Developing Mobile Peer-to-Peer Networks
Dec. 2, 2003 -- Can you hear me now?
For cell-phone users, that simple question often points to one of the most frustrating things about cell phones or any other wireless communication network — the hassles of maintaining a connection to a central and often distant transmitter.
While that can be maddening for cell-phone users, the situation can be even worse for mobile workers who are increasingly switching to wireless 802.11, or Wi-Fi, data networks.
As some Wi-Fi users know from experience, the quality of wireless connections between computers and so-called access points, or APs, can vary greatly.
A Wi-Fi PC in a home living room, for example, might have no trouble connecting to an AP in the den. But a Wi-Fi laptop in the kitchen might not connect because of the pipes in the walls that separate the room from the den.
So several companies are working on new schemes called "mesh networking" that might ease the aggravation. And one start-up company, PacketHop in Belmont; Calif., has hashed out a particularly interesting way to form these mobile meshes.
Making a Mesh of Things
PacketHop's technology is a set of specialized software that can be embedded into mobile devices and wireless network transmitters. The computer code allows each device within the wireless network — whether a cell phone, laptop, handheld computer — to act as a "node," or relay point, and to route digital data along to other compatible devices nearby.
Data bounces, or "hops," from one device to another until it reaches the device that requested the information. If a node becomes unavailable due to interference or other problems, other nearby devices can pick up the slack in relaying the bits of data.
Since PacketHop's special software essentially creates a "mesh" of connections between all the devices, the wireless network becomes more robust at delivering data where it needs to go.
"The routing [software] in each device discovers instantly neighboring devices and broadcasts to [it]," says Michael Howse, chief executive officer of PacketHop. "The net effect is that it creates a topology of peer devices."
Simply Solid Software
Although there are other companies working on similar mesh networking solutions, Howse claims PacketHop has several distinct advantages.
For one, the company and its technology is a spin off of SRI International, a Menlo Park, Calif., research firm with more than 30 years of experience in developing network protocols — including those used by the ARPANET, the military data network that was the predecessor to the Internet.
In fact, PacketHop claims the technology is similar to the interconnected approach used in the wired world of the Internet. "The idea," says Howse, "was to create an instant, mobile Internet."
What's more, since the technology is just a small software code of about 500 kilobytes, Howse claims it can be easily added as part of any common wireless communication scheme — including Wi-Fi and cellular networks.
"We create a whole new capability," says Howse, "not reinvent the wheel."
Wait and See
J. Gerry Purdy, principal analyst with wireless research firm MobileTrax in Cupertino, Calif., says development of mesh networks is a natural progression for wireless communications.
"I think it's an important development in making wireless networks more useful," says Purdy. "It's just taking what [known technology] works in the wired world and make it work wirelessly."
But, he cautions it remains to be seen if and when mesh networks will develop.
"Most of the mesh networking effort today is in an evolutionary stage," says Purdy. And while SRI's technology looks interesting, "[PacketHop] is a true start-up company and haven't been able to demonstrate a true [commercial] system."
PacketHop's Howse says the company is confident in the technology and has recently raised $5 million from U.S. Venture Partners, SRI and others. And he says the company is already building a mesh network for a "big trial" in the first quarter of 2004 with an undisclosed commercial customer.