What is Network Latency? | Glossary | A10 Networks

The Causes of Network Latency

When a client sends a request to a server across the Internet a complicated series of network transactions are involved. A typical request path will have the client sending the request to a local gateway which in turn routes the request to a sequence of routers, through firewalls and load balancers, and finally to the server. Each step or “hop” involves:

• Receiving of the request
• Decoding the protocol
• Figuring out where the request should go (or even whether to route the request at all)
• Possibly modifying the request to meet protocol or routing requirements
• Sending the request onwards to the next device

All of this takes time, so each hop introduces a delay. Network latency is the total time, usually measured in milliseconds, required for a server and a client to complete a network data exchange. Even if there are no intermediate hops, which is never the case for communications across the Internet, latency is still involved because the request has to traverse layers of software and hardware at each end.

How Latency is Measured

There are two ways to measure network latency:

• Time to First Byte (TTFB) which is how long it takes for a server to receive the first byte of a client request
• Round Trip Time (RTT), the time it takes for the source to send a request and get a reply from the server

If we’re concerned about how network latency affects application performance, then the Round Trip Time is what we care about. On the other hand, if we’re trying to optimize Internet of Things (IoT) transactions we’ll usually be more concerned about Time to First Byte latency.