Solitary backup tape spurs fresh MPE plea
A Hybrid Solution to Staying and Going

What's MPE got to do with emulators?

Thoroughbred-horsesCompanies that want to use their MPE/iX applications a long time might count their timelines with two eras: Before Emulator, and After Emulator. The B.E. period left the MPE/iX user locked to Hewlett-Packard hardware and waiting for upgrades to HP boxes. The A.E. era uses virtualization via Charon to permit many beefy Intel boxes do the MPE/iX work. But what does MPE/iX code have to do with the magic of Charon? Not much, which is a good thing.

There's a stubborn story we hear about how the gem of MPE's source code is at the heart of what Charon does. What a virtualization engine like the Stromasys product delivers is a new capability for Intel hardware. An Intel box can pretend to be a PA-RISC processor, thanks to the software engineered by the creators of similar products for the Digital market.

But Charon doesn't rely on MPE/iX secrets to do this magic. It's like thinking a jockey is the being who's running a 2-mile racetrack course. He's the rider, and the horse in this metaphor is Charon. The basic design of Charon products, like the ones that virtualize the Sun Sparc systems and the PDP systems of DEC, creates the expertise for booting up Intels like they're 3000s. Nobody expects the ancestry of the jockey to play a role in making the horse faster. We don't sit in the grandstands to watch jockeys hoof it around the track.

Source code for MPE/iX was a big part of the push for emulation that started in 2002. The thought at the time, though, was that MPE itself could be taken on as a third party tool. That somehow something could pretend to be MPE, instead of pretending to be the PA-RISC hardware. Emulation is a broad term. Virtualization is more accurate, and these days, almost as well known thanks to titans like VMware.

OpenMPE fought the good battle to get HP to release source code for MPE/iX. It won, but the group never got a copy of the source for itself. The reasons run to finances and organization of the board of directors, struggles we've documented starting six years ago. Then 2011 dawned as one of the dimmest years for the 3000, just dumped off HP's support and with no apparent future for fresher hardware. The only advocacy group was reeling from lawsuits among its leaders. For the last two elections of its board, the number of nominees was the same as the number of directors. Few people could volunteer with gusto. Some of the best were already there, or had already done their bit.

It's a good thing indeed that MPE/iX source code plays no role in the technology of Charon. The code is not open source, as OpenMPE wanted. (That's the whole reason it's called OpenMPE, instead of FreePA-RISC.) What the community got was a license from HP, not freeware. If Charon needed MPE/iX gems, it would mean HP retains control over a product that will let newer hardware take over when the battleship-grade HP iron hits the scrapyard. Nobody wants Hewlett-Packard to have a say in MPE/iX host hardware futures. The vendor's had enough impact on those. In the days of A.E. the virtualization miracle needs nothing from MPE/iX but a stable copy of the OS to carry in its saddle. The horsepower comes from an independent company, the kind that's kept the MPE/iX journey on course.

Comments