back to article You've got a yottabyte on your hands: How analytics is changing storage

For all of its advances, the IT sector’s first five decades could be characterised as the electronic storing of systems of record. The move to the electronic era saw paper ledgers, tax returns, bank statements stored and archived safely and legally on machines instead of paper files. Rows of data were worked on in …

COMMENTS

This topic is closed for new posts.
  1. Anonymous Coward
    Anonymous Coward

    "Advanced analytics means using methods to pull meaning from data that will allow for accurate"

    And it will fail. It's thousands of years that mankind is trying to pull meaning by summing empirical observations to allow for accurate predictions, and every time it fails.

    It wasn't until people like Galileo who showed that empirical observation can be faulty, and you need to abstract the problems and understand the real "laws" behind them, that real Science got a big boost towards the right direction.

    But now we are back, haruspices, ehm "data scientist" try to look at coincidences or ephemeral data and build prediction on them. Economists have been trying the same approach for years too, and utterly failed too.

    Still the "kings" or "CEOs" will believe them because it's a great tool to avoid responsibilities, "the gods themselves, ehm, the data themselves told me to act this way!"

    Don't get me wrong, statistics is a great tool as long as you use it properly. Instead we are wasting enormous resources to store haystack over haystack of data because someone told them there's a gold needle within - where and why, nobody knows. And maybe there's no needle at all.

    That's instead of working on real problems about very complex systems, gaining a true understanding, and find real solutions.

    1. elDog

      Re: "Advanced analytics means using methods to pull meaning from data that will allow for accurate"

      Agree with the preferability of having knowledge of the problem space and formulating your hypotheses and experiments (queries?) against that knowledge.

      However, it has become extremely difficult for a small research group or even a large corporate department to have enough subject-matter-expertise on board to be able to have that knowledge. We all can turn to the cooperative groups that may cover more of the ground but this becomes a chore in choosing what to use and who to ask.

      I think the role of big data and data science is more to show what signals there _might_ be in the noise. Much of the data collected in the LHC or other massive scientific efforts is collected to use for possible future research. Many a blip has led to some impressive discoveries based on previously stored and massively analyzed data.

      Don't forget, we wouldn't have discovered ET if all that SETI data was not collected and processed in our 8088's, etc.

      1. This post has been deleted by its author

      2. Anonymous Coward
        Anonymous Coward

        Re: "Advanced analytics means using methods to pull meaning from data that will allow for accurate"

        There's a difference in how CERN and other scientific institution collect and use data, and the "big data" approach. The former ones don't try to use data for any "prediction". They use data to confirm actual theories, or, if any discrepancies arise, they look for new one that fit the data. And still, they may be wrong an never confirmed. When discrepancies in the Mercury orbit were found, the first attempt was to look for an inner planet, because it worked before, and that's how astronomers worked. But there was none, and a better Gravity theory was needed, from someone who was not an astronomer.

        "Big data" looks an attempt to work more like economists work - try to use statistics to fit data somehow, or identify outliers, and then extrapolate that to predict future occurrences, without actually understanding what really happened. Look at how it worked well in protecting us from big financial crisis...

        We should be worried because the impact it can have in our life. If they process big medical data, and without a clue about what they really mean, they try to predict that you or I will develop illness X in Y year, and based on that you employer, insurance, etc. take critical decisions, they are no better than haruspices who decided who had to be sacrificed to obtain favours from the gods, looking at a dead animal liver. Same with software that attempts to predict criminal behaviours.

        I would prefer those stack of money to be invested in real researches, and then use data to confirm or not it. Hoping to be sting by a needle while processing huge haystacks looks a stupid attempt to me.

        But as long as there is someone who thinks it can make easy money because nobody could confirm or deny the results obtained by those "magical" processing, they keep on.

  2. Anonymous Coward
    Anonymous Coward

    So how many hours of porn can you store?

  3. Denarius Silver badge
    Devil

    stored safely and legally

    Bwahahahah, Fantasy. Stored illegally and unsafely more like it. Just ask any 5Eyes, Merkin TLAs and no doubt their Chinese and Russian equivalents along with banks,scammers, those pestiferous companies that want a lot if personal info for no customer business reason and other finance fraudsters. Adding criminals would be redundant. All of them enemies of ordinary people.

    IMNSHO deep data mining is something the LHC and similar scientists can use. As for the rest, mostly another run of the Expert Systems failures.

    1. Brian Miller

      Re: stored safely and legally

      What was the latest thing I read? Oh yes, data left in full public view of world + dog. Or data on HDD inside safe that was stolen. And so on.

      Now there's a yottabyte of sensitive garbage for the taking by some miscreant. How about not slurping it up in the first place?

  4. find users who cut cat tail

    Four words

    Law of diminishing returns.

This topic is closed for new posts.

Biting the hand that feeds IT © 1998–2020