Whether network performance is good is based on if users are satisfied or not; however, it is hard for service administrators to understand users' actual network experience by only monitoring device status like CPU or bandwidth utilization, etc. with operation tools. It is common that there is nothing abnormal about devices, but users keep complaining about the efficiency. It is because users and operation tools are in different location; when users connect to a website, they need to pass many route points, and if any of them makes the connection unstable, the latency will lead to negative perception.
N-Partner currently has two ways to monitor network latency. The first one is sending ICMP ping packets to collect round trip time (RTT) information of each network node, and the other is imitating people’s behavior of visiting the monitored web services and recording responding time to do analysis.
Users’ devices go through several network nodes, including L2/L3 switches, core switches, firewalls/security gateways, and so on, while accessing a service. As a result, the best practice of collecting round trip time (RTT) data is to send ping packets from users’ devices to each node and record their latency in order to find out which node has the longest one.
Web service monitoring monitors the performance of five stages in web browsing: DNS Query and Responseà TCP Connectionà SSLà Responseà The First Page Download
N-Reporter/N-Cloud of N-Partner can send ICMP ping packets from their management IP to collect RTT data; also, they can send access requests to web services and record the latency of every stage. To know users’ actual network experience, administrators can deploy N-Probe in different places, such as in OA network, offices of branches, wireless LAN, telecommunications facilities of ISP, etc. By monitoring the connection status of different nodes, administrators can have an instant, comprehensive understanding about web service performance. N-Probe can monitor both RTT with ICMP ping and web service performance.
N-Reporter/N-Cloud keep receiving RTT and web service performance information, do analysis, and make reports. There are two kinds of alert threshold. One is set manually, and when the threshold is exceeded, alerts will be sent; the other is dynamic threshold the system set automatically with AI learning, and alerts will be sent when latency abnormally surges. What is more, with prediction function, N-Reporter/N-Cloud can predict the possible ICMP ping RTT in the next few days to several months, sending notification when the performance is getting worse.
