Does WiFi introduce latency?

[FightingChance]

Simply, do the latest WiFi standards (b and/or g) introduce an extra step of latency versus wired connections?

i.e. If I pinged a game server at 40, would WiFi be the same? Would it be 45, 90? Would it be variable?

[Gorganzola]

The latency in WiFi is directly proportional to the distance between the transmitter and receiver, and the obstructions between the transmitter and receiver.

The latency will be variable only if the transmitter and/or receiver are moving, if there are interfering signals from another source and if the obstructions vary between the transmitter and receiver.

The fact of the matter is that if you are playing a game where any delay could be death, WiFi isn't the what you should be using.

Yes, copper incurs latencies directly proportional to the length used, and from external interference but these latencies are dramatically less then what WiFi incurs. Plus, electrons travel remarkably faster then radio waves.

[SupDawg]

Originally posted by FightingChance
Simply, do the latest WiFi standards (b and/or g) introduce an extra step of latency versus wired connections?

i.e. If I pinged a game server at 40, would WiFi be the same? Would it be 45, 90? Would it be variable?

I would have to agree with Gorgo,

If ping is the thing, then you don't want wireless.

With wireless, it specifies that each packet sent to the wap must have an acknowledgment back. Wifi actually has a nasty habit of packet collisions. So, yes it will be slower. Slower because of radio vs electric, and slower because each packet must have a receipt of delivery to ensure error free networking.

[SKI]

Actual latencies are reported at around 3-6 ms/hop for a WiFi link, given a high signal/noise ratio and low utilization. Collisions from high traffic, Low s/n ratios, etc will increase retransmissions and latency. Many PCs on a WLAN would have difficulty gaming together or over the Internet, for obvious reasons. WiFi is a shared media. It behaves similarly to shared Ethernet, where speed/performance is directly related to the number of hosts attached. I've been thinking about this a bit more, and decided that a related issue for gamers is the number of PCs gaming online over the same WAN link. Since this is a single full-duplex path, adding users, say several PCs gaming on the same server at another site over DSL or Cable, would reduce performance and increase latency for all users, especially on the outbound (upload) side with its limited bandwidth. Packetloss brought up another related issue. XP's Zero Configuration service for wireless further complicates the latency issue for applications sensitive to delay. See: http://www.abxzone.com/forums/showth...957#post496957

I'd differ with Gorgo slightly, and say that the speed at wich wired and wireless signals travel is less an issue than is signal propogation and translation. Transmitter and receiver ops, link handshaking/overhead, antenna physics, and conversion from Ethernet to radio back to Ethernet are going produce the vast majority of latency. At 50 ft the difference in signal speed through media (radio/copper) is no more than 2.6X10^-5 milliseconds... not very much overall, at the distances we're talking about.

The bottom line is that to minimize latency you want to look for a router and modem that tests fast, and use wires. Possibly one would be even better off with a modem directly connected to NIC (& S/W firewall, of course, for security).... you could possibly pick up another few milliseconds.