Monday Apr 20, 2009
Wednesday Feb 06, 2008
Thursday Jan 17, 2008
By Josh Simons on Jan 17, 2008
With MySQL's deep database experience and Sun's extensive hardware and software portfolio and expertise, there are huge possibilities for the kinds of solutions we can bring to market by combining forces. Practically everything Sun does is relevant for such solutions. Here are a few of them:
- We build highly-threaded, high core count SPARC microprocessors and we have partnerships with both AMD and Intel that give us access to their microprocessor roadmaps, giving us the freedom to innovate at the system level over a wide range of product and application requirements.
- We build highly scalable disk and tape subsystems based on innovations within Sun and on the many years of experience in the storage and archive arena brought to Sun through our acquisition of StorageTek in 2005.
- We have Solaris and all of its technologies and capabilities, including our work on Project Indiana to create an OpenSolaris binary distribution and to make the Solaris user experience more welcoming for customers accustomed to GNU/Linux. And when I say "we have Solaris" I also mean we have the expertise within Sun to make Solaris sing on all our hardware platforms--from highly scaled SPARC systems to Intel and AMD systems with a wide range of capabilities.
- And we have the ability as a system company to innovate across our entire hardware and software stack to deliver highly-optimized solutions for our customers. And now with MySQL joining the family, we can extend those benefits right up into the database layer. There are very few companies that have the ability to do this to the extent Sun does.
I also see this acquisition as part of a continuing evolution (or perhaps "enlightenment") on Sun's part with respect to Open Source. I've seen several stages of Open Source involvement at Sun, starting with "Contributor" with our release of OpenOffice and other code to the open source community, moving to "Practitioner" with the creation of the OpenSolaris community, and now moving on to what I think of as "Champion" with our recent acquisition of Cluster File Systems and, soon, MySQL. I can't think of another large corporation that has embraced open source and community development as broadly and aggressively as Sun. And now we are putting our money where our mouth is, as we say in the US. We've gone beyond merely using and contributing to open source software to active investment in significant areas of the open source ecosystem. Of course, for the world to see these acquisitions as a positive development for open source, we need to prove to you that we will be capable and tasteful stewards of these products and technologies. I believe we've demonstrated this so far with our handling of the CFS acquisition by continuing to actively support and extend the Lustre parallel file system roadmap and in particular by our firm statements concerning the recognized importance of Lustre to our many Linux-using High Performance Computing customers. Moving Lustre to Sun was good for Lustre customers and good for Sun. I'm confident the same will be true of MySQL.
By the way, "MySQL" is pronounced My-Ess-Que-Ell, not My-Sequel. Says so here. I've never met anyone who pronounces it incorrectly, but I hear you are out there...
Monday Jun 18, 2007
By Josh Simons on Jun 18, 2007
Sun is holding its next internal engineering leadership conference in early October. These conferences bring together Sun's Fellows, Distinguished Engineers, Technical Directors, and a small number of other invited participants for two days of technology presentations and discussions. This particular conference will be held on Monterey Bay in California in early October.
Any employee may submit an extended abstract for consideration. We are especially interested in technologies that enhance support of red shift applications or improve our effectiveness and efficiency for blue shift applications. Eco-oriented topics are of interest as well. These areas are not meant to be all-inclusive as there is a tremendous amount of other relevant and interesting work being done within R&D. Deadline for abstracts is July 9th.
You should have seen at least one internal email message from Rich Zippel, IAS Organizing Committee Chair, about the conference. If you haven't and are interested, send me email (joshua dot simons at sun dot com) and I'll send you a copy.
Thursday Mar 01, 2007
By Josh Simons on Mar 01, 2007
John Fowler mentioned today at an internal Systems meeting that we were more forthcoming than usual about our processor roadmap at the Sun Analyst Summit held on February 6th in San Francisco. He wasn't kidding.
Check out his slides [PDF] for a roadmap of when Niagara 2, Victoria Falls, ROCK, etc. will be available, and for a basic idea of what these chips are if you don't already know.
People in Systems are kicking butt to deliver an interesting and innovative array of products over the next several years. It's nice to see some of this work acknowledged, if only via a brief glimpse behind the curtain.
Monday Dec 11, 2006
By Josh Simons on Dec 11, 2006
For engineers, the patent process may seem to end once a patent has been granted, but that is far from the truth. You inventors should understand that your work and your innovations continue to play important roles for Sun by highlighting Sun as an innovator and through other means.
