Size: 859.00 kB
Release date: Mar 10 2020
Platform: Linux,Linux Console,Linux Gnome,Linux GPL,Linux Open Source
Publisher’s Site: https://www.tcpiplabs.com
Publisher’s Product Page: https://www.tcpiplabs.com/Tmarlin_linux_traffic_tool
As a host program, Tmarlin can meet almost all your needs for VM & Docker & Hypervisor & X86 server traffic.
Tmarlin is a pure C program, extremely low consumption and high performance, it can reside in the background to perform continuous run.
Tmarlin can analyze TCP/UDP/HTTP/SQL sessions, output 20 KPIs and 10 latency KQIs for every session in JSON by FIFO, capture packets to PACP by FIFO, or output JSON or Packets to remote path.
Tmarlin can be used independently, and play an important role in latency positioning and traffic traceback after an O & M or security incident.
At the same time, it is also very suitable as a front-end traffic collection and data governance tool for cloud O & M, and cloud security big data analysis solutions.
The main functions of tmarlin include:
1. Real-time analysis of the host’s NIC / IP / IPs / Subnet / IP: port traffic, export 30 kinds of KQI / KPI to JSON for every session;
2. Rich time type KQI: network latency, application reponse time, handshake time, payload transfer time, retransmission time;
3. The KPI/KQI data can be FIFO circularly recorded in the local JSON file, or sent to the user data collector in real time;
4. Record Raw Packets to the local PCAP file in a FIFO loop, support packet size packet cutting function; Raw packets could forward to the remote path with VXLAN protocol in real time;
5. Automatically analyzes VXLAN / GRE protocol traffic;
6. It can be combined with Tdolphin Cloud Traffic Monitoring or user’s big data platform to realize real-time online monitoring of abnormal traffic and performance.
The most important premise above is that when you deploy Tmarlin, you don’t need to worry about t’s resource consumption and impact on other applications, because it only occupies 50MB of memory, and even at 500Mbps, it only occupies 8% resource just in only one vCPU.