Extending the Internet to the Bottom of the Ocean

Researchers test acoustic/WiFi network on Lake Erie near Buffalo

Wireless networking vendors have promised (and often delivered) access to the Internet from anywhere.

To technologists, that translates as “anywhere it’s possible to get a wireless signal.” To end users, it translates as “why can’t I get my e-mail from inside this cave?”

That lack of nuance in the expectation of end users has driven corporate technology divisions, wireless equipment vendors and cellular providers to litter Corporate America with repeaters, femtocells, WLAN access points and wireless range extenders that radiate the Internet into almost every train, plane and automobile and every corner of every corporate sub-basement, lunchroom or remote executive retreat.

Wireless and cellular networks cover even vacation beaches and extend over the ocean to ships at sea but not, so far, under the ocean.

A team of researchers at the University of Buffalo believe they’ve solved at least the technical problem of how to push wireless networking signals for long distances through the deep ocean to connect offshore oil and gas platforms, floating and underwater tsunami sensors and other remote facilities without having to bounce signals off a satellite first.

Radio waves tend to be smothered or distorted by travel through water; most ocean-based sensors use acoustic waves instead, which link sensors into underwater acoustic sensor networks (UWASN).

Tsunami sensors on the sea floor use acoustics to send data to surface buoys, which convert them to radio to forward them by satellite, for example. Each set of acoustic-networked devices uses a different networking scheme, however, making it difficult to have one set of sensors relay data sent by another, or allow access in the opposite direction so land-based users without special connections can see what’s happening on the sea floor.

“A submerged wireless network will give us an unprecedented ability to collect and analyze data from our oceans in real time,” according to Tommaso Melodia, leader of the research team whose paper “The Internet Underwater: An IP-compatible Protocol Stack for Commercial Undersea Modems” will be presented at a conference on underwater networking to be held in Taiwan Nov. 11-13 by the Association for Computing Machinery.

The difficulty of converting data from TCP/IP networks to the idiosyncratic protocols of UWASN-connected devices – not to mention the difficulty of getting signals to the devices in the first place – mean underwater sensors may be able to send update data to relays, but cannot normally be reprogrammed, reconfigured or even closely monitored once they are deployed, according to the paper describing the team’s results. (PDF)

It is possible to tunnel TCP/IP traffic to UWASN devices, but doing so creates overhead, increases what it already an almost unworkable latency due to the slow speed of signals in seawater, and uses extra power from sensor batteries with a lifespan that is already too limited, the paper explains.

Melodia and a team of graduate students from the University of Buffalo designed a low-power IPv4/IPv6-compatible networking protocol that uses very low power, compresses headers, is tolerant of fragmented data and connection delays, allows bi-directional communication with (and reconfiguration of) existing underwater sensors and is compatible with standard TCP/IP networks and IP router proxies.

The approach is more than a simple translation from one networking medium to another. It leaves the higher-level TCP/IP networking protocols intact, but adds an adaptation layer between the data-link layer and network layer that compresses headers, changes packet size, transmission time-out settings and other requirements to be compatible with slower underwater transmissions.

Unlike most other approaches, the new protocol is not tied to any specific application and includes the ability to auto-configure packet size and other transmission characteristics and automatic route-control using a set of border routers.

It also allows sensors whose connectivity depends on very low Layer 2 mesh routing and which keep their own routing tables to continue transmitting that way, then elevate transport from Layer 2 to the Layer 3 IP networking layer through a “virtual link.”

The team tested the implementation using a Linux-based driver, both PC and ARM-based computers and a Teledyne Benthos SM-75 Modem. They sealed two network nodes in 40-pound waterproof cases, dumped them into Lake Erie near Buffalo and transmitted instant-messaging signals from the application IPTUX from one to the other. They were also able to transfer files using FTP from an underwater client to server.

While the work is specialized and primarily designed for undersea scientific applications, it could make possible the extension of an Internet of Things, encourage the development of other power-conserving networking protocols or expand the number of sensors looking for pollution, tsunamis and other underwater events with serious consequences on land.

“We could even use it to monitor fish and marine mammals, and find out how to best protect them from shipping traffic and other dangers,” Melodia said in the University’s announcement. “An Internet underwater has so many possibilities.”

Image: University of Buffalo/ Douglas Levere

Post a Comment

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

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>