8 posts by 5 authors in: Forums > CWNA - Enterprise Wi-Fi Admin
Last Post: October 31, 2005:
  • Appreciate some clarification here:
    RTS/CTS is only in 802.11b/g network right?
    If one is using 802.11a, then there's no RTS/CTS involved, right or wrong?

  • Hi Lim:

    IEEE 802.11 RTS/CTS is independent of the 802.11 physical layer and so applies to 802.11a OFDM as well as 802.11b HR/DSSS and 802.11g ERP.

    I hope this helps. Can you add your location to your forum profile? Thanks. /criss

  • Criss,

    RTS/CTS is specified in 802.11-1999, 802.11a supplement specifies an additional PHY, and 802.11g-2003 specifies CTS-to-Self. Since a and g are amendments to 802.11 can I assume that 802.11a networks use CTS-to-Self?



  • Hi Moe:

    This is a wonderful question.

    The IEEE 802.11g amendment to the 802.11 standard introduced a new clause 19 in which is defined the Extended Rate Physical (ERP) PHY.

    The IEEE 802.11g amendment added to clause 9, MAC sublayer functional description, a new sub-clause 9.2.11 in which is defined the protection mechanism CTS-to-self thus.

    9.2.11 NAV distribution -- When a node needs to distribute NAV information, for instance, to reserve the medium for a transmission of a non-basic rate frame (that may not be heard by other nodes in the BSS), the node may first transmit a CTS frame with the RA field equal to its own MAC address (CTS-to-self) and with a duration value that protects the pending transmission, plus possibly an ACK frame. The CTS-to-self NAV distribution mechanism is lower in network overhead cost than is the RTS/CTS NAV distribution mechanism, but CTS-to-self is less robust against hidden nodes and collisions than RTS/CTS. STAs employing a NAV distribution mechanism should choose a mechanism such as CTS-to-self or RTS/CTS that is appropriate for the given network conditions. If errors occur when employing the CTS-to-self mechanism, STAs should switch to a more robust mechanism.

    There is no mention here of ERP, or 802.11a OFDM. There is barely even a reminder to use a low data rate modulation more likely to reach and be decoded by earlier generation Wi-Fi equipment that share the same frequency space.

    I hope this helps. Thanks. /criss

  • Criss,

    Since the 802.11g added to the MAC sublayer functional description and 802.11a added the OFDM physical layer specification and there is no mention of ERP & OFDM I interpret this to mean that 802.11a networks may take advantage of CTS-to-Self.

    Has anyone seen 802.11a networks were CTS-to-Self is used?

  • The point of RTS/CTS in the original standard was for use in cases of hidden nodes. There could be physical barriers or distances between non-AP stations that caused the hidden nodes.

    When 802.11g showed up, we had the same problem, but the issue causing the hidden node was a difference in modulation (e.g. languages spoken by each station). The RTS/CTS protocol could still be used in the situation where ERP (802.11g) STAs needed to let non-ERP (802.11b) stations know that they need to be quiet during an OFDM transmission. This is accomplished through transmission of the RTS and CTS frames using DSSS modulation instead of OFDM.

    CTS-to-Self was introduced for same general purpose, but the with advantage of less overhead and latency. The disadvantage is less coverage around the cell.

    802.11a stations have the issue of hidden nodes when the cause is physical (barriers or distance), but not the cause of differing modulations (OFDM is the only modulation used in the U-NII bands by Wi-Fi gear).

    The question you posed is a good one, and one that I've never specifically looked into (just because it hasn't come up), but I would bet that CTS-to-Self could be used in 802.11a if the mfr implemented it in their chipset. CTS-to-Self functionality is automatically implemented by the radio card as needed and it is built in on a per-mfr basis.


  • Hi Moe:

    I agree with Devin. Whether OFDM (802.11a) stations use CTS-to-self is up to the chip vendors.

    The term "hidden node" appears once, anecdotally, in two of the relatively recent IEEE 802.11 amendments. The term was coined apparently by the IEEE 802.11 Handbook authors in the early days and has become popular although inherently confusing. Some authors show the "hidden node" crouched behind a wall.

    The standard considers no station to be "hidden" from any other. The problem rather is this -- given two stations A and B within range of each other, and a third station C, the third station may be within range of only one of the first two. These stations may be any combination of access point and non access point, and may be from the same basic service set (BSS), two BSSs, or three BSSs. "Within range" means the same frequency and power levels sufficient for carrier sense.

    In this trio there is no one station that is the "hidden" one. For example imagine the above stations A, B, and C in a straight line. When A transmits to B it must cope with being out of range with C. Likewise when C transmits to B it must cope with being out of range with A. But when B transmits to either A or C it is master of its fate.

    I hope this helps. Thanks. /criss

  • In all the 802.11a APs I've seen, RTS-CTS is an option. This includes Linksys, Cisco, Extreme, Proxim and Colubris. Very few client software utilities have RTS-CTS as an option for 802.11a or 802.11b/g networks. I believe Cisco's client utility does and I know Netgear's a/b/g utility does. The Windows, Dell, Funk Odyssey and Apple Airport utilities do not.

    As far as actually enabling it, I've only tried it on a Linksys andCisco network for 802.11a. I'm assuming it would work for the rest.

Page 1 of 1
  • 1