Please create an account to participate in the Slashdot moderation system

 



Forgot your password?
typodupeerror
Data Storage The Internet Transportation Hardware

Boeing 787s To Create Half a Terabyte of Data Per Flight 213

Posted by samzenpus
from the peanut-tracking dept.
Qedward writes "Virgin Atlantic is preparing for a significant increase in data as it embraces the Internet of Things, with a new fleet of highly connected planes each expected to create over half a terabyte of data per flight. IT director David Bulman said: 'The latest planes we are getting, the Boeing 787s, are incredibly connected. Literally every piece of that plane has an internet connection, from the engines, to the flaps, to the landing gear. If there is a problem with one of the engines we will know before it lands to make sure that we have the parts there. It is getting to the point where each different part of the plane is telling us what it is doing as the flight is going on. We can get upwards of half a terabyte of data from a single flight from all of the different devices which are internet connected.'"
This discussion has been archived. No new comments can be posted.

Boeing 787s To Create Half a Terabyte of Data Per Flight

Comments Filter:
  • by Anonymous Coward on Thursday March 07, 2013 @04:07AM (#43102451)

    What could go wrong?

    • by ls671 (1122017) on Thursday March 07, 2013 @04:18AM (#43102491) Homepage

      Hopefully, they meant a TCP/IP connection, not "Internet" connected ;-)

      • by jklovanc (1603149) on Thursday March 07, 2013 @05:00AM (#43102639)

        I was hoping that too but the following quote seems to indicate there is some access from the ground.

        If there is a problem with one of the engines we will know before it lands to make sure that we have the parts there.

        I just hope they can only see information and have no control from the ground.

        • by ls671 (1122017) on Thursday March 07, 2013 @05:17AM (#43102693) Homepage

          I just hope they can only see information and have no control from the ground.

          Yeah, right, it seems to ring a bell for me, let's see... OK, let's say: like having a read-only access to a web-site?

          Hopefully the plane pushes the data if it reports in real-time and the plane doesn't have any listening sockets accepting connections on some kind of wireless network. Pilots could also transmit problem reports through radio...

          • by drakaan (688386) on Thursday March 07, 2013 @08:20AM (#43103433) Homepage Journal
            Instrumentation has been part of the stack of internet protocols for many years. It's called SNMP, and it's certainly possible to implement as read-only.
            • by h4rr4r (612664)

              There have been lots of SNMP security bugs over the years.

              Syslog is what you want. The plane exports syslog to a listening box and that is it.

              • by jeffmeden (135043) on Thursday March 07, 2013 @10:07AM (#43104417) Homepage Journal

                Do you want to be the guy responsible for 500GB of syslog messages about every flight? Talk about making your eyes bleed...

                Mar 07 04:08:02.040 UA565 altimeter: altitude 36455.5 feet
                Mar 07 04:08:02.050 UA565 altimeter: altitude 36455.6 feet
                Mar 07 04:08:02.060 UA565 altimeter: altitude 36455.7 feet
                Mar 07 04:08:02.070 UA565 altimeter: altitude 36455.6 feet
                Mar 07 04:08:02.080 UA565 altimeter: altitude 36455.5 feet
                Mar 07 04:08:02.090 UA565 altimeter: altitude 36455.6 feet
                Mar 07 04:08:02.100 UA565 altimeter: altitude 36455.7 feet
                Mar 07 04:08:02.110 UA565 altimeter: altitude 36455.6 feet

          • by wvmarle (1070040) on Thursday March 07, 2013 @08:32AM (#43103513)

            Not likely that it'd be streamed in real time.

            500 GB per flight, say six hours of flight time for an average flight, and I'm at almost 24 MB/s on data production. That's 240 Mb/s. 4G mobile phone can't do that, and when flying you'll often be out of reach of a mobile phone network. So you need satellite - while that may be able to handle the data, it's costing you an arm and a leg. Besides, most of the data is not that interesting.

            What would be viable, though, is for errors to be automatically transmitted. Like an engine that detects an anomaly, indicating it needs maintenance, that such a little bit of information is forwarded to the ground.

            • by Rich0 (548339)

              Yup, makes way more sense to just get a 2kb event list sent during flight, and then the dispatchers can have the right maintenance team on the tarmac when the plane pulls in. They can plug in an ethernet cable and download whatever raw diagnostic logs they need to speed up the repairs.

