This is not a NASA Website. You might learn something. It's YOUR space agency. Get involved. Take it back. Make it work - for YOU.
Commercialization

NASA Does The Right Thing – Then It Does The Wrong Thing

By Keith Cowing
NASA Watch
November 26, 2012
Filed under , , ,

How the government can turbocharge private-sector innovation, Gigaom
“Traditionally, NASA attempts to commercialize and otherwise transfer the good work done in its research labs to the public by two means: directly auctioning its patents to the private sector, or maintaining the patents but actively choosing not to enforce them if doing so would impede innovation. NASA claims over 1,200 success stories in this regard, and there’s plenty to show for it. But arguably no single NASA patent has had the same kind of market-disrupting effect that OpenStack has had merely by opening the doors to the community and letting the market drive development and adoption. That’s food for thought.”
Keith’s note: Of course, NASA’s response to the potential of OpenStack? NASA CIO Linda Cureton walked away from OpenStack – while industry has embraced it. And you wonder why NASA cannot figure out how to keep sensitive data off of laptops that are continually stolen? Clearly some management changes are needed in this regard. Check out her blog – its full superficial treatment of important IT issues and pop management babble. Clueless.
NASA CIO Dumps NASA-Developed Open Stack, earlier post
Previous IT posts

NASA Watch founder, Explorers Club Fellow, ex-NASA, Away Teams, Journalist, Space & Astrobiology, Lapsed climber.

8 responses to “NASA Does The Right Thing – Then It Does The Wrong Thing”

  1. sunman42 says:
    0
    0

    Chris Kemp was the main man behind OpenStack when it was Nebula. I’m going to assume he knows what he’s talking about when he says it has value, but — you can claim it was because they were clueless, or out of touch — for whatever reason, no one at NASA could find anything that scaled well under OpenStack and made economic sense. It wasn’t just the CIO. To no one’s surprise, the Nebula cluster was also less than optimal for high-performance computing, because its 10GigEthernet hardware connecting CPUs was an order of magnitude slower than the Infiniband hardware used in high-performance compute clusters (HPCC) already in use at NASA. As a result, NASA CIOs are left scrambling to discover a use to which they can put any sort of cloud, in pointy-haired boss mode, just so they can say they’re using a cloud.

    Right now, a lot of what NASA does other than HPCC involves very, very large amounts of publicly available data, and it has yet to be proven that anyone’s cloud architecture can serve Petabyte-sized archives economically (though Amazon’s new Glacier service, which is rumored to be tape-based, may change that). Not even Google wants to touch those, yet. The endorsements cited come mostly from storage hardware vendors. I wonder what their interest might be.

    I still think there are business applications within NASA that could be optimized for cloud architectures, but the guy who led the development effort for OpenStack, smart and knowledgeable though he is, may not be the most unbiased judge of the tradeoffs. But what he has to say about open source vs. licensing is worth listening to.

    • hikingmike says:
      0
      0

      “less than optimal for high-performance computing, because its 10GigEthernet hardware connecting CPUs was an order of magnitude slower than the Infiniband hardware used in high-performance compute clusters (HPCC) already in use at NASA”

      Is there anything about the OpenStack system that requires using the slower network hardware? It’s my understanding that OpenStack is a software project (and isn’t nearly as specific in the hardware as you imply). From wikipedia – “The mission of the OpenStack project is to enable any organization to create and offer cloud computing services running on standard hardware”. It is even compatible with Amazon Web Services so you wouldn’t technically need any hardware.

      • sunman42 says:
        0
        0

        Believe you misread the comment; it states that _the Nebula cluster_ was less than optimal, not OpenStack, which is irrelevant to some HPCC applications — it’s built for big data, which not all HPCC applications use.

        • hikingmike says:
          0
          0

          Ah ok, I see. I didn’t know much about Nebula. So OpenStack shouldn’t actually be affected by the issues with Nebula cluster, though there may be some psychological effect with people remembering that and equating that with cloud.

  2. Anonymous says:
    0
    0

    NASA helped get OpenStack started.  That is good that the government provided research dollars to do so.  Then someone at NASA decided that Amazon could provide a better service.  Fine.  We pay them to make that decision.

    Keep in mind, OpenStack is a means for an organization to do cloud services as opposed to buying a service. The organization still has to provide the labor and infrastructure.  If a commercial service works better economically so be it.  It reminds me of when you were all over NASA for Spacebook.  Remember what you said then?

    “Another reinvented wheel that needed to be uninvented. I can only
    imagine what they spent to create and maintain this bad copy of
    Facebook.”

  3. Steve Whitfield says:
    0
    0

    Other than “rented storage” usage, I just can’t envision anyone transitioning to OpenStack.  If someone was starting an entirely new project, then they might consider using more of OpenStack — if it met their requirements — but how many new major projects which might have been users have begun since OpenStack became available.  It’s only been around a couple of years and they are revising it 2 or 3 times a year, so I would be hard pressed to call it proven.

    Also, considering the amount of work that has to be done on the user’s end, I think they are going to have to be compatible with a lot more selections of OS, hardware and software before people are going to commit.  My gut feel is that it’s just not going to happen; the system requirements appear to be just too confining.

    Steve

  4. hikingmike says:
    0
    0

    “Today, OpenStack is the fastest-growing open-source project in history (eclipsing even contributions to the Linux kernel)”

    Wow that’s impressive. I’m not sure how they are measuring that (commits?), but with pretty much any possible measure it’s very interesting.

    OpenStack seems like a really exciting project for sure.

  5. hikingmike says:
    0
    0

    Dell Announces Private Cloud Built on OpenStack
    http://slashdot.org/topic/d
    Also a senior CS project I saw was using ambitiously using OpenStack but they had too much of their time taken up by getting that to work so they may dump it and go non-cloud.