back to article CERN: Build terabit networks or the Higgs gets it!

CERN's openlab has published a document explaining stuff it wants someone to invent so it can get the most out of the large hadron collider (LHC) and its other instruments. The Whitepaper on Future IT Challenges in Scientific Research (PDF) ) sets out the problem, namely that exotic kit was needed to cope with the petabyte-per …

COMMENTS

This topic is closed for new posts.
  1. Michael H.F. Wilkinson Silver badge
    Happy

    Alternatively

    They need faster data reduction methods. We are actually working on much faster processing methods to extract the information from the raw data, so the raw data need not all be stored before processing. They do use preprocessing methods for reducing storage now, but existing methods do not scale well, so for ever faster data rates better, lower complexity algorithms are needed.

    Not that I would mind kit with the specs they want. I want that too. And not to run Crysis!

    1. A Non e-mouse Silver badge

      Re: Alternatively

      Data reduction only works if you know what to throw away.

      I've read several times where researchers re-analysised old data, and made new discoveries.

      1. Michael H.F. Wilkinson Silver badge

        Re: Alternatively

        True, as a scientist you do not like throwing away anything, but with projects like LOFAR, SKA, and the like there is little other option. One trick in the case of LHC-like is to store the detected tracks in a parametric way rather than storing every point that made up the track. That makes a big difference. Might we miss things? Certainly, can we store all the data for later reuse? Not at the moment.

        You really want to be able to do both. Processing faster to get the information from the data and store every bit quickly.

    2. roselan

      Re: Alternatively

      You need to realise that 1 PB/s is 0.02% of what is detected. 99.8 or so % is discarded on detectors "boards" already, before it hits the first cable.

      1. Michael H.F. Wilkinson Silver badge

        Re: Alternatively

        You need to realise that 1 PB/s is 0.02% of what is detected. 99.8 or so % is discarded on detectors "boards" already, before it hits the first cable.

        Very true.

  2. RISC OS

    "A revival of the SETI at Home concept of using individual desktop computers"...

    In a post snowden world I can't see many people wanting to signup for something like this anymore

    1. Destroy All Monsters Silver badge

      Really? People still buy routers from Cisco or use Skype, you know,

      Your BOINC stats

      1. frank ly
        Happy

        I'd be happy to put my old laptop in the DMZ and let CERN use it.

  3. Suricou Raven

    “operate, in a high-radiation, high magnetic field environment, at bandwidths exceeding 100Gbit/s and without losing a single frame”

    Sometimes, you've just got to put up with the fiber trailing across the room.

    1. Martin-R

      Fibre has its own problems in high radiation environments - basically it goes cloudy

      See eg http://en.wikipedia.org/wiki/Radiation_effects_on_optical_fibers

      1. Destroy All Monsters Silver badge

        I am over 30 and what is this?

        THIS UNIVERSE SUCKS!

  4. Lusty

    I suspect a transcription error

    20GB/s works out at 160Gbps so not sure what they are doing between those 1000 100Gbps connections and the storage but seem to be binning a lot of data!

    Sure they didn't say 20 Petabytes? My MacBook can sustain 4GB/s so 20 would be pretty easy to do with a small Violin array PCIe connected.

    1. Anonymous Coward
      Anonymous Coward

      Re: I suspect a transcription error

      Violin 6000 series supports max 4GB/s read. Cern currently have 10GB/s peak capacity. They want 20GB/s (write not read). PCI Express 4 can support up to 31GB/s but writing that to long-term storage is another matter. As for your MacBook supporting 4GB/s sustained write - now THAT must be a transcription error

  5. Anonymous Coward
    Anonymous Coward

    Catchy title -- The Higgs already had too much.

    And while we are at Santa's list, me, I want a red Porsche for five quid, so I can get the most out of the M25.

