HP’s The Machine Open Source OS: Truly Revolutionary
Last week, Hewlett-Packard (HPQ) announced its plan to build a revolutionary new type of computer called The Machine. And here's what makes it truly revolutionary, in all senses of the word: The Machine will run an open source operating system developed in universities, as well as Linux and Android.
Last week, Hewlett-Packard (HPQ) announced its plan to build a revolutionary new type of computer called The Machine. And here's what makes it truly revolutionary, in all senses of the word: The Machine will run an open source operating system developed in universities, as well as Linux and Android.
Most of the buzz so far about the project known as The Machine, which HP made public last week, involves radically innovative types of storage in the form of "memristors." Those are non-volatile memory that could store a data center's worth of information on a single computer or phone using virtually no power, according to HP. Leveraging non-volatile memory in a practical way for server, desktop and mobile computing is the core goal of the project.
But there's something else revolutionary in the works, too. That's HP's plan "to build a new operating system all open source from the ground up, optimized for non-volatile memory systems," to power The Machine, according to Martin Fink, HP CTO, director of HP Labs and general manager of HP's Cloud business unit.
Actually, HP wants to build three new, apparently distinct operating systems for The Machine. One will be based on Linux and another on Android (which is also a form of Linux, technically), while the third will be the original work of HP in conjunction with academic institutions around the world. "We want to reignite in all of our universities around the world operating system research which we think has been dormant or stagnant for decades," Fink said.
That initiative is what makes The Machine truly revolutionary on the software front. And when I say revolutionary, I'm thinking both of the older definition of the word, when it referred to things that were cyclical—i.e., that revolved—as well as the modern sense evoking radical upheaval. The new open source OS HP wants to build will do both things.
By promising to engage with universities to develop the OS, HP is restoring the focus to the academic environments that gave rise to Unix (including the University of California at Berkeley's variant, BSD, upon which Apple's (AAPL) OS X is based) and the other operating systems that powered computers for decades, long before private companies took the helm. HP apparently thinks Microsoft (MSFT), Apple and the like are no longer innovative enough to build the OS software of the future, and that universities can do a better job, as they did decades ago.
At the same time, HP is being revolutionary in the other sense of the word by seeking to rupture the co-dependency between PC vendors and Microsoft that has formed the basis for the hardware market since the 1990s. However The Machine's OS shapes up, it will not be Windows, and HP will not have to depend on Microsoft to sell its products. That would be a first, since no major PC maker today offers a serious alternative to Windows-based machines (Dell's paltry assortment of Ubuntu-powered computers notwithstanding).
Of course, for now The Machine remains a nebulous concept, not something HP actually will be selling anytime in the near future. Machines based on The Machine concept may never even fully materialize. But if they do, they'll revolutionize hardware and software in all respects.
“the other operating systems
“the other operating systems that powered computers for decades, long before private companies took the helm.”
Pardon?
When I got into the business in 1974, virtually all computers were running proprietary operating systems, made by the hardware manufacturer.
There were exceptions, such as the Amdahl range of plug-compatibles, running IBM OSs and a few university OSs running on proprietary hardware (e.g. MTS the Michigan Terminal System running on IBM mainframes).
True some of this university research fed into the proprietary OSs, but the article implies that once upon a time all (or most) computers ran OSs from universities, which simply is not the case.
And where is the author’s supporting evidence for his last sentence? Not in this article.
It would be considerably
It would be considerably exciting and refreshing to see a large commercial, proprietary company like HP engage in innovative computer technology, as the VarGuy intimated was the state many years ago.
However the VarGuy and every other person in this country should not be so naive to think that HP is divorced in any way from the throws of Microsoft. It was only 6 – 7 months ago that HP acted as “front man” for Microsoft in attacking the City of Munich, Germany about their decision of adopting Linux in replacement of Windows, most deceptively by reporting a “false” – unsubstantiated and unverified – savings with Windows over Linux, and denigrating the City’s own findings of Linux savings, since which proven correct. A leopard does not change it stripes so quickly, if at all.
