Change log from v03 (June 29) to v04 (July 28):

In the rough order of significance:

1) Section 1.3 "Terminology Used in this Document"
For "Split MAC Architecture": 
[old]¡°¡­implement the delay sensitive MAC services (like the control frame processing) for IEEE 802.11, while tunneling all the management and data frames to AC for centralized processing.¡±

[new]¡°¡­implement the delay sensitive MAC services (including all control frames and some management frames) for IEEE 802.11, while tunneling all the remaining management and data frames to AC for centralized processing.¡±

2) Section 3.2 "Security" (note: for autonomous architecture)
[old]"As per problem #4 (in problem statement), the only security issue in this architecture is mutual authentication between the WTP and the Ethernet infrastructure to ensure that WTPs cannot be easily stolen and deployed in unprotected networks. This can be ensured by existing mechanisms such as, for instance, 802.1x between the WTP and the Ethernet switch it plugs into."

[new]One of the security issues in this architecture is the need for mutual authentication between the WTP and the Ethernet infrastructure. This can be ensured by existing mechanisms such as, 802.1x between the WTP and the Ethernet switch it plugs into. Another security issue with this architecture is the very fact that the WTP is most likely not under lock and key, but does contain secret information in order to communicate with the backend systems, such as AAA, SNMP, etc. Due to the common management method used by IT personel of pushing a "template" to all similar devices, theft of such a device would potentially compromise the wired network. 

3) Section 4 "Centralized WLAN Architecture": 
At the end of first paragraph, delete "A Network Management Station can be considered part of the Access Controller, but typically the latter contains additional functionality beyond that of the former. "

Instead, add the following to the end of last paragraph "The AC(s) may also choose to implement some of the control functions locally while providing interfaces to access other global network management functions which are typically implemented on separate boxes, such as a SNMP Network Management Station and an AAA backend server (e.g., Radius Authentication Server)."

4) Section 4.5 "Remote MAC" First paragraph, delete ", and improves WTP manageability" at the end.

5) Section 4.6 "Comparisons of Local MAC, Split MAC and Remote MAC"
Added the following paragraph at the end of this section:
"Each of the three architectural variants may be advantageous in certain aspects for certain deployment scenarious. While Local MAC retains most of the STAs state information at the local WTPs, Remote MAC centralizes most of the state into the backend AC. Split MAC sits somewhat in the middle of that spectrum, keeping some state information locally at the WTPs, and the rest centrally at the AC. Many factors should be taken into account to determine the exact balance desired between centralized v.s. decentralized state. The impact of such balance on network manageability is currently a matter of dispute within the technical community. "

6) Section 5 ¡°Distributed Mesh Architecture¡± 3rd paragraph:
[old]Any global configuration or policy
   change can be better served in a coordinated fashion if an AC exists.
   For example, a centralized management entity can be used to update
   every mesh node's default configuration; it may also be more
   desirable to leave certain functions (such as user authentication) to
   a centralized AC (Access Controller), with an alternative being to
   distribute them across the mesh nodes.
[new] Some global configuration or policy
   change may be better served in a coordinated fashion if some form of
   Access Controller (AC) exists in the mesh network, even if not the
   full blown version of the AC as defined in the Centralized WLAN
   Architecture.  For example, a centralized management entity can be
   used to update every mesh node's default configuration; it may also
   be more desirable to leave certain functions such as user
   authentication to a single centralized end point (such as a RADIUS
   server), but mesh networks allows the possibility of each mesh AP to
   directly talk to the RADIUS server.  This reduces single point of
   failure and takes advantage of the client distribution in the
   network.

7) Section 4.1 old title ¡°Interconnection Topology between WTPs and ACs¡± is changed to ¡°Interconnection between WTPs and ACs¡±.
Throughout the document, ¡°topology¡± (when referred to L2 or L3 connection) is replaced with either ¡°connection¡± or ¡°interconnectivity¡±. 

8) 4.2 old title "Overview of Three Centralized WLAN Architectures" is changed to "Overview of Three Centralized WLAN Architecture Variants"

9) 5.1 "Security" (note: for mesh)
Added "Each node that is part of the mesh must be fully trusted for the mesh to be secure. " in the first paragraph.

10) removed all appendix sections & references to the individual architecture submissions

11) Editorial changes (to fix typos, grammar, better sentence/paragraph flows) throughout the document