March 23, 2010 TurboHercules: The Next Step
By Roger Bowler, Creator of Hercules and Co-founder of TurboHercules
(Posted in News & Blogs section of turbohercules.com on 23 March, 2010)
The growth of the Hercules mainframe emulator from a one man project into a production-capable system demonstrates the power of open source software development, which IBM has rightly recognized by embracing Linux as a key part of their future strategy.
As the founder of the Hercules project, I can state with confidence that our emulator is in no way an enemy of IBM. In fact, the Hercules project is made up of some of the biggest mainframe fans on the planet. We are people who have spent our entire careers learning the ins and outs of this architecture, and we want nothing more than to see it thrive far into the future. Mainframes are now so deeply embedded in the infrastructure of modern society that they are too important to be left in the hands of a single company (IBM).
Applications and business data that run on IBM’s mainframe software are among the most valuable in use by the world’s largest corporations and government agencies. By IBM’s own estimation, $5 trillion worth of applications run on IBM mainframes. These highly proprietary applications, which have typically been developed over many years and embed an incalculable wealth of company-specific business knowledge, cannot be easily migrated to other platforms. Even with all the talk about Linux on the mainframe, the reality is over 90% of mainframe applications still use native mainframe operating systems such as z/OS and languages such as COBOL, PL/I or Assembler.
The Hercules open source project team created the Hercules emulator in order to give the owners of this hugely valuable installed base of mainframe applications freedom of choice in the hardware systems used with this software. The extraordinary performance and reliability offered by the current generation of multi-core 64-bit processor technology from Intel and AMD make emulation a viable alternative for many mainframe applications. There is no reason why mainframe application owners should not be allowed to benefit from these technologies when they match their needs. That they are not presently available is due solely to IBM’s needlessly restrictive licensing policies regarding the use of z/OS and other IBM mainframe system software on non-IBM hardware. We at TurboHercules abhor all forms of software piracy, and have the utmost respect for the valuable intellectual property embedded in IBM’s mainframe software. We ask only that IBM allow the customers who purchase that software to deploy it on the hardware platforms of their choice.
IBM claims that the mainframe represents only a tiny percentage of worldwide server shipments. But this isn’t quite the whole story because as IBM never tires of repeating, the mainframe is an entirely different animal from Linux, UNIX or Windows servers. No other architecture in the world performs as many business critical transactions each day or stores as large a share of the world’s most important corporate data. Viewed in this light, the fact that IBM today owns 100% of the IBM-compatible mainframe market cannot be explained away as something not worthy of concern. This monopoly position is without precedent in the history of the mainframe – indeed, for decades mainframe customers benefitted from the competition resulting from manufacturers of mainframe hardware compatible with the IBM architecture. But no more. Today, thanks to its use of lawsuits, preemptive takeovers and threats against would-be competitors, IBM is the only vendor in the world allowed to sell hardware systems compatible with applications that run on its vast suite of mainframe software (e.g. z/OS, z/VM, CICS, IMS, DB2).
The outcome that we at TurboHercules hope for is a return to the competitive market for mainframe technologies that existed in the ‘80s and ‘90s, where IBM licensed its operating systems to customers of the Plug Compatible Mainframe (PCM) manufacturers such as Hitachi and Fujitsu/Amdahl.
To this end, I wrote to IBM in July of last year asking that it allow customers to acquire z/OS licenses for use with the Hercules open source emulator, with the understanding that pricing and conditions would be set “by IBM at the sole discretion of IBM on reasonable and fair terms.” After four months of silence, in November 2009 I received a reply from Mr. Mark Anzani, CTO of IBM’s mainframe division, that not only rejected our request, but went on to accuse Hercules of “infringing” IBM’s intellectual property.
Let me state as forcefully as possible that, regardless of one’s stance on the morality of a giant IT vendor such as IBM threatening patent action against a small open source project such as Hercules, we do not believe that the Hercules emulator violates IBM intellectual property. We wrote back immediately to Mr. Anzani to express our surprise that, after ten years during which Hercules has acquired thousands of users around the world, including many within the ranks of IBM itself, IBM has now suddenly discovered that we are violating its intellectual property. In our reply we asked that IBM identify the specific property we allegedly violated in order that we could investigate that claim. In the unlikely event that IBM’s assertions were found to have merit, we asked further that IBM consider adding such intellectual property to their much-publicized and deservedly admired “non-assertion” pledge to the open source community – the “IBM Statement of Non-Assertion of Named Patents Against OSS” described in http://www-03.ibm.com/press/us/en/pressrelease/21846.wss.
Given this situation, TurboHercules has reluctantly taken the step of filing a complaint asking the European Commission to restore free and fair competition to the IBM mainframe market. Mainframe customers should be permitted to run the applications and data that they own, and in many cases developed, on the computer hardware of their choice. It is my sincere belief that TurboHercules will contribute to the growth and longevity of the mainframe ecosystem upon which so many depend.
Reader Comments (7)
Take a look at history: a ruler once was rescued in a battle by one of his soldiers, who - other than a philosopher advised him - went up to his ruler for a reward: he lost his head as rulers do not like those who saw them in a week position, so: will IBM still like FOSS ? And: you have a perception of friendship very different from that of big business: in fact they (going strong again) do not know this term - it is for weenies. All they would know is competition, defeat and victory. Try to line up with the crowd out there so that the threat of bad publicity might improve the situation.
Sorry for the typo: weak/week. And I might add "...victory, sometimes compromise".
I am confused by the blurring of roles between "Hercules" the open source project, and "TurboHercules" the (presumably) for-profit entity for commercializing Hercules. I see IBM responded to the for-profit entity with their patent admonition; how does this represent an attack on the open source project? I don't recall any part of the IBM patent pledge pertaining to commercial for-profit entities.
The IBM press release talks about things like
andHowever, the pledge itself only refers to certain intellectual property that might be necessary to implement more than 150 standards designed to make software interoperable. The actual wording includes
So the pledge itself applies specifically to commercial for-profit entities. What it does not do is cover any and all open-source software; it only covers software implementing any of the specified standards. Now, I'm sure it would be trivial for Hercules to implement some code that specifically complies with XML1.1 (one of the listed standards) but that would only address two of the patents they're referring to in their claim against TurboHercules. I'm quite certain that IBM would be willing to withdraw their complains on those two specific patents, after all there's 104 others in the complaint.
Also worth noting is that TurboHercules filing a complaint with the EC does not appear to in any way be asserting any Necessary Claims against any Covered Implementations of IBM or of any third party - the complaint is based on monopolistic behaviour, not any assertion of IP violation on IBM's part.
Nice article, I was looking for these kind of information on Hercules mainframe emulator.
Please continue writing....
I enjoyed reading your blog. Keep it that way.
Please please THROW it all out ;; redefine jump on board;;;;;
z390.org implemented NEW System/Z z196 on x86 & x64 in JAVA on win & Linux or any other J2SE platform in Dec of 2010
When I implemented NEW System/Z z196 on x86 & x64 in z390 HLASM & zCOBOL w/DB2 Express-C in Early 2010
My CLOUD was deployed/tested z390 HLASM & zCOBOL NEW System/Z z196 on x86 on CloudShare Pro in Dec 2010/Jan 2011
Sort of THROWS all this OUT the DOOR does it not...
Next thing people will say they did IT