Last week I wrote about the basics of peer-to-peer alerting for mobile devices and how peer networks might facilitate the dissemination of alerts. I also published Peer-to-Peer Alert Systems Part II that provided the several examples of mesh networks. In this article, I wrap up the discussion by presenting examples of star networks as well as combined mesh/star networks.
To recap from Part I: in a star network, devices gather information from a central node that is recognized as the master. A device can leave a network, join a new network with a new central node, and start receiving new information.
A combination of these two types can be quite powerful. A device can receive information from a central node and share that information with its peers if they don’t have it as well. This can help keep information timely and authoritative.
There are many existing systems using peer-to-peer networks that may be valuable to consider when improving peer-to-peer alerting. The first examples use star networks to monitor the safety of workers. The second examples use a combination of the two networks to take advantage of each peer’s unique trove of information.
Examples of Existing Star Networks
Accenture Life Safety Solution
Last April, Accenture announced that Marathon Petroleum Company LP would deploy the wireless-based Accenture Life Safety Solution at its Robinson, Illinois refinery. The solution integrates Wi-Fi and location-based technologies with multi-gas detectors to allow companies to remotely monitor incidents in diverse locations. In particular, it is designed for environments such as refineries, chemical plants, and other locations where dense steel infrastructure can make wireless communications difficult.
Employees who work within dense steel structures often have difficulty communicating with their central offices or with each other. Accenture provides a dual solution: fixed Wi-Fi access points that provide integrated wireless coverage and mobile Wi-Fi access points that are installed in trucks to provide flexible coverage for workers moving around larger open units. The Wi-Fi units, acting as a central node for the star network, then connect to the Marathon’s central communications room by cellular networks.
The Department of Homeland Security has been working with Honeywell to develop GLANSER: Geo-spatial Location, Accountability and Navigation System for Emergency Responders. GLANSER is capable of tracking firefighters in a multi-story building within one to three meter’s accuracy, showing an incident commander what floor a firefighter is on, what area of the floor the firefighter is working in, and whether the firefighter needs help. Each firefighter carries a geospatial locator unit (GLU) that tracks his movements during an emergency situation. The GLU communicates with other GLUs as well as the incident commander’s base.
Using a combination of inertial sensors, magnetic sensors, altimeters, Doppler velocimeters, and ranging radios, GLANSER does not need to rely on GPS to pinpoint a firefighter. This is extremely important when firefighters go into buildings and no longer have line-of-site with GPS satellites.
By considering each example, developers of alerting networks can learn the benefits of using a star network over a mesh network. Whether it’s tracking workers in a wireless-hostile environment or monitoring firefighters’ positions in three dimensions, star networks provide an extra layer of resiliency that can ensure alerts are disseminated.
Examples of Combined Mesh and Star Networks
The term “peer-to-peer” has long had association with less savory practices such as illegal file sharing and bit-torrenting. The technology itself, however, can be quite efficient for delivering large quantities of data. Swarmplayer is an example where the technology is used to deliver legitimate video much more quickly and efficiently that traditional streaming. In fact, Wikipedia’s media arm, Wikimedia, has embraced this technology for delivering video content to its visitors.
Rather than offloading videos from a central bank of HTTP servers (which eats up a lot of bandwidth), Wikimedia has developed Swarmplayer to grab whatever part of the video it can from other network peers rather than directly from Wikimedia’s HTTP servers, greatly reducing the operating costs for the Wikipedia foundation. Swarmplayer then pieces together the video in its logical order for the viewer.
Vehicles that Report the Weather
Finally, alongside the vehicle safety networks described above, the U.S. Department of Transportation is researching ways that vehicles can collect and share weather-related data elements. Vehicles moving along a specific stretch of roadway can share barometric pressure, ambient air temperature, relative humidity, rain levels, and pavement temperatures with each other and with a central node. Car systems can use the information from its peers to predict dangerous conditions like icy roads. Weather observatories can use the “observations” from multiple vehicles to make much more accurate predictions about more complex weather systems.
In each case, a central node collects information that is shared among peers then aggregates that information to flesh out the greater context. Alert developers who are interested in crowdsourcing intelligence gathering for emergency situations can look to these examples for guidance on how to efficiently collect information.
Peer-to-peer networks provide a tremendous opportunity for sharing alerts as well as important detailed information without relying on complex and expensive infrastructures. The precedence has been set by several related industries. It’s up to the alerts and warnings community to figure out how to best use their insight.