              I could even see an in-flight request for specific logs to be sent, but it isn't like you can change an engine in-flight so I'm not sure how much time that would really save.

              Saving multi-TB of data on looped storage for such

              • by wvmarle (1070040)

                Hardening shouldn't be too hard - don't modern black boxes store a lot of digital info about the flight already? With advancing technology I would expect it to be totally feasible to have all this info stored in the black box, in addition to the regular plane's computers.

                • by cdrudge (68377)

                  I believe there are 88 parameters that are required to be recorded by modern flight data recorders, although they have the option to record far more. However I would be surprised if they dumped the same amount of data to the FDR that they would send to a quick access recorder [wikipedia.org] type of device.

            • by acoustix (123925)

              Well, technically the article only says that the large amount of data could be created. It doesn't say that all data created is sent over a network. I would guess that the system doesn't send all data, and the rest of the data could be downloaded once the plane has landed.

          • Ther is a differance between a censor that can be read and an actuator that moves the flaps.
        • by AmiMoJo (196126) *

          There is already a broadcast-only system in place for sending this kind of telemetry back to HQ. Air France was getting warnings about their flight that went down over the Atlantic telling them that the air speed sensors were producing conflicting data (nice of them to share that with the pilots...)

          This could be a similar sort of thing using UDP. It could also be two way - lots of servers accept TCP/IP connections for management and telemetry. It will be via a dedicated satellite link most likely and hopefu

          • by Joce640k (829181)

            Yep. Some engine manufacturers already monitor all their engines continuously, in real time.

            Rolls Royce was the first, I think GE have started doing it too.

            It's not a big leap from monitoring the engines to monitoring the entire aircraft.

        • by 0123456 (636235)

          I was hoping that too but the following quote seems to indicate there is some access from the ground.

          Fault data is already sent over ACARS (a low-bandwidth text-message-like protocol) from existing aircraft. When the avionics detects a fault it sends a message so the maintenance staff will be ready to fix it after the plane lands.

      • by leuk_he (194174)

        eeeh, Almost everything is internet connected. They tell you that the nuclear power central is not connected, but if you really audit it you will find there is some kind of double vpn to connect it.

        The international airports have their own internal networks, but in some ways they are always connected to the internet. There are some vpn and firewalls between your home desktop and the plane, but in the end it is all connected.

      • Hopefully, they meant a TCP/IP connection, not "Internet" connected ;-)

        Whatever it is I hope it's an encrypted connection. That would put them one step ahead of the military [slashdot.org].

      • I do seriously hope that too. Connecting flight critical hardware to an open network is asking for trouble.

        Quite frankly, if their CISO didn't veto that and got shot down, he should be fired. Out of a cannon.

        • by Wolfrider (856)

          --Bah, almost all they have to do is put the plane on IPV6 and it will automatically be more secure, since no one is using it :-P

    • by Richard_at_work (517087) <richardprice.gmail@com> on Thursday March 07, 2013 @05:24AM (#43102721)

      This isn't exactly new - Rolls Royce, GE and Pratt & Whitney all do "power by the hour" rented engines, which are permanently connected (allowing for coverage issues) to data receiver centers which manage them. If they need maintenance, are running hot or have a vibration issue (its amazing how much you can discern about an engine due to its vibration levels), the engine manufacturer can determine before the flight has even ended whether or not the engine needs that maintenance at that point, needs replacing, or can suffice until the aircraft can be rotated to a full maintenance center to be swapped out.

      On PBTH engines, its typical that the airline will be called by the manufacturer to report the issue before the crew flying the aircraft ever notices anything.

      Its also a service engine purchasing airlines can select.

    • by Anonymous Coward on Thursday March 07, 2013 @06:58AM (#43103045)

      wget http://www.vigrin-atlatic.com/flight/12321/manage.aspx?flags=down&gear=up&parachutes=deploy

    • by LWATCDR (28044)

      Slashdot not knowing what they are talking about. They are not going to Intranet connect flight controls. They will NETWORK them. Probably using a variation of firewire or maybe ethernet. Slashdot news for wannabe nerds and and clueless politicos.

