Overview

Internet content operation (integrated management system of data center) supports the analysis and scheduling of resource introduction by taking the analysis of popular key resources, normalization analysis of poor quality resources and normalization analysis of competitive products as objects, cache/CDN, scheduling and introduction as methods. According to the exploration of Internet resource scheduling analysis methods and test results, with the goal of supporting traffic management, promote the optimization of Internet resource scheduling of operators, focus on hot business, key business, poor quality business and large traffic business, promote independent cache operation, gradually establish an introduction system, and form a closed-loop mechanism.

Function

IDC global real-time monitoring: presenting the overall customer, service province and internal network, service external network, total return source, internal return source, external return source and other traffic and changes.

IDC customer security behavior detection: Provides a refined management function for IDC customers, analyzes the nature of ICP customers, monitors their compliance and bandwidth audit of operational business based on website content, open ports, traffic, click through volume, etc., and provides insight into distribution content and abnormal resource blocking control.

IDC content refinement operation: Analyze the introduction of high-value customer resources by IDC service providers, optimize content introduction and resource scheduling strategies; For key ICP guarantees, present the network traffic distribution of each ICP customer and the overall network composition of ICP customers.

IDC Quality Analysis Traceability: Analyze the quality situation of IDC data centers, links, user locations, user types, and other dimensions, such as the latency, retransmission rate, and broadband situation of data center customers. Establish an evaluation positioning optimization cycle mechanism to promote quality improvement and optimization.

IDC intelligent operation and maintenance: provide multi-level, multi-dimensional, and multi method alarm monitoring for servers, switches, firewalls, DPI probes, and big data clusters within IDC from multiple perspectives and angles; And provide intelligent batch jobs that can be viewed, traced, and recorded for all execution processes.

Characteristics

A quality analysis system based on unidirectional messages: A quality analysis system based on unidirectional messages has been established, avoiding the incomplete situation of only extracting quality from bidirectional messages and achieving comprehensive measurement of network quality.

A data quality analysis standard for DPI system has been developed. By monitoring the input and output statistics of multiple modules, comparing data changes horizontally, monitoring and analyzing data quality, and generating alarms in case of abnormalities.

Based on massive big data processing and machine learning algorithms: relying on the StreamCloud big data platform, providing users with PB level storage capacity through cluster storage systems, timely and on-demand online dynamic expansion, and achieving fast access to files; Capable of mining video and log files based on machine learning algorithms, effectively utilizing the computing resources of storage devices.

Millisecond level blocking and control capability: supports docking with probe devices with multiple bandwidths such as 10G and 100G, using DPDK technology to quickly receive and parse messages; Quick matching mode, quickly and accurately matching sealing strategies; Block the rapid transmission of messages. The delay is in milliseconds, effectively blocking and ensuring a blocking rate of over 99.9%.

Architecture