Forum

Hi
Please clarify me the below question:

Why Fragmentation threshold is 2346 bytes in 802.11 as compared to 1500 bytes in 802.3 Ethernet protocols.

An 802.11 frame has a maximum size of 2346 bytes. So setting the fragmentation threshold to 2346 effectively disables fragmentation. Although generally most wireless frames will contain an IP packet which has a maximum size of 1500 bytes.

One thing that should be remembered is that the IEEE 802.11 protocol is a Layer 2 data link layer protocol, just like PPP for example. It was designed to be "Layer 3 protocol independent". Although IP is the most commonly used layer 3 protocol in Wi-Fi systems, in theory many other layer 3 protocols could be used.

This then brings up the next question......if we look at the frame structure of an 802.11 frame, we do not see any "type" field. So how does the Access Point for example know what is encapsulated inside the frame ?

A SNAP header is commonly used to provide that "type"information. SNAP type fields cover the vast majority of layer 3 protocols. SNAP was developed due to the physical size limitations of the original Ethernet type fields.

If some manufacturer had some weird and wonderful layer 3 protocol that was not covered by the huge list of SNAP types, it would be a fairly straightforward job to introduce a proprietary adjustment to the protocol.

802.11 is really mainly concerned with "flight time" i.e. the physical transmission of frames through the air from say STA to AP or vice versa or STA to STA. One of the reasons for the max length ( and there are a few of them ) was to prevent one station from "hogging the airspace". There are modern techniques of airtime fairness etc, but that was one of the reasons for sizing.

Once we hit the AP, we have Distribution System Services and Integration Services. Perhaps we need to go onto an Ethernet/IP network. The IS will help with frame translation....possible jumbo frames etc etc.

Dave