『nitrogen flow meter micro flow|flow meter distributor』Related information(clamp on meter|electromagnetic meter|venturi meterrotameter|orifice meter|ultrasonic flow meter|mass flow meter|coriolis mass flow meter|coriolis flow meter|magnetic flow meter|magmeter flow meter|magflow flow meter|mag meter flow meter|electromagnetic flow meter|vortex flow meter|turbine flow meter|thermal mass flow meter|thermal flow meter|rotameter flow meter)
1. DeepFlow Data Flow
DeepFlow Data Flow parsing DeepFlow data flow is a tedious and detailed process that involves multiple modules and steps to ensure efficient processing and analysis of network traffic. The following is a detailed analysis of the DeepFlow data flow process:
1. Data capture and injection Data capture: The data is first captured by the recw_engine in the dispatcher module of the network. Data injection: Dispatcher calls the inject_ceta_cacket() method of flow_map to inject the captured data into the flow generator module (mainly flow_map).
Secondly, the Flow Generator module processes and generates relevant data: The Flow Generator module is responsible for generating relevant tagged_flow data, L7 performance data (l7 perf data), and application log data (such as httplog, mysqlog, etc.). Data transmission: These data are then sent to the collect module for further processing. 3. Tagged Flow Data Processing Scenario: When a new flow_node is updated or an expired flow_node is deleted, the generation and transmission of tagged flow data will be triggered. Path: The data flows to QuadrupleGeneratiorThread through the output queue of the flow_map class (carried in by the dispatchers flow_dender parameter). In QuadrupleGeneratiorThread, data is further processed and a second level flow meter is generated. Aggregation and transmission: The second level flow meters are then aggregated into minute level meters. These data are sent to the DeepFlow server data plane through four threads (quadle_generationor, l4_flow-aggr, second_completor, minte_completor). 4. L7 performance data processing scenario: When adding or updating TCP/UDP flow_node, L7 performance data will be sent. Path: The dat
Fifth, Application Log Processing Scenarios: When adding and updating TCP/UDP flow_node and flow_close, as well as when the inject_flush-ticker method sends a fixed tail and EBPF packet timeout, application log data will be generated and sent.
. Path: The data flows to the SessionAggregator through the out_Log_queue of the flow_map class (which is passed into the log_Snder by the dispatchers log_output_queue parameter). Aggregation and Sending: The SessionAggregator is responsible for aggregating HTTP and DNS requests and responses to form a session. In the end, these session data are sent to the server data plane through proto_1og_stender. 6、 Flow recognition and aggregation process: Flow recognition mainly converts packages into flownodes for subsequent processing. Stream aggregation: Using the update_tcp/udp_node() method, each packet is processed according to the state machine to make the flownode enter the corresponding state. For TCP flows, more detailed state tracking and statistics will also be conducted. 7、 At the end of the entire processing flow, a stream log will be generated and sent for subsequent analysis and auditing. The processing flow of DeepFlow data flow is an external to internal process that involves multiple modules and steps. FlowMap manages the entire flow table, finds or creates FlowNodes; FlowNode contains TaggedFlow and other
Hot Blog
Related AD
