Simple Setup Guidelines for Meru’s Plug & Play Wireless LAN

Simple Setup Guidelines for Meru’s Plug & Play Wireless LAN


In this video, I want to talk about how you
can very simply setup Meru’s Plug and Play Wireless LAN. Now, the key elements of Meru’s
network are threefold. We have access points, we have controllers and we have management
servers. Now, to start with, I’m going to take a very simple example where you just
have access points and controllers. Now, access points don’t work in isolation, they only
work in the context of being managed by controllers. So let’s assume that you have a wired network.
I’m going to take, very simply, sort of think of this as a wired network that has many subnets.
Each of these slices effectively is a subnet. So this is a backend network, and what you
want to do is build for yourself a wireless LAN that is an integral part of this wired
network. So the way we build our network is that our wireless LAN is an overlay, it does
not impact your underlying backend, and regardless of which backend network you have, Cisco,
GenePro, Extreme, Brocade, whoever, we will operate on top of that. Go ahead and put yourself a controller, and
if you choose, you could set the controller IP address to a name and map that to DNS.
Now sprinkle around access points. In a companion video, we’re going to talk about how access
points set themselves up and how SSIDs or wireless services are visible on access points
and how they map to VLANs. In this video, all I want to discuss is how quickly the system
can be setup as a plug and play. So so far, step one, you plug for yourself a controller
with one port. You gave an IP address, and if you want as an option, you set for yourself
the IP address and name to the DNS. Now you plug in access points. When an access point
comes up, it first tries to discover a controller without any configuration on its layer 2 network.
So it does a layer 2 broadcast on the wired network, and notice for these two access points,
this controller will respond. Now there are more complex configurations
where you can do layer 3 preferred or layer 2 preferred, this controller may choose not
to respond to all of them. So right now, we’re just going to take a very simplistic example.
So for these two access points the controller responds, and they setup for themselves a
layer 2 tunnel. This is an Ethernet and Ethernet layer 2 tunnel. That’s what these two access
points do. And once they connect to the controller, the controller recognizes them, it makes sure
the software is appropriate, and then if the software is not appropriate — in other words,
it’s not the same rev — then the controller does an automatic upgrade if it is so configured
and then it starts downloading all of the wireless LAN policies. For these access points,
the layer 2 discovery fails. In that case, by default, they’re configured to go into
DHCP. So they will acquire a DHCP-based IP address, and as part of DHCP it is possible
using option 43 to specify this controller’s connectivity end point. This IP address. If
it is specified as part of option 43, these access points get their IP addresses because
they’re on different subnets, and let’s just pick this access point as an example. Once
it gets an IP address, it’ll look for option 43, if it exists, it knows what this connectivity
endpoint is, it tries to connect, and then sets up for itself an IP tunnel. This happens
to be IP — actually, UDP in IP. So let’s call this a UDP in IP tunnel. So all the management
frames and the control actually go as UDP, and they happen to be encapsulated in IP,
so that you’ll essentially be able to transmit frames between the access points and controllers.
So the idea here is this is a layer 2 tunnel, this is a layer 3 tunnel. Now it is possible not to specify anything
in DHCP but actually to leverage DNS. The way that happens is if option 43 doesn’t kick
in, then this access point goes and looks for WLAN-Controller, a well-defined name,
which of course if you want can be customized on the access point. But for a zero touch,
it will go look for WLAN-Controller, and you can map this name to this IP address, in which
case it does a DNS resolution and looks for the controller, at which point it then establishes
for itself a tunnel. Now if this controller goes down, there’s a few different ways of
actually setting the system up for redundancy. We’ll cover more in a redundancy-based video,
but again, in terms of the zero-touch deployment you can have a DHCP server provide multiple
IP addresses — controller IP addresses. So it can provide a primary or backup secondary,
or in DNS you have WLAN-Controller2. So you can setup for yourself a second IP address.
So this is one way where, without the controllers backing each other up, you can have access
points go to a primary controller maybe in this network, and a secondary controller that
lives in an entirely different cloud. Of course, if you want controllers to back each other
up, the sort of traditional way of doing this is you actually have n+1 redundancy, so you
have multiple controllers where if one controller goes down, others stay up. So we have multiple
ways in which you can establish connectivity, or network connectivity, one where controllers
transparently back themselves up or in fact access points have visibility into multiple
controllers, potentially in entirely different networks, so they connect to the preferred
and if that doesn’t work out they can connect with the secondary network. Now you can ask
the question, what happens if I have an access point far way? For example, in one of your
homes? Or in a bridged mode where you are potentially connected over the Internet? So,
let’s assume in this particular case that you have a home network. You’re an employee,
you have an access point from Meru and you take it home. Now we’ll talk about how services are instantiated,
and this is remotely managed in a separate video. This is not really a video about the
remote AP capabilities or the remote access point capabilities. I really want to focus
on the zero-touch configuration part of it. And the idea is when you have this access
point and it comes up, again, of course, unless you have another local controller at home
it’s not going to discover this controller on the layer 2 subnet. If you get an IP address,
notice that so long as this IP address or some NAT equivalent is visible to it either
by DNS or by DHCP, this network, or by configuration, it really doesn’t matter. All this access
point needs to know is how to get to an IP that maps to this IP address. So either it
gets this IP, which is visible, or it gets to a NAT’ed address that in fact gets mapped
to the controller IP. So we work across NATs as well as double NATs. So the access point
will get an IP address, it will try to look for this particular IP, which as we talked
about you can get in multiple ways, and then we establish a tunnel and the point of this
discussion is that this tunnel can be established either on the local campus network or even
across a remote backbone. Thank you.

Danny Hutson

1 thought on “Simple Setup Guidelines for Meru’s Plug & Play Wireless LAN

Leave a Reply

Your email address will not be published. Required fields are marked *