RE: [WIRELESS-LAN] Meru Question

[Dave Molta]

The Meru "good-neighbor" issue involves some technical questions about how Meru uses the duration field of 802.11, which provides virtual carrier sense. Although I am a little disappointed that Cisco is trying exploiting this issue through their sales channel (I bet less than 5% of Cisco salespeople actually understand the issue and it looks a lot like the old IBM FUD strategy), I think it is a potentially serious issue in some environments and we are working with both companies to verify/refute the claim. We're currently waiting on Meru to provide us some AP's so we can do some further testing. As Frank Bulk pointed out, this probably isn't a huge issue for college campuses where you effectibly control the airwaves but it may be a significant issue in multi-tenant office buildings.   With respect to the 11n issue, Mike raises an interesting point about the challenges of implementing 11n's channel-bonding capabilities in the 2.4 GHz band. Personally, I don't think too many enterprises will choose to do this unless they REALLY need the performance. It's more likely that bonding will be used to pump up the numbers on the boxes of retail products in order to move more product off the shelf at Best Buy. Consumers are often swayed by these performance numbers, even though the vast majority are throttled by their Internet connection rather than any performance bottleneck on their LAN.   While it is true that Meru has some advantages because they are more active in scheduling the medium, it's also valid to ask whether the overhead associated with managing WLAN traffic will cause problems for them in terms of controller scalability. You take the bad with the good.   On a related note, I recently learned that Aruba is currently recommending a 4-channel model (1,4,8,11) for dense 2.4 GHz deployments, arguing that the performance benefits of one more channel offsets the co-channel interference issues. Is anyone doing this?   dm From: Ruiz, Mike [mailto:[EMAIL PROTECTED] Sent: Tuesday, June 27, 2006 8:58 AM To: WIRELESS-LAN@LISTSERV.EDUCAUSE.EDU Subject: Re: [WIRELESS-LAN] Meru Question Don,             As a Meru user I can personally tell you that Meru’s system does not negatively impact any other access points unless you put them on overlapping channels or use the rogue suppression.  As far as the “bug” this is simply not true, and I can provide more detail regarding this if you want but didn’t want to bore anyone.  There are lots of tests here and independent tests to verify the first.  Likewise Meru uses Atheros technology and 100% 802.11 standards compliant client side technology.                          My perspective on 802.11n is that Meru is most uniquely positioned to make 11n a workable reality.  Forget the fact that they will continue to eliminate co-channel interference and contention across cells making the bandwidth promised by 11n a reality.  The real core of what makes 11n work is that each channel uses more bandwidth.  Thus in the 24GHz space you will essentially need two of the three available channels to serve 11n.  Well if you’re using 1 and 6 or 6 and 11 what are you left with for neighboring cells?  A coordinated design that can overlap without interfering will be required unless another “band-aid” solution like micro-cells is developed.  Or you can move the 5Ghz space, cut the number of channels in half and then be faced with all the problems plaguing 802.11g today.  It’s consistently amazing to me that vendors tout 11n as a solution when problems like the crash in available bandwidth when 3 or more users come online remains a reality.    Cheers, Mike   -- Michael Ruiz Network and Enterprise Systems Engineer Hobart and William Smith Colleges                 From: Donald R Gallerie [mailto:[EMAIL PROTECTED] Sent: Monday, June 26, 2006 3:47 PM To: WIRELESS-LAN@LISTSERV.EDUCAUSE.EDU Subject: [WIRELESS-LAN] Meru Question   Here at the University at Albany, we had Meru come in and give us an overview on their wireless offering.   From our vantage point, it does appear that Cisco is pushing the controller-based system so we decided to look at other vendors in this space.  As part of this effort, we asked Cisco to come in and give us an overview of their offering as if they didn’t already have a presence on campus.   One of the items that came up had to do with Meru’s method of distributing timeframes to clients (don’t know if I’m phrasing this correctly).  The Cisco engineers said that Meru’s methodology works well in a Meru-only rollout but that they would negatively impact other, non-Meru access points. Additionally, the said that there is a “bug” in the current 802.11b/g standard that Meru takes advantage of and that it may not be there in future (802.11n) standards.   Not that I would doubt anything Cisco says but has anyone heard any similar remarks or can anyone expand on Cisco’s claims?   Thanks….   Don Gallerie The University at Albany ********** Participation and subscription information for this EDUCAUSE Constituent Group discussion list can be found at http://www.educause.edu/groups/. ********** Participation and subscription information for this EDUCAUSE Constituent Group discussion list can be found at http://www.educause.edu/groups/. ********** Participation and subscription information for this EDUCAUSE Constituent Group discussion list can be found at http://www.educause.edu/groups/.