    Potential missing staff in some areas is a separate issue, and educational programmes are not designed to make up for it. On-the-job learning and training are not separated but dynamically linked together, benefiting to both parties. In my three years of operation, I have unfortunately witnessed cases where CERN duties and educational training became contradictory and even conflicting.

    http://ombuds.web.cern.ch/blog/2013/06/lets-not-confuse-students-and-fellows-missing-staff

    "How should we make it attractive for them [young people] to spend 5,6,7 years in our field, be satisfied, learn about excitement, but finally be qualified to find other possibilities?" -- H. Schopper

    Indeed, even while giving complete satisfaction, they have no forward vision about the possibility of pursuing a career at CERN.

    This lack of an element of social responsibility in the contract policy is unacceptable. Rather than serve as a cushion of laziness for supervisors, who often have only a limited and utilitarian view when defining the opening of an IC post, the contract policy must ensure the inclusion of an element of social justice, which is cruelly absent today.

    http://staff-association.web.cern.ch/content/unsatisfactory-contract-policy

    And a warning to non-western members:

    "The cost [...] has been evaluated, taking into account realistic labor prices in different countries. The total cost is X (with a western equivalent value of Y) [where Y>X] source: LHCb calorimeters : Technical Design Report

    ISBN: 9290831693 http://cdsweb.cern.ch/record/494264

  6. adeater

    Sure... be an open book.

    In Geneva, the very first meeting of the Coordinating Committee for Intercontinental Research Network (CCIRN) was in May 1988. This committee was the first attempt to harmonize the inter-regional operation of the emerging world-wide research network.

    The second meeting took place in October 1988 at a summer resort in Western Virginia, sad and grey this particular autumn. The Americans turned up in force. Bill Bostwick, from the Department of Energy was the Chairman, Barry Leiner from the Department of Defense and Vint Cerf were present. The European representatives were thin on the ground: a German and British representative plus Francois Flückiger.

    In 1991, 80% of the internet capacity in Europe for international traffic was installed at CERN, in building 513.

  7. Anonymous Coward
    Anonymous Coward

    The bigger the better

    Our Internet link at work was upgraded to deal with the large amounts of data we download from CERN. My porn^H^H^H^H important research downloads so much quicker thanks to CERN.

  8. Rol

    Just come out and say it.

    We want to go analogue!

    I wonder how much analogue information could be buffered in a parsec or so of super-conducting cable looping around the site?

    Just a thought.

    1. Truth4u

      Re: Just come out and say it.

      "I wonder how much analogue information could be buffered in a parsec or so of super-conducting cable looping around the site?"

      Some.

    2. Christian Berger

      Re: Just come out and say it.

      Well it's around 3 years of information at the rate you can put it in. The rate you have there is difficult to estimate. Such long lines without any resistance actually might have very nasty properties. For example if it has a resonant frequency, it'll be undamped making it very narrow band as the only dampening happens at the receiver. That's why loudspeakers are made deliberately dampened.

      Of course any realistic chance would be to use fibre and amplify the signal every 100 kilometres or so. However since you insist on going analogue, amplification will always introduce noise. So after perhaps 10 or so amplifications your signal will have disappeared below the noise floor. Analogue fibre systems exist, they are the backbone of cable companies. Even when they claim they are digital, they actually transmit the analogue signal containing their DVB-C muxes over fibres. That's way cheaper than placing DVB-C modulators everywhere.

      The most realistic chance you've got are satellites. Those are, and probably will be for the foreseeable future, be analogue devices. Plus geostationary satellites give you a round trip time of a quarter of a second.

      Still if you don't insist in using a delay line as storage, you can use analogue hard disks. Those were used in early non-linear video editors as a preview storage:

      http://www.youtube.com/watch?v=npExd1D6mtI

      BTW here's the largest device I've seen using delay line storage. It's an electronic standards converter made by the BBC. It can store a whole field of video!

      http://www.bbc.co.uk/rd/publications/rdreport_1970_37

This topic is closed for new posts.

Other stories you might like