Symptom: The application appears sluggish. It may freeze on a certain screen or halt while accessing data. Often, the network is blamed for these issues.
Cause: Exonerating the network in application performance problems can assist server maintenance personnel to take out the guesswork and isolate the issue to the right place. Many issues can cause an application to slow down. Among the most common causes are server backups occurring during production hours, slow response from database servers, and packet loss on the network. From a network technician’s point of view, the most important thing to determine is whether the problem is caused by the server or by the network. To determine this, a capture of application traffic can be collected from a client machine. Look for any retransmissions between the client and server. If retransmissions exist, there is packet loss on the network, which severely affects application performance. If there are no retransmissions and connectivity between client and server is established, the problem is likely in the server and can be monitored from that perspective.
Resolution: Although packet analyzers can be very difficult to use when tracking down a problem, they often are equipped with simple counters displaying TCP retransmissions. Use this counter to assist in determining if there is packet loss on the network between client and server. Look for Ethernet errors (FCS Errors, Alignment Errors, or Late Collisions) on any switches and routers between client and server that could be causing this packet loss. If there are no errors, watch for packet loss on the WAN because of excessive utilization across the link.