Patents have traditionally encouraged individuals and organizations to innovate by granting exclusive rights to inventors. Sun has pursued two complementary paths by building both a strong patent portfolio (see the latest IEEE Spectrum patent survey) and by building a strong track record of participation and donation to the open source community. Our patents give us certain rights for our commercial products, but they also play an important defensive role when negotiating patent cross-licensing agreements with other companies.
Over the past year or so, I've been acting as a technical advisor in a series of cross-licensing meetings with another company that will not be named. We've spent considerable time in these meetings discussing patents whose inventors are friends and colleagues of mine at Sun. I thought I'd explain briefly how such meetings work for those not familiar with the process.
In negotiating a cross-licensing agreement, each side seeks to demonstrate to the other the strength of their patent portfolio by mapping one or more of their patents onto the other company's products. For each such attempt, a proof package is presented. The proof package is a set of slides that introduce a patent and the demonstrate in detail how the other company's product(s) implement every element of a representative claim chosen from the patent. To map successfully, all of the multiple elements of the claim must be shown to exist within the product that is alleged to infringe.
Over a series of meetings, rebuttals to proof packages will also be presented. As you'd guess, a rebuttal is an attempt to demonstrate why a previously-presented proof package was incorrect. It might be incorrect due to a misunderstanding of a product's functionality, or due to a demonstration that the product in question was created using prior art.
The aim in these meetings is to score as many credible hits as possible in order to demonstrate the strength and relevance of one's patent portfolio so as to put one's company in a position of strength when negotiating the particulars of the cross-licensing agreement. Patents are the essential bargaining chips in the negotiation.
There's a bit of a MAD flavor to all of this. Hits that map onto a company's products can be expensive, depending on the product and the impacted revenue. The most effective way to neutralize such a situation is by scoring hits in the other direction to achieve a balance of sorts. The logic of this is simple enough: Patents are important as a defensive weapon because without them Sun would be at an extreme disadvantage in situations like these. Basically, if someone is shooting at you, you had better be able to fire back with something of equal or better weight. While the defensive aspect isn't pretty (can't we all just get along?), it is a reality that Sun cannot afford to ignore. It's a tough world out there, boys and girls.
Friday Oct 06, 2006
By Josh Simons on Oct 06, 2006
Things are hopping in OpenSPARC Land.
As you may know, Sun released the T1 processor design under GNU open source license earlier this year. This is the full design of our latest 32-thread, multicore SPARC CPU--the same CPU that is used in our new Sun Fire T1000 and T2000 servers.
What you may not know is that Ubuntu Linux has been available for Sun's T1-based T1000 and T2000 systems since May, and at least 800 sites are running Ubuntu on SPARC systems. In addition, now that support for the T1 has made it back into the mainline Linux kernel (2.6.17), Gentoo has recently begun offering a distribution that supports these T1-based systems.
On top of all the technical news, a Community Advisory Board has been formed to guide the OpenSPARC community's efforts. It has five members, two of whom are from Sun. Details on the CAB and other OpenSPARC developments are summarized in this recent Sun press release. My favorite quote from the press release: "Since its introduction last March, there have been over 3,500 OpenSPARC T1 hardware downloads..." Hardware downloads! Beam me up, Scotty.
In addition, an online article in the new Zealand edition of Computerworld ponders the relevance and value of open source hardware efforts. Worth a read if you are wondering what impact open hardware projects like OpenSPARC might have in the future.
Monday Sep 18, 2006
By Josh Simons on Sep 18, 2006
For many people the question "What's the strategy"? is equivalent to "What's the roadmap?" In their view strategy is all about the What. In fact, there are other equally important elements to consider.
A successful strategy must embrace the Who and the How as well. It is only with competency in all three of these areas that a company will be able to deliver a truly WOW experience to its customers. Conventional wisdom is to lump How and Who together as Execution rather than Strategy issues, but that's a bug because the three aspects are intimately related and should be considered together.
A company that pays attention to the Who ensures through hiring and skills development that employees will be ready and able to deliver the What. In practice, this requires an understanding of future skills requirements as they relate to the corporate strategy in addition to any current requirements. And skills are not the only issue. Those employees who will execute on the What need the appropriate context on which to base the myriad decisions they will need to make during a product's design and implementation phases. In particular, they should have an understanding of the state of the art as practiced by competitors and, if appropriate, the research community. Add to that a firm grasp of their customers' needs and an ability to distinguish when innovation is appropriate and when it is not.