    • by tibit (1762298)

      Nothing critical on those planes uses internet in any shape or form. The closest you get to internet-anything is the TTE (Time Triggered Ethernet) -- that's what interconnects the avionics together, probably the FADECs too. About the only commonality with "internet" is that it uses the Ethernet for lower 2 or 1.5 OSI layers, depending on who you ask :)

  • by Anonymous Coward

    It would be cool if one could play back that data in a flight simulator to recreate the flight.

    • by Anonymous Coward on Thursday March 07, 2013 @04:46AM (#43102569)

      Sufficient data for playing back flights in simulators has been recorded by the black boxes for a couple of decades already. It's done to evaluate the crew's actions (and compare those with what they've been trained to do) as well as to test if alternative actions could've prevented the accident. A good example was Swissair 111 (in 1998) which had a fire on board and crashed before the crew could land the aircraft. The crew didn't even try to get it to the nearest airport as fast as they could because they wanted to dump fuel, prepare the cabin and were reluctant to land at the very closest airport at first since they wanted one with Swissair mechanics. As part of the investigation crews in simulators got exactly the same fire scenario at the same point and among other things tested if they had been able to reach the airport if they didn't dump fuel and land overweight. In those scenarios too, however, it would've taken too long from the point the fire was first detected.

  • by stoploss (2842505) on Thursday March 07, 2013 @04:12AM (#43102469)


    dave@console:~ ssh dave@hal-787
    [dave@hal-787 ~]$ echo "1" > /dev/landing-gear-doors
    echo: I'm sorry, Dave, I can't do that.
    [dave@hal-787 ~]$ sudo echo "1" > /dev/landing-gear-doors
    dave@console:~ Connection to hal-787 lost.

    • Re: (Score:3, Informative)

      by Anonymous Coward

      Admittedly sudo echo "1" creates the most powerful "1" there is, but it's still not quite enough to affect permissions checks on the redirection.

    • by Agent ME (1411269) <agentme49@@@gmail...com> on Thursday March 07, 2013 @04:34AM (#43102541)


      [dave@hal-787 ~]$ sudo echo "1" > /dev/landing-gear-doors

      You don't want to run echo as superuser; a regular user can echo 1 to the program's own stdout just as easily as superuser. The shell is what opens the output file (/dev/landing-gear-doors), so you either need to run the shell as superuser or have a different program as superuser which opens the file. Either of these will work:


      [dave@hal-787 ~]$ sudo sh -c 'echo "1" > /dev/landing-gear-doors'
      [dave@hal-787 ~]$ echo "1" | sudo tee /dev/landing-gear-doors

      • Re: (Score:2, Insightful)

        by ls671 (1122017)

        He,he, I kind of never got used to sudo. On all systems I ever had access to, simply typing "sudo bash" gives me the old root shell...

        • sudo -s is fewer characters than sudo bash.

          sudo make me a sammich

          • by jkflying (2190798)

            If you're short on characters, plain old su will save you a few more...

            • by ls671 (1122017)

              It doesn't work because root has no password set on a typical sudo system.

              ~# grep root /etc/shadow
              root:!:15997:0:99567:5:::

              I use my user (not root) password to execute sudo bash. Thanks to the guy who gave a hint about sudo -s!

              • You could just use "sudo su"
                • Re: (Score:2, Informative)

                  by Anonymous Coward

                  As far as I know, "sudo -i" is the best option. This way also the home directory is changed to /root, so that the configuration files (or history files) in the home directory of your regular user aren't suddenly owned by root.

        • Re: (Score:2, Informative)

          by Anonymous Coward

          If you find yourself doing commands needing sudo, try this.

          sudo !!

          !! will repeat the last command.

      • Re: (Score:2, Funny)

        by Anonymous Coward

        Just log in as root. no sudo needed.

        and then do
        shutdown -h now

        That gives the pilot something to do!

    • by RDW (41497)

      Even attempting to activate the manual override can be confusing for those without prior unix experience:

      http://xkcd.com/912/ [xkcd.com]

  • by scsirob (246572) on Thursday March 07, 2013 @04:13AM (#43102473)

    Connecting flight controls to "The Internet" would be the stupidest of all ideas. If they do this, anyone getting on board would be a candidate for the Darwin awards.

    I'm sure they meant to say that all these systems are networked together, using ARINC or other aviation network technologies.

    • there is a difference between internet (any internetwork) and the Internet (a worldwide publicly accessible system of interconnected computer networks)
    • Connecting flight controls to "The Internet" would be the stupidest of all ideas. If they do this, anyone getting on board would be a candidate for the Darwin awards.

