wireless thinking networks electronicspecifier

Upload: cyrus1484

Post on 08-Aug-2018

212 views

Category:

Documents


0 download

TRANSCRIPT

  • 8/22/2019 Wireless Thinking Networks ElectronicSpecifier

    1/2

    Wireless

    July 2013 |July 2013 |Electronic Specifier DesignElectronic Specifier Design

    Thinking networks

    PhotoCredit:UniversityofMichiganlow-power,smartsensorPhoenixsystem:

    DaeyeonKim(http://www.engin.umich.e

    du/newscenter/feature/smallsensor/)

    WWhenexaminingFlame, the

    most sophisticated malware thathas appeared to date, investigatorsdiscovered an interesting feature: Flame can steal andtransmit data from computers that have no Internetconnections.

    It does this by using unsuspecting humans for bi-

    directional data transport. The process begins bycopying itself to every digital storage device that itencounters, including USB sticks and external harddrives. When humans hand-carry portable digitalstorage devices to unconnected PCs and desktops,Flame copies itself into the new computers and startsstealing data. At this point, traditional spyware wouldhave to upload the loot to a remote Internet server, butFlame is a bit more sophisticated; if theres no Ethernetconnection Flame just waits for the next external storagedevice to come along. When one does, Flame copiesitself again and brings along a copy of the stolen data. Itrepeats the process until it finds a computer with anInternet connection, and then it starts transmitting.

    Looked at in a certain way, you could say that Flameuses its human data mules as high-latency Ethernetconnections. You could also argue that its an earlyexample of a promising new tool in datacommunications: Disruption Tolerant Networking (DTN).

    New approach to networkingThe old network model assumed that the bulk of anetworks intelligence was located in controllingcomputers. If connections were lost, data would be lostas well. So network designers tried to get as close as

    possible to whatthe telcos famously

    refer to as Five Nines(99.999 percent) uptime. The

    closer you got to perfection, the harder itwas to achieve the next incremental improvement andthe more expensive things became.

    That model is changing; downstream devices arentnecessarily dumb anymore. As integrated circuits

    become steadily smaller and more powerful, and assoftware becomes more sophisticated, it is becomingeasy and cost-effective to distribute localised intelligenceall across the network.

    The University of Michigan, for example, hasdeveloped a low-power, smart sensor system thatdemonstrates many of the key principles that will beemployed in the smarter, thinking networks of thefuture. At just nine cubic millimeters its the size of aVitamin C tablet, but is solar powered, has aninternal battery and radio, and is equipped with itsown processor called the Phoenix (photo). Theprocessor employs a unique power gatingarchitecture and an extreme sleep mode to achieveultra-low power consumption.

    Smart network nodes like the Phoenix system willprovide network designers with opportunities thathavent been available in the past. Smart nodes willrequire less bandwidth, for example. Theyll beequipped with situational awareness programming thatconsiders parameters like power, network availability

    and the status of surrounding nodes to makeindependent decisions about whether there is any needto log in and report. Smart network nodes will also beable to collect data, time stamp it, log it, and like

    Reproduced with PermissionReproduced with Permission

    http://www.electronicspecifier.com/http://www.electronicspecifier.com/http://www.electronicspecifier.com/http://www.electronicspecifier.com/http://www.electronicspecifier.com/
  • 8/22/2019 Wireless Thinking Networks ElectronicSpecifier

    2/2

    Electronic Specifier DesignElectronic Specifier Design | July 2013| July 2013

    Flame report thedatawhenever anetworkconnectionbecomesavailable. Even ifthe network isdown for minutes orhours, this DisruptionTolerant Networkingmodel will ensure thatdata is not lost, thus freeing designers from theneed to pursue Five Nines uptime.

    With their increased efficiency and intelligence, andtheir ability to hibernate when they have no usefulinformation to report, smart network nodes will requirefar less power than their dumb predecessors. Whencombined with advances in power harvesting, like thetiny solar panel demonstrated on the Phoenix sensorsystem, this will give thinking networks the ability toextend the network edge to include locations andapplications that would previously have beencompletely inaccessible. The thinking networks of the

    future will include network nodes that are completelyindependent of the power grid.

    Hardware for thinking networksLike Flame, with its human data mules, thinking networkswill use multiple techniques to transmit and deliverinformation. Single vendor solutions and proprietarydata communications protocols have already become athing of the past. Users are demanding network-wideinteroperability and the ability to make use of any datacommunications option that may be available. Legacyserial devices are being Ethernet-enabled with Wi-Ficonnections using both embeddable and external Wi-FiAccess Points. Where a fibre optic buildout would beimpractical, designers are deploying cellular routers thatcan establish network nodes anywhere theres cellulartelephone coverage. Thinking networks will continue thetrend, using various combinations of copper cable, fibreoptics, wireless and cellular data transmission. And asthe network nodes get smarter, the methods used totransmit data will become increasingly irrelevant. Like

    Flame, thinking networks will find their way around

    obstacles, takeadvantage of

    whateverconnectionoptions areavailable,see to it thatthe data getswhere it

    needs to go, andensure that it arrives intact.

    None of this will make life anyeasier for network designers. Why? More

    and more network nodes will become independent ofcable connections and the power grid. So even though

    network designers will be able to invest less time andenergy in pursuing perfect uptime, theyll be forced tostart thinking about power budgets. Recommendationsfor things like 30-day battery replacement cycles will beunacceptable. Network designs will have to use powerefficiently enough to ensure that remote network nodesand devices will never go dark. Traditional always oncommunications schemes wont get the job done.

    Pros & ConsDistributed network intelligence and integratedcommunications infrastructures will be engines forincreased efficiency and productivity. Itsestimated that 50 billion devices will be network-enabled by 2020. Many of them will be M2Mdevices that communicate with one another toresolve problems with no need for humanintervention, much the way smart meteringrelieved utilities of the need to send employees outto visit the meters in person. Thinking networks willultimately be able to deliver just about any data,just about anywhere, and the transmissionmethods involved will be completely transparentto the end user, whether that end user is amachine or a human being.

    But the same hardware and software that helpdata travel across thinking networks will greatlycomplicate network security issues. Flame malwarehas already found a way to access the Internetfrom locations in which no Internet connectionexists. We will soon live in a world where Internet

    connections are virtually everywhere.

    Author profile: Mike Fahrion is the Director of Product

    Management of B&B Electronics 8 More from B&B ElectronicsReproduced with PermissionReproduced with Permission

    http://www.electronicspecifier.com/http://www.electronicspecifier.com/http://www.electronicspecifier.com/company/BB-Electronics-Europe.asphttp://www.electronicspecifier.com/company/BB-Electronics-Europe.asphttp://www.electronicspecifier.com/http://www.electronicspecifier.com/company/BB-Electronics-Europe.asp