Cisco To Unveil Wireless Mesh Hardware

An anonymous reader writes "CRN is reporting that Cisco will enter the wireless mesh networking fray next week. Since aquiring Airespace Cisco has been working hard to bring their own mesh technology to fruition. The new solution will target businesses who wish to move the traditional Wi-Fi network outside and possibly cover large regions."

Comments

Re:Why hardware?

[RingDev]

I attended a future of wireless conference in Madison, WI a few weeks ago, they had a keynote speaker from Cisco who talked about this very technology. The things its hardware can do is impressive. And yes, a lot of it could (and is) taken care of via software. These mesh nodes can detect other nodes and alther their signal strength to improve coverage. They can adapt to cover for another node that has gone down. They can immediatly detect unauthorized access points and alert NetAdmins of a potential security issue. These systems are self contained, you don't need to wire each one to the network, they just need juice except for the gateway which can run off of power over ethernet. No expencive wire pulling. No manual configuration. Just plug it in and go.

-Rick

Re:Why hardware?

[tosspot1]

I find this announcement from Cisco quite puzzelling. It's not really a mesh - more of a star. A true mesh means that each node can talk to many other nodes. In Cisco's proposed solution they have a central node attaching (via 802.11a) to remote nodes in what seems like a star configuration.

In a true mesh topology only one node requires a wired connection, and the traffic can be routed through nodes which attach to this node to nodes further away. If I understand Cisco's idea of a "mesh" the end nodes (the ones in the street lights) etc cannot route anything to any other nodes.

Now the interesting thing is that people have been building mesh networks for a few years now. They take the very inexpensive WRTG54G, and put a linux image on it. From that point on it's a matter of using open source software (available in various shapes and forms) to make this very inexpensive box into a true mesh node.

I think another point which is being lost here is that people (including municipalities) want to roll these things out at low cost. Because nobody wants to pay huge money to access them, and the coverage of an individual node is so small, the cost per access point must be low, and I mean REALLY low (like under $100). Cisco will never provide a solution for these costs - that's just not how they work.

So in the Cisco solution we have to also have a tower, so we can blast out 802.11a signal to all the nodes - also not very desirable. What if we want to cover a few streets and don't have proper line of sight? Suddenly it means multiple towers and escalating costs. A true mesh network hands the signal between adjacent nodes so line of sight isn't always needed to cover an area.

There are people out there rolling out working mesh networks right now, yet we are reading how there is big demand for this "new" technology. Please... Or am I missing something here?

The only problem I see with the existing solutions based on the WRTG54G is that it isn't an outdoor device. What we need is for someone to design a little box with very lower power consumption that can handle outdoor environments, and still keep it as low cost...

Just my two cents' worth.

Re:Why hardware?

[saridder]

Actually it really is a mesh. The "controller" is just for management traffic. All Cisco AP's can now use LWAPP, which let's the AP's get all it's management, configurations and control traffic from a centralized device. This means all security configs, polity configs, etc.. are pre-defined on a controller. No pre-configuring each AP before deployment; you just plug in an AP like you would a light bulb and it works.

On the data plane side, they truly are mesh and can talk to any other device in that mesh.