty_flink.py \
 python experiment_run_e2e_no_queryengine.py \
   experiment.name=e2e_no_qe \
   experiment.config_file=/path/to/config.yml \
-  cloudlab.num_nodes=4 \
-  cloudlab.username=myuser \
-  cloudlab.hostname_suffix=myexp.cloudlab.us \
+  providers.cloudlab.num_nodes=4 \
+  providers.cloudlab.username=myuser \
+  providers.cloudlab.hostname_suffix=myexp.cloudlab.us \
   prometheus.local_config_dir=/path/to/prometheus/config
 ```
 
@@ -284,8 +284,8 @@ python experiment_run_sketchdboffline.py \
 # Flink performance testing with aggregation scaling
 python experiment_run_flink_with_different_num_aggregations.py \
   experiment.name=flink_perf \
-  cloudlab.username=myuser \
-  cloudlab.hostname_suffix=myexp.cloudlab.us \
+  providers.cloudlab.username=myuser \
+  providers.cloudlab.hostname_suffix=myexp.cloudlab.us \
   experiment_variants.flink_aggregations.config=/path/to/config.yaml \
   experiment_variants.flink_aggregations.aggregation_id=0 \
   experiment_variants.flink_aggregations.min_aggregations=1 \
@@ -334,9 +334,9 @@ python experiment_run_e2e.py \
 python experiment_run_e2e.py \
   experiment.name=test \
   experiment_type=cloud_demo \
-  cloudlab.num_nodes=4 \
-  cloudlab.username=user \
-  cloudlab.hostname_suffix=exp.cloudlab.us \
+  providers.cloudlab.num_nodes=4 \
+  providers.cloudlab.username=user \
+  providers.cloudlab.hostname_suffix=exp.cloudlab.us \
   prometheus.local_config_dir=/path/to/config
 ```
 

asap-tools/experiments/config/config.yaml

@@ -5,12 +5,17 @@
 experiment:
   name: ???  # REQUIRED: Human-readable experiment name
 
-# CloudLab configuration
-cloudlab:
-  num_nodes: ???  # REQUIRED: Number of CloudLab nodes to use
-  node_offset: 0  # Starting node index (default 0, allows running multiple experiments in parallel)
-  username: ???  # REQUIRED: Your CloudLab username
-  hostname_suffix: ???  # REQUIRED: CloudLab experiment hostname suffix
+# Infrastructure provider configuration — activate exactly one provider:
+#   CloudLab (default): fill in the cloudlab fields below via CLI overrides.
+#   Local machine: uncomment the local section below (leave cloudlab as-is).
+providers:
+  cloudlab:
+    num_nodes: ???  # REQUIRED for CloudLab: Number of CloudLab nodes to use
+    node_offset: 0  # Starting node index (default 0, allows running multiple experiments in parallel)
+    username: ???  # REQUIRED for CloudLab: Your CloudLab username
+    hostname_suffix: ???  # REQUIRED for CloudLab: CloudLab experiment hostname suffix
+  # local:  # Uncomment this section to run on the local machine instead
+  #   home_dir: ???  # REQUIRED: absolute path to home directory on this machine
 
 # Logging and debugging
 logging:
@@ -57,7 +62,7 @@ streaming:
   use_kafka_ingest: false  # Use Prometheus remote write by default, set to true for legacy Kafka ingest
   # Prometheus remote write configuration
   remote_write:
-    ip: "${remote_write_ip:${cloudlab.node_offset}}"
+    ip: null  # set at runtime via provider.get_node_ip(node_offset)
     base_port: 8080
     path: "/api/v1/write"
     enable_optimized_source: true # Use optimized Prometheus remote_write source (10-20x faster)

asap-tools/experiments/experiment_only_ingest_path.py

@@ -30,11 +30,6 @@
     "local_experiment_dir", lambda: constants.LOCAL_EXPERIMENT_DIR
 )
 
-# Register custom resolver for remote write IP based on node_offset
-OmegaConf.register_new_resolver(
-    "remote_write_ip", lambda node_offset: f"10.10.1.{node_offset + 1}"
-)
-
 KAFKA_NUM_TRIES = 5
 CONTROLLER_LOCAL_OUTPUT_DIR = None
 CONTROLLER_REMOTE_OUTPUT_DIR = None
@@ -83,6 +78,7 @@ def main(cfg: DictConfig):
 
