US 12,445,361 B2
Inference-oriented method and system for application flow and index vectorization
Fenglin Zhi, Shanghai (CN); and Xiaohua Cai, Shanghai (CN)
Assigned to SHANGHAI NETIS TECHNOLOGIES CO., LTD, Shanghai (CN)
Appl. No. 17/926,622
Filed by SHANGHAI NETIS TECHNOLOGIES CO., LTD, Shanghai (CN)
PCT Filed Dec. 17, 2021, PCT No. PCT/CN2021/138960
§ 371(c)(1), (2) Date Nov. 20, 2022,
PCT Pub. No. WO2022/206040, PCT Pub. Date Oct. 6, 2022.
Claims priority of application No. 202110352102.X (CN), filed on Mar. 31, 2021.
Prior Publication US 2023/0186122 A1, Jun. 15, 2023
Int. Cl. H04L 43/04 (2022.01)
CPC H04L 43/04 (2013.01) 10 Claims
OG exemplary drawing
 
1. An inference-oriented method for an application flow and an index vectorization, comprising:
Step S1: quantizing an original index information of related devices with known network flows and current flows in a distributed system into first tuples of a first same type to obtain known network flow information tuples and index information tuples;
Step S2: quantizing unknown network flows in the distributed system into second tuples of a second same type to obtain unknown network flow information tuples;
Step S3: forming a network according to a time window and a priori association relationship based on the known network flow information tuples, the index information tuples, and the unknown network flow information tuples, and sampling a network description information by a random walk;
Step S4: obtaining network flow vectors and index information vectors by an embedding algorithm based on the network description information; and
Step S5: performing a root cause analysis and an index prediction according to the network flow vectors and the index information vectors, wherein
each of the known network flow information tuples, the index information tuples, and the unknown network flow information tuples comprises a code and a time, wherein the code is a first positive integer id, and the time is a flow starting time point or an index sampling time point.