      I'm sure they meant to say that all these systems are networked together, using ARINC or other aviation network technologies.

      TFS says "an internet". A network -> network connection is an internet connection, regardless of whether it's routed to "the internet".

    • by Sique (173459)
      If it uses the IP (Internet Protocol), then it's probably an internet.
    • Another possible meaning is that it's purely dataloggers monitoring every device he describes. So there's no way to control anything, just a way to monitor it.

      • by rioki (1328185)

        True that, by implementing push only the plane is safe. Let's hope they implement proper authentication, since injecting fake error reports may not hurt the plane, but can definitely hurt the airline.

  • by Anonymous Coward on Thursday March 07, 2013 @04:19AM (#43102493)

    Sources say it's an XML dump. Maybe 100KB of actual data in there.

  • by verifine (685231) on Thursday March 07, 2013 @04:53AM (#43102597)

    Finally it makes sense, the plane (which is often in the clouds) generates data (which is stored in the cloud.)

  • Because all 787s are grounded. Perhaps Boeing should have concentrated on the basics for a bit longer instead of the frivolous nonsense that is nothing more than a tickbox on a marketing sheet.

    • by jimicus (737525) on Thursday March 07, 2013 @05:10AM (#43102675)

      I don't know that it's all that frivolous.

      Your average airline is running on razor-thin margins. They do NOT want a plane grounded for any longer than is absolutely necessary - because a grounded plane isn't earning any money. If an airliner can signal any faults several hours before it lands, the maintenance crew have advance warning so they know exactly what to look at (and maybe even have parts available) the instant it touches down rather than have it sitting on the tarmac waiting for parts to arrive.

      • A grounded plane not only isn't earning money, it costs and arm and a leg to just stand there.

        I doubt they're happy with this.

    • by DerekLyons (302214) <(fairwater) (at) (gmail.com)> on Thursday March 07, 2013 @07:45AM (#43103239) Homepage

      Maintenance data is far more than a 'tickbox on a marketing sheet', it's the absolute bedrock for efficiently operating a large fleet of... well, anything. Cars, trucks, planes, etc. That airlines and airframe manufacturers can and do collect and analyze tons of maintenance and operational data is a large part of why air travel is so safe and (relatively) cheap.

  • by martin (1336) <<maxsec> <at> <gmail.com>> on Thursday March 07, 2013 @05:23AM (#43102717) Journal

    Have been doing this for years. They constantly stream data to RR HQ and theres a team of highly experienced people watching the data. http://www.youtube.com/watch?v=zPIYBgZNrsg&sns=tw [youtube.com]

    • by martin (1336)

      Quite - everything seemed to happen v quickly once the oil pipe broke. And yes the engine should be able to contain a broken blade

      Allows get issues with new engines and the original Trent almost broke RR (its development costs being the major reason fir nationalising it in the 1970s from what I remember

      • by pittance (78536)

        And yes the engine should be able to contain a broken blade

        Slight correction: the engine should be able to contain a failed fan blade. A turbine disc (or the one third disc pieces it breaks into), which is what failed on the flight here, acts like god's own laser beam and is not containable and so the aircraft will always have to be designed to tolerate that type of failure as it duly did. In fact the regulations require uncontained fan blade failures to be checked as well, even though they shouldn't happen.

  • What percent of that "data" is bad scripting, really bad mark-up and AdSense shit? Is there maybe an address like "m.part_number.plane_number.location.internalsite.boeingcorp.com" so the people who fix the problems can get the info without being tracked and spammed by OK Cupid and 100 different tool sellers?
  • by Anonymous Coward on Thursday March 07, 2013 @06:13AM (#43102885)

    "We can get upwards of half a terabyte of data from a single flight". Well, provided they're actually able to fly, which is not the case, last time I checked.

  • 0.5 TB about how your flight is late or cancelled and there's nothing anyone can do about about anything ever.

