|
|
 |
home > May/June 2007 issue > article
|
|
 |
|
| Stephanie Diani / WPN |
|
| “It’s an attempt to push the envelope. It’s not a complete redesign, but an extension of what already exists.” — UCLA Professor Mario Gerla |
|
Manet revisited
By John Moore
Special to Defense Systems
Government, industry and academia look to devise a new approach to mobile ad hoc networks on the battlefield
The military is looking for a new twist on an old concept in wireless networking: mobile ad hoc networks.
Mobile ad hoc networks, or Manets, are self-configuring peer-to-peer networks. Each communication device in a Manet is a network node that can interchangeably act as the source of a message, a relay point or a destination. As users and radios or personal digital assistants move about, the network topology continually shifts. The technology’s advocates say Manets can keep parties talking when network infrastructure is impaired or nonexistent — as often happens on the battlefield.
The Manet concept dates to the 1970s and the Defense Advanced Research Project Agency’s Packet Radio Network initiative. The wireless experiment has continued. Since the mid-1990s, pilot applications have ranged from tactical battlefield communications to emergency response.
Now, some 30 years after its initial foray, DARPA is revisiting Manets. The agency’s Strategic Technology Office asked for input on “revolutionary, paradigm-shifting wireless communication and networking concepts.” The office’s request for information acknowledged that Manets, although promising, face limitations in their current form.
“Implementing the Manet concept for efficient use in tactical operations ... has proved challenging,” the RFI states.
Meanwhile, DARPA isn’t alone in mulling the future of Manets.
Researchers in civilian agencies, industry and academia are also on the case. Various initiatives target areas such as performance and scalability. The integration of nodes contained in high-speed vehicles represents another problem under consideration.
Taken as a whole, the collective research may shed light on wireless technology’s evolution in general. For Manets, the inquiries present an opportunity to build on decades of development.
“Periodically, DARPA comes out with a clean-slate, blue-sky request for new ideas,” said Mario Gerla, a professor at the University of California at Los Angeles’ Computer Science Department. “It’s an attempt to push the envelope. It’s not a complete redesign, but an extension of what already exists.”
Weak links
Manets bump into a number of deployment barriers, and DARPA’s RFI includes a list of concerns. “Current network implementations have been plagued with insufficient performance and network lifetime, limited scalability, excessive network overhead, and significant security vulnerabilities.”
The Manet’s nature creates performance issues. As nodes move about, their disposition constantly changes and their arrangement at any given point in time affects how the network functions.
Manets “are extraordinarily sensitive,” said John Stine, an engineer at the Operations Research and Systems Analysis department of Mitre’s Acquisition and Systems Analysis division.
For example, an increase in the density of nodes in a Manet will cause the capacity per user to decline, Stine said. That’s because with more crowding, nodes compete for the same channel.
A decrease in Manet density produces another effect: loss of connectivity. A message on a Manet travels from one neighboring node to another until it reaches its destination. The multihop communication, however, may break down in sparsely populated networks.
The National Institute of Standards and Technology established a Manet test bed a couple of years ago to explore multihop communications. “One problem was that if you allow unrestricted mobility, there is no guarantee of connectivity,” said Nader Moayeri, manager of the Wireless Communications Technologies Group at NIST.
The reliance on intermediate nodes could affect communication in an emergency response setting. Moayeri, who focuses on public
|
|
 |
|
| Photo Illustration by Sam Votsis |
|
| Using the Dynamic Manet On Demand protocol, a mobile ad hoc network could be attached to the Internet by setting up a gateway to route and forward packets destined for nodes with the network. (Source: Internet Engineering Task Force) |
|
safety applications, said a first responder inside a building could lose contact with an incident command center outside of the building if there are no nodes in between.
Then there’s the problem of exposed nodes. A node located on a mountain may become saturated as network nodes separated by the landscape feature use the exposed node to communicate back and forth. The node must wait for all sides to be silent before it can access the channel, Stine said.
Scalability is another stumbling block associated with Manets. Gerla said the current crop of experiments typically involve 20-node to 30-node networks and, occasionally, 100 or more. But the military, Gerla said, has talked about Manets with 10,000 nodes.
Routing protocols, which guide network packets from node to node, haven’t yet come to grips with massive Manets, some observers said. Routing schemes that scale to thousands of nodes have been proposed but not implemented.
“On paper it all looks good, but no one has actually deployed them,” Gerla said.
Another Manet challenge is maintaining connectivity among nodes in high-speed vehicles. The DARPA RFI encourages next-generation Manet proposals that offer a “greater than 90 percent sustained packet delivery ratio for tactical communications between nodes moving at speeds up to 60 mph.”
“Mobility combined with large scale is difficult,” Gerla said.
Incremental Progress
Industry, academic, and government researchers have been refining Manet technology and exploring workarounds for the network’s limitations. (See sidebar, Page 27)
But Pat Ryan, director of defense initiatives at Cisco Systems, said no single protocol would resolve every Manet issue.
“There is no silver bullet ... protocol that is going to solve the mobility problems as we see it,” he said.
Cisco is working on what it calls a transportable service-oriented network architecture, or TSONA. Ryan called TSONA a framework for developing and delivering applications in a mobile, dynamic environment.
Against this backdrop, Cisco is pursuing a Manet concept that seeks to circumvent the scalability issue by breaking networks into smaller chunks.
Instead of connecting 5,000 nodes, a Manet builder may instead opt to link smaller communities of interest, Ryan said. Those smaller communities may then connect to a larger community of interest.
Stine also cited the practicality of each node focusing only on nearby network topology. “The problem is the expectation that every node be able to track the full topology of the network,” he added.
But when the nearby topology is out of communication range, the Manet breaks down. A NIST technology demonstration addresses such situations. The demo involves the use of matchbox-sized intelligent relays that a user drops at intervals as channel quality diminishes, Moayeri said. The trail of bread crumbs — actually Mica2 Motes from Crossbow Technology — provides intermediate nodes between the user’s communication device and the destination node.
Moayeri said the objective of potential solutions such as intelligent relays.
So Manet progress continues in a step-wise fashion, despite DARPA’s call for revolutionary wireless concepts.
Gerla, for one, isn’t anticipating a revolutionary technology that will dramatically change Manets.
“It’s less revolution and more incremental progress,” he said of Manet’s development.
|
|
|
|
![1105 Media [happiness]](/images/ds1_pntmlogo.gif "1105 Media [happiness]"){kind=small}
