r&dBreaking down tech—present and future
64 MAXIMUMPC september 2007
D
ata networks enable millions of devices to
share and exchange data and break down
complex problems into bite-size chunks. But
the term “network” is a mighty big umbrella
covering a number of topologies, including
peer-to-peer, Token Ring, client-server, and
Star. We’re going to focus on one of the new-
est varieties: the mesh network.
Although it’s garnering a lot of attention
in today’s market, thanks to several novel
wireless products, the basic tenets of mesh
networking have been around for some time.
In the early 1990s, the U.S. military identified
a need for a secure wireless broadband data
network that could function on the battlefield,
where little infrastructure existed. Such a
network had to be able to be set up quickly,
expand automatically whenever new devices
were granted access to the system, and allow
extended distance between wired connec-
tions. This idea of an ad hoc peer-to-peer
network forms the basis for today’s mesh net-
work technology.TEACHING OLD NODES
NEW TRICKS
The key difference between a mesh network
and the more familiar point-to-point wireless
network is that each node in a mesh network
can act as a self-contained relay station. This
contrasts with a point-to-point topology, in
which information flows from a central hub out
along spokes to isolated nodes. In such an
arrangement, routing information is confined
to the central hub, which limits how far any
single node can be deployed. Position a node
too far from the hub, and it will lose its con-
nection to the network.
Mesh networks are more decentralized.
Every node is connected to every other node,
with each one capable of behaving as a relay
station; information passes from one node tothe next until it reaches its ultimate destination.
This not only allows for much larger networks
covering greater areas but also enables the
network to heal itself if one or more nodes
becomes incapacitated.
In addition to being self-healing, mesh
networks can be self-forming, meaning that
nodes can automatically discover one another
and link up when they’re within range. In this
way, a mesh network can expand without
human input—simply introduce a node to the
coverage area and both the network and its
coverage expand.
Mesh networks can be either wireless or
wired—in fact, the world’s largest network,
the Internet, is itself a mesh network. Most
smaller-scale deployments, however, are of
the wireless variety. Building a wired mesh
network with just five nodes, for instance,
requires 20 hardwired links in order to con-
nect every node directly to every other node.
Building a hardwired partial-mesh network, in
which direct connections exist only between
the nodes that exchange the most data, would
still require an enormous amount of cable.COMMERCIAL APPLICATIONS
You’ll find mesh-networking concepts
deployed in a variety of tech products. The
Sonos Digital Music System is an easy-to-
deploy audio system that uses mesh network-
ing to stream music to multiple rooms within
a home. Hardwire one ZonePlayer to your
LAN, and it will stream music to as many as
31 other ZonePlayers. The hardwired unit tagseach stream with information about which
location it’s intended for and broadcasts it to
every unit within range. Each unit that receives
the stream examines the data and plays it if it’s
the intended target or passes it on to the next
node within range if it’s not.
The competing home-automation proto-
cols, ZigBee (IEEE 802.15.4) and Z-Wave (pro-
prietary to Zensys, Inc., but broadly licensed),
operate in a similar fashion. Products using
these technologies have been deployed for
lighting control, home security, access con-
trol, and many similar applications. Z-Wave
networks are based on low-power radio
transceivers installed in dimmers, switches,
garage-door openers, remote controls, and
other devices.
When a command is issued from a remote
control in the bedroom to turn on a light in the
backyard, it hops from one Z-Wave device to
the next until it reaches its destination. If the
network determines that the most direct path
is blocked—by a stainless-steel refrigerator
in the kitchen, for example—it will automati-
cally map an alternate path to its destination.
When the targeted light switch receives its
command, it turns on the light and sends an
acknowledgement back down the same path
to the remote.
MIT’s Roofnet project applies mesh net-
working to the 802.11b/g protocol in an effort
to extend wireless broadband Internet access
across the city of Cambridge, Massachusetts.
The project deploys Wi-Fi antennas on roof-
tops to extend and share a single broadbandWhite Paper: Mesh Networking
Decentralized mesh networks
are spreading across the globe
like kudzu—but their impact
is much more beneficial.
by GORD GOBLEIn a point-to-point network (left), a central hub services all the nodes in the network. In order for one node to communicate
with another, data must first travel to the hub and then from the hub to its final destination. In a mesh network (right), data
can travel between nodes without first having to return to the server. Mesh networks can cover much larger areas than
point-to-point networks because each node can act as a relay station.How tHey work Point-to-point wireless vs. mesh networks
POINT-TO-
POINT
WIRELESS
NETWORKWIRELESS
MESH
NETWORK