  • by Required Snark (1702878) on Thursday March 07, 2013 @07:34AM (#43103177)
    In less then two minutes on Google I found this article with a description of the 787 on-board network: http://www.avionics-intelligence.com/articles/2011/06/boeing-787-avionics.html [avionics-i...igence.com]

    The Core Network, which is standard on the 787, uses computing servers and networks based on commercial open standards. It also has a variety of third-party applications to manage the onboard data flow to improve airline efficiency. The Common Data Network (CDN from Rockwell Collins is a, bi-directional copper and fiber optic network that utilizes ARINC 664 standards and protocols to manage the data flowing between the 787's onboard systems. It is based on Ethernet technology and enabled for avionics systems. The CDN has higher data rates, expanded connectivity, and reductions in overall aircraft weight when it is contrasted with point to point topologies, Rockwell Collins officials say.

    Another quick search on ARINC 664 yields the following: http://en.wikipedia.org/wiki/Avionics_Full-Duplex_Switched_Ethernet [wikipedia.org]

    AFDX is a next-generation aircraft data network (ADN). It is based upon IEEE 802.3 Ethernet and utilizes commercial off-the-shelf hardware thereby reducing costs and development time. AFDX is one implementation of deterministic Ethernet defined by ARINC Specification 664 Part 7. AFDX was developed by Airbus Industries for the A380, initially to address real-time issues for flight-by-wire system development. A similar implementation of deterministic Ethernet is used on the Boeing 787 Dreamliner. AFDX bridges the gap on reliability of guaranteed bandwidth from the original ARINC 664 standard. It utilizes a cascaded star topology network, where each switch can be bridged together to other switches on the network. By utilizing this form of network structure, AFDX is able to significantly reduce wire runs thus reducing overall aircraft weight. Additionally, AFDX provides dual link redundancy and Quality of Service (QoS).

    So both the Airbus 380 and the 787 use COTS hardware and Ethernet, as does the Internet. Although slightly sloppy, describing the network as an "internet" is technically correct. Asserting that the data is "bloated XML" or that their is bad scripting, spam or cookies involved is grossly stupid.

    I have worked with previous ARINC formats, and the data is very compact. It fact, it is positively cryptic, and generally you use software to turn it into a more human friendly form, like a line graph. So if there is a half terabyte per flight, it is all "real" data. Any of the posts that assume otherwise are a combination of arrogance and ignorance, which is typical for what passes as comments on Slashdot these days. Hence my sig:

    • by Muad'Dave (255648)

      ...deterministic Ethernet defined by ARINC Specification 664 Part 7.

      That rang a token-bus bell, so I dug a little deeper. It seems they borrowed the token-bucket concept [wikipedia.org] from ATM [wikipedia.org] to get their deterministic behavior. Pretty clever.

      On a side note, how I wish we'd standardized on ATM-to-the-premises!

    • Asserting that the data is "bloated XML" or that their is bad scripting, spam or cookies involved is grossly stupid...Any of the posts that assume otherwise are a combination of arrogance and ignorance, which is typical for what passes as comments on Slashdot these days.

      Although I'm always tired of the "what could possibly go wrong" posts (and was tired of them back in '03 too, slashdot has always had those posts)...Dude, the xml comment was a joke, and a pretty funny one at that. As I literally laughed out loud after reading it, it never even occurred to me that somebody else might read that post and think it was a serious concern that 100kb suddenly transformed into half a terabyte because of xml tags. Until I read your post that is.

      You should go in search of your sense

    • by HungWeiLo (250320)

      AFDX and the Rockwell CDN is not just "COTS hardware and Ethernet". AFDX/664 is a protocol based on UDP packets coupled with hardware redundancy, real-time data transfer, and packet time/sequencing verification. Having the technology based on UDP packets makes it possible to use existing RJ-45 cables and cheap switches for development.

  • And lets hope the server controlling the engine is not written in PHP or something

    • There are only bad programmers. The real problem is to ensure the whole system is not reachable from the Internet, otherwise terrorists may not even have to take the plane...
  • John Deere has a variety of satellite-guided systems that can be implemented, and there are a few methods to monitor and program firmware over a wireless connection (I don't know the exact communications medium, it's not my field). Suggesting that there is 500+GB isn't unlikely, because I use CAN to interact with the hardware that we test, and a few seconds of reading a few variables can easily be 1MB.

    Here's my quick number-crunching output:
    500GB / 5 hours (estimate average flight including prep) = 100GB /

  • by PPH (736903) on Thursday March 07, 2013 @12:06PM (#43105869)

    Don't worry. It's got battery backups.

    Oops!

Power corrupts. And atomic power corrupts atomically.

Working...