A company can stumble as it executes its strategy even with the appropriate vision and people in place if attention is not paid to How products are developed. The methodologies matter. Achieving an appropriate level of quality, delivering a product quickly, and doing so without killing your development team is all part of the How. Other aspects of How that can yield benefits are maintaining a culture of data (e.g. a six sigma orientation), assessing the value of new development approaches like extreme programming or agile methods, and introducing formal inspection processes for code or other work products, to name a few. A focus on process for process' sake is not useful. Instead, focus on finding that level of methodology and process which is appropriate for the task at hand.
Friday Sep 08, 2006
By Josh Simons on Sep 08, 2006
Barely five months after we released the eight-core, 32-thread OpenSPARC T1 design under open source, Simply RISC (whose website has been temporarily taken off line, as of this writing) has released the design of the open source S1 processor, a stripped down version of the OpenSPARC T1 intended for use in PDAs, set-top boxes, etc. SWEET!
More details available here.
Thursday Dec 08, 2005
By Josh Simons on Dec 08, 2005
Open-source hardware? Yeah, open-source hardware. I mean: YEAH!!! Open-source hardware!! If you missed it, we announced earlier this week that we'll be open sourcing the UltraSPARC T1 processor early next year. In other words, our latest and greatest, coolest and hottest, SPARC processor. Downloading a chip--and for free--kind of a funky idea. OpenSPARC.
For those not familiar with how hardware is designed this may seem an odd concept, but it is actually pretty straightforward. At least conceptually. Hardware designers write pile and piles of code in a hardware description language (HDL) which is then essentially "compiled" and used to create an actual, physical chip. We used Verilog for the UltraSPARC T1 design. Here's an example of what Verilog looks like from the Wikipedia.
So who's going to play in this new sandbox? I don't think we know for sure yet. I'd expect universities to give it a serious look since students don't often get to look at a full, industrial-strength, cutting-edge microprocessor design in an academic environment. It would be cool to see someone offer a course in which students make interesting modifications to the basic framework. With all of the framework already there, I'd expect to see some pretty funky and creative mods, even within the span of a single semester.
On his blog David Johnson says he doesn't think hobbyists will have too much use for OpenSPARC, but he does allow that someone could take it as a base and, by ripping out a lot of instructions, create a simpler custom processor design that they might then burn into an FPGA. Interesting idea. I could imagine some three-letter agencies I know thinking the same thing.
One thing is sure. If you give people a hugely powerful and flexible tool with all the knobs and buttons exposed, there's no telling what they'll come up with. And watching that is going to be tremendous fun.
Sunday Nov 13, 2005
By Josh Simons on Nov 13, 2005
She is an excellent speaker and I found the Ajax section particularly exciting and the Web 2.0 piece less so. Probably because I feel Ajax is the early embodiment of something important, while Web 2.0 is a squishier concept--more an attitude than anything else.
So, what did learn? First, Ajax is what is enabling some of the new style of web interfaces that have started popping up. Panning by clicking and dragging in Google Maps, for example. Netflix is starting to incorporate Ajax into its interface. Web interfaces are starting to act more like desktop interfaces.
So why is it exciting? If Ajax merely helps web apps move closer to what's already available for apps on the desktop, what is the big deal? The big deal is that you can now see an inkling of a future in which desktop operating systems and applications don't matter anymore. The browser really begins to be capable of becoming the future desktop replacement. And, while Ajax might not be exactly the right solution in the end, you can see the shape of where we are headed.
A desktop born of open standards and based on open standards. Applications that live in the network--applications and data that are accessible from everywhere. Looking out a few years, I see a broad and promising vista (with a small 'v') where a good, solid, standards-compliant browser is all you'll need to access your applications.
Hagan's slides and supporting video clips are here.
- Mirror, Mirror
- DTrace Deep Dive in Boston this Week!
- Sun Microsystems Alumni
- Rest in Peace
- Barbie's Next Career
- Igniting the Earth's Atmosphere
- Virtualization for HPC: The Heterogeneity Issue
- Sun Grid Engine: Still Firing on All Cylinders
- Sun HPC Consortium Videos Now Available
- You Put Your HPC Cluster in a...WHAT??