MONITORING NETWORK SWITCHES IN DATA CENTERS
Network availability and performance are critical parameters in determining the proper operation of LAN, MAN or WAN. Malfunctions in network switches adversely affect the productivity of companies, therefore their proactive monitoring is an important element in the work of the administrator. Here we provide a short overview of methods used nowadays for monitoring of network switches.
All types of switches can be monitored using SNMP. Monitoring can provide information within the port: port availability status and information about transmitted packets. In addition, we can monitor equipment performance metrics: CPU usage, RAM usage, etc.
NetFlow, sFlow, jFlow Protocols
NetFlow is a Cisco protocol running on the switches of the company, sFlow protocols and jFlow are similar technologies developed by competitors. These protocols provide information about a stream of data flowing through the network devices, providing detailed insight into the performance and network bandwidth. Because the data is pre-aggregated, the use of this protocol is easier than using a packet sniffer.
Packet sniffing : monitoring using the monitoring port
The outer packet sniffer (usually built into NMS system) examines all network data packets sent through a special monitoring port in a switch. This port in a switch sends a copy of all network packets from different port (or ports) of a switch. Such packets are then analyzed by the NMS system. Out of the three switch monitoring technology, this one creates a highest load on CPU and network.
The central point of the supervised switches environment is IT infrastructure and network monitoring system (NMS). The system aggregates data from the monitored points, provides powerful capabilities for analyzing and visualizing the information collected, and transmits alerts about incidents and failures.
SMS Alerts as an effective notification of failures
A key element in the course of automatic detection of an incident or failure is early as possible and effective notification of the occurrence of the event. For this purpose, data center administrators often use SMS channel. Due to the very good responsiveness to SMS messages (incoming SMS is considered a high priority by the customer compared to other channels, type of e-mail, instant messaging) and versatility (SMS does not require any dedicated application) it is an often used channel for sending alerts about incidents or failures. In order to shorten the critical path (minimize number of devices between NMS server and GSM/3G network), one can use hardware SMS gateway with built-in GSM/3G modem. Such device allows to send SMS alert directly from NMS to the GSM network excluding external Internet Service Provider.