     # Create infrastructure provider
     provider = create_provider(cfg)
+    args.remote_write_ip = provider.get_node_ip(args.node_offset)
 
     local_experiment_root_dir = os.path.join(
         constants.LOCAL_EXPERIMENT_DIR, args.experiment_name

asap-tools/experiments/experiment_run_clickhouse.py

@@ -69,19 +69,19 @@
   python experiment_run_clickhouse.py \\
     experiment_type=clickhouse \\
     experiment.name=my_bench \\
-    cloudlab.num_nodes=1 \\
-    cloudlab.username=myuser \\
-    cloudlab.hostname_suffix=myexp.cloudlab.us \\
+    providers.cloudlab.num_nodes=1 \\
+    providers.cloudlab.username=myuser \\
+    providers.cloudlab.hostname_suffix=myexp.cloudlab.us \\
     experiment_params.dataset.local_data_file=/path/to/hits.json \\
     'experiment_params.query_groups[0].sql_file=/path/to/clickbench.sql'
 
 Baseline + sketchdb:
   python experiment_run_clickhouse.py \\
     experiment_type=clickhouse \\
     experiment.name=my_bench \\
-    cloudlab.num_nodes=1 \\
-    cloudlab.username=myuser \\
-    cloudlab.hostname_suffix=myexp.cloudlab.us \\
+    providers.cloudlab.num_nodes=1 \\
+    providers.cloudlab.username=myuser \\
+    providers.cloudlab.hostname_suffix=myexp.cloudlab.us \\
     experiment_params.dataset.local_data_file=/path/to/hits.json \\
     'experiment_params.query_groups[0].sql_file=/path/to/clickbench.sql'
 
@@ -148,9 +148,6 @@ def _expand_groups(groups):
 OmegaConf.register_new_resolver(
     "local_experiment_dir", lambda: constants.LOCAL_EXPERIMENT_DIR
 )
-OmegaConf.register_new_resolver(
-    "remote_write_ip", lambda node_offset: f"10.10.1.{node_offset + 1}"
-)
 
 
 def _run_query_client(

asap-tools/experiments/experiment_run_e2e.py

@@ -47,11 +47,6 @@
     "local_experiment_dir", lambda: constants.LOCAL_EXPERIMENT_DIR
 )
 
-# Register custom resolver for remote write IP based on node_offset
-OmegaConf.register_new_resolver(
-    "remote_write_ip", lambda node_offset: f"10.10.1.{node_offset + 1}"
-)
-
 
 @hydra.main(version_base=None, config_path="config", config_name="config")
 def main(cfg: DictConfig):
@@ -63,15 +58,9 @@ def main(cfg: DictConfig):
     # Create infrastructure provider
     provider = create_provider(cfg)
 
-    # Route the remote-write IP through the provider (127.0.0.1 for local mode,
-    # CloudLab's 10.10.1.x scheme otherwise) before anything resolves
-    # cfg.streaming.remote_write.ip (Args does, below).
-    OmegaConf.register_new_resolver(
-        "remote_write_ip", provider.get_node_ip, replace=True
-    )
-
     # Convert config to args-like object for backward compatibility
     args = config.Args(cfg)
+    args.remote_write_ip = provider.get_node_ip(args.node_offset)
 
     if provider.is_remote():
         local_experiment_root_dir = os.path.join(

asap-tools/experiments/experiment_run_exporters_and_prometheus.py

@@ -22,11 +22,6 @@
     "local_experiment_dir", lambda: constants.LOCAL_EXPERIMENT_DIR
 )
 
-# Register custom resolver for remote write IP based on node_offset
-OmegaConf.register_new_resolver(
-    "remote_write_ip", lambda node_offset: f"10.10.1.{node_offset + 1}"
-)
-
 
 @hydra.main(version_base=None, config_path="config", config_name="config")
 def main(cfg: DictConfig):

asap-tools/experiments/experiment_run_grafana_demo.py

@@ -43,11 +43,6 @@
     "local_experiment_dir", lambda: constants.LOCAL_EXPERIMENT_DIR
 )
 
-# Register custom resolver for remote write IP based on node_offset
-OmegaConf.register_new_resolver(
-    "remote_write_ip", lambda node_offset: f"10.10.1.{node_offset + 1}"
-)
-
 
 @hydra.main(version_base=None, config_path="config", config_name="config")
 def main(cfg: DictConfig):
@@ -60,6 +55,7 @@ def main(cfg: DictConfig):
 
     # Create infrastructure provider
     provider = create_provider(cfg)
+    args.remote_write_ip = provider.get_node_ip(args.node_offset)
 
     args.forward_unsupported_queries = True
     print("Forcing forward_unsupported_queries to True for Grafana demo")

asap-tools/experiments/experiment_run_sketchdboffline.py

@@ -17,11 +17,6 @@
     "local_experiment_dir", lambda: constants.LOCAL_EXPERIMENT_DIR
 )
 
-# Register custom resolver for remote write IP based on node_offset
-OmegaConf.register_new_resolver(
-    "remote_write_ip", lambda node_offset: f"10.10.1.{node_offset + 1}"
-)
-
 
 def run_sketchdboffline(args, output_dir) -> subprocess.Popen:
     cmd = (