RUG talk notes emulator licensing, recovery
April 24, 2014
Second of two parts
When CAMUS held its recent user group conference call, MB Foster's Arnie Kwong had advice to offer the MANMAN and HP 3000 users about the CHARON emulator for PA-RISC systems like the 3000. A more complex environment than HP's decade-old 3000 hardware is in place to enable things like powerfail recovery while protecting data. And readying licenses for a move to the Stromasys CHARON 3000 emulator means you've got to talk to somebody, he said.
"Everybody is pretty helpful in trying to keep customers in a licensing move," Kwong said. "If anyone tells you that you don't even have to ask, and that you're just running a workalike, that would be a mistake. You have to have an open and fair conversation. Not doing so, and then having a software problem, could be a fairly awkward support conversation. You can't make the assumption you'll be able to make this move without any cost."
If you create secondary processing capacity through CHARON, you'll have to execute new licenses for those licenses. But most of the third party vendors are going to be pretty reasonable and rational. We've all known each other for decades. People who do lots of IT procurement understand straightforward rules for handling that.
Kwong said that CHARON prospects should make a catalog of their MPE software applications and utilities, and then talk to vendors about tech compatibility, too.
In manufacturing IT in particular, its cost has been declining recently. "Short of somebody paying $10-15 million to re-engineer around SAP, or Infor's other products, most of the incremental spending in the MANMAN and 3000 environments have been to extend life. People do a lot of stuff now on Excel spreadsheets and SQL Server databases around the ERP system. We look to see if the 3000 is the essential piece, and often it is. We look at what other things are affected if we change that 3000 piece."
Kwong said that MB Foster has not done MANMAN-specific testing against its in-lab CHARON installations yet.
Data integrity questions came up from Mike Hornsby, who wanted to know about comparison in using transactional testing to evaluate possible data loss. Of the HP 3000's powerfail environment, Kwong said, "it's been one of the key strengths of the 3000 environment in particular." The tests at MB Foster haven't revealed any data loss. Kwong didn't dismiss the possibility, however.
"This is theory, but I'll say this: One of the things you have at risk during the crash recovery process is either in the CHARON emulator, or the underlying infrastructure in the cloud environment that you're running it in." In this meaning of the word cloud, Kwong was referring to the VMware hosting that's common to the 3000 CHARON experience.
"In those instances you could have failures that were never in their wildest imaginations considered by the folks who built this software-hardware combination. I have not seen anything personally in our testing where things have been horrendously corrupted, rolled over and died. But inherently in the environments they're running, there are assumptions of database logfiles, and particularly in certain key files and so forth, where your warmstart processing can be at risk."
When such failures occur — and they can happen in HP's provided hardware — "You have the same predictability in an emulated environment as you do in the 3000 hardware environment. I don't think I'd lose a lot of sleep over it." However, networking and storage architecture issues are different for the emulated MPE hardware than for HP's native hardware, he added,
But application expenses take the forefront over hardware and platform issues at the sites where MB Foster has discussed transitions of any kind. "When you take the context where the 3000 is running from a business standpoint, yes, you have licensing issues for maintenance and so forth," Kwong said. "But as a total percentage of the cost to the enterprise, the application's value and the application's cost to change anything, usually begins to predominate.
"It's not the fact that you have no-cost terminals and low-cost hardware anymore, it's what that application's power brings you. We've seen that newer managers who come in from outside at these sites with stable HP applications have vastly different expectations for what the application's going to deliver — also, different demands for the applications portfolio — than people who've been there for decades running the same architecture. The platform discussions usually aren't major economic drivers.
"Running a 3000 application in another environment, such as Windows or Linux, is never zero, although it's cheaper to do that in a Stromasys environment. We need to carefully consider the hardware scalability performance availability, and certain kinds of communication and networking interfaces that aren't qualified for use in the Stromasys environment yet."
"We look at how to approach the problem of migration and its processes. In talking to our customers and concerns they have at small one-person shops with boxes running for 20 years, a move will take a year or two years to do. People that we talk to say they're gotten by for a long time without having to pay the kind of money needed to migrate to SAP or Oracle, or FMS or JD Edwards. Those alternatives are on the list of things they look at.
"Few people are talking about development stages for the kinds of complex environments the folks on this call represent. The days of large scale development have pretty much gone by the board. Everybody's talking about what kind of capacity they can buy, and what kind of features can they buy, rather than concentrate on what kinds of things they could move to the new environment.
"For them, the Stromasys approach says they'll leave their software base the same and go to new hardware, essentially. There are a lot of business assumptions and a lot of applications assumptions that might change because you're running in that new hardware environment. Things that were always based on the 7x24 capability, running without a lot of staff expense — all of those things are now open to question and rethink. We encourage people to take a step back and look at their business planning assumptions and business models, because that's the foundation for why they bought the 3000 in the first place".
Kwong he believes most of the users on the call could agree HP didn't do badly by them in the initial offering of high-value, investment-protected systems. Now that the system is into its second decade beyond HP's exit announcement, protecting that value deserves some fresh assessment.