I suspect that this project may be a ruse or ploy to leverage against any new (higher) Windows pricing structure from Microsoft, and Redmond’s favourtism stance toward Dell – nothing more.
HP has an abysmal track record of starting, and then quickly abandoning every technology project that did not favour or include Microsoft, especially those involving Free/Open Source Software (FOSS) like Linux
Yes and no. And this is for
Yes and no. And this is for the writer too.. The memristores, are now a reality mainly driven by HP. There are actually memristore components. The problem looks like the mass production , but it wont take ever.
Its like changing from typewriter to computers. And it was only the crazy dream of a nut guy. And its a reality.
It can take even 5 or 10 years to reach the market. But we are gonna live a new reality. And it’s coming. Face it guys…
Universities did not create
Universities did not create UNIX it was created by a small Bell Labs team based around two people Ken Thompson and Dennis Ritchie. Both employed at Bell Labs part of the AT&T Corporation.
Except for the parts of Unix
Except for the parts of Unix developed at UC Berkeley, which were significant enough to win BSD independence in the courts. It was a group effort developing back then.
@wanderson – I agree with you
@wanderson – I agree with you entirely – except that leopards have spots not stripes. Maybe you’re thinking of tigers?
Reading the recent
Reading the recent announcements from HP, Martin Fink is influencing them in a very positive manner. Since he became HP CTO and director of HP Labs, he has led significant change in the hardware, software and solutions that HP is investing in. While The Machine is futuristic and will be interesting to watch develop, HP has made a number of recent announcements that are here and now. A sizeable commitment to OpenStack, multi-vendor SOC’s powering HP Moonshot systems, expansion of the ConvergedSystem range and launched HP Helion, a portfolio of private, public, and hybrid cloud consisting of market-leading HP hardware, software, and services. This is good news for the Industry, as HP is getting back it’s inner geek.
HP should be even more
HP should be even more REVOLUTIONARY by paving the way for a true cognitive computer OS as it is true that the memristor will be a key ingredient of the hardware of such machines loosely based on “neuromorphic” circuits.
What we need is to break the dichotomy and bottleneck of programming the computer vs using it.
What I envision for future OS is something converging in the long run to the Her movie OS. You will not program anymore nor “use” the computer but e.g. you will assign to the computer goals and an agenda (constraints) to attain them, being a (former) programmer, you will tune the computer resources to better solve the problem you’ll submit to it. At the other end of the spectrum, as an user, you will “teach” the computer by providing samples, tips, advices, analogies etc. to have it better focus on the goals you submit to the machine. But as in humans, such computers will engage in a never ending loop, basically continually refining and profiting from its previous experience to successfully engage in new challenges, possibly engaging the experience of other computers over the cloud to accelerate this process.
I agree with your idea of
I agree with your idea of “cognitive” computing. collapsing the memory hierarchy like the memrister does is a big step forward. Biological intelligent neural systems do not have different types of memory, neurons, glia, neuromodulators, neurohormones all work togeter in an integrated way to create short term, long term memory, as well as “ALU” functions (analogue of flow of control, etc). What I have yet to find out about The Machine is whether the memristor combines computing functions with memory functions (I think early articles years ago about “memristors” implied that would be the case, but, understandably because of proprietary issues, I havent found out yet whether the HP design does this). If so, this would be an even more giant step forward, and more isomorphic to neuronal behavior (we don’t have some neurons just for “memory” and others just for “logic and flow of control” — all do both, or better said, both are higher level descriptions we APPLY to the aggregate of functions inplemented by neurons.)
I think I am going to write to HP and try to find out what sample pre-SDK material might be available to research programmers. (I am an AI theorist, and ex programmer.)
Anybody who knows more might reply to this message and start a thread.
Three cheers for HP for taking a chance and betting the farm. Only way progress gets made. Win or lose, they are doing something admirable by going for it.
maybe i miss understood this
maybe i miss understood this describes the formation of OpenBSD and Opensolaris now called OiOS both were developed into Apple and solaris by commercial interests