June 27, 2013
Backing up proves an emulator just works
Proving the concept of emulation for MPE operations is becoming popular this year. To offer evidence, longtime managers of 3000 servers check out the mundane as well as the specific tasks that drive their companies. Backup is a backbone of real IT -- and one evaluator shared his pleasure in watching the Stromasys CHARON HPA/3000 product improve on such an essential mission.
The process is somewhat different than on a physical HP 3000. First off, you can do backups while people are still on the 3000, if you have backup software to support that. When you configure the emulator, you specify a virtual tape drive, similar to the way you specify the virtual disc drives, with each virtual device pointing to a file in the Linux environment. Then, when you run MPE's STORE command, CHARON puts the data in the file associated with that virtual tape drive. When the backup is done, you can copy that file (using standard Linux commands) to some other backup media for archival.
One very nice thing I found is that CHARON doesn't ever run out of 'tape' on a backup. It just keeps growing the file as needed. When I configured our emulator environment, I configured the tape drive at 8GB, thinking that would be enough. However, when I finished the software install and had copied our test data, I had about 10GB worth. When I did the full system store, Charon successfully backed up everything and expanded the virtual tape drive size to be 10GB.
Later, when I did just a SYSGEN to the virtual tape drive, the file was only 5GB. No more having to worry about what tape density you're using -- and no more getting the 'please insert next tape' message on a backup.
Backup is just the latest example of "it just works," the motto that the emulator prospects come away with once they're done with a proof of concept. A serious number of them will be using the product to extend the life of MPE applications that are destined for replacement. Until that day, everything has to be backed up. Of course, the real test of any backup process is to restore your data.
To do an MPE restore, you find the Linux file that corresponds to the backup tape you desire, copy it back to the emulator directory (using the appropriate Linux commands), name it the same as the virtual tape drive you configured in Charon, and do an MPE restore. Charon reads the file as if it were a backup tape and finds the appropriate MPE file to restore. It works the same way for SYSGens, except that it creates a bootable image that you can boot from later.
At this stage, everything works as expected with CHARON and its backups. "You don't need to 'pre-build' the tape file before a STORE," says product manager Paul Taffel. "CHARON rewrites any pre-existing file before starting a Store, and the file will then grow as large as needed. The :DEVCTRL command must be used to put a virtual tape online before any STORE or RESTORE operation."
The 3000 manager who was proving the emulator concept was satisfied. But it will be later in the year before the emulator takes over the 3000 hardware's work.
"All in all, we were pleased with what we saw," he said. "But when an internal project needed more resources, I was pulled off all of the other projects I was working on, including the CHARON testing, to devote 90 percent of my time to this other project. We were almost ready to begin the procurement process, once I had verified that PowerHouse Web worked. I hope to resume testing when my current project is finished."
June 26, 2013
Steps for Doing a Final HP 3000 Shutdown
There might be another HP 3000 in your company's future, but at some point every system, homesteaded or migrated, will have to be shut down. Here's the steps to do the job professionally.
A customer asked how to turn off an HP 3000 once and for all. While this is a sad time for the IT expert who's built a career on MPE knowledge, doing a shutdown by the numbers is in keeping with the rest of the professional skill-set you can expect from a 3000 manager. I am reminded of the line from Citizen Kane. "Then, as it must for every man, death came to Charles Foster Kane." Nothing escapes death, but a proper burial seems in order for such a legendary system.
Chris Bartram, who stocks a Technical Wikipedia (TWiki) for the 3000, offered all the details of turning off an HP 3000. "I have performed last rites for a 9x8 server at a customer site," he says, "and have been through the exercise a couple times before."
His steps did not include SOX requirements, but "that might be useful," he said in his usual modest introduction. There are 10 steps Bartram details before switching off the 3000's power button.
Bartram reported that he first purged all accounts except sys, hpspool, and 3000devs (and had to log off all jobs, shut down the network, and disable system UDCs to do that). Then:
2) Reset/blanked all system passwords (groups, users, accounts)
3) Purged all groups from SYS account that I could (aside from in-use groups) as well as all users except MANAGER.SYS,OPERATOR.SYS, MANAGER.HPSPOOL.
4) Went through PUB.SYS listf (file by file) looking for anything that might be a job stream or contain user data (or anything not critical to keeping the system up) and PURGEd it
5) Went into VOLUTIL and condensed my discs
6) Created a group called JUNK.SYS (you would need to do this on each volset; this box only had the system vol set)
7) Wrote and ran a short script that copied NL.PUB.SYS (the largest file remaining on the system) into JUNK.SYS in a loop using filenames A####### and X####### until all disc space was used up
8) Typed the command PURGEGROUP JUNK.SYS
9) Went into NMMGR and changed IP addresses on the box to something bland/different; including the default gateway (also deleted any entries in the NS directory if there are any)
10) Sequentially PURGE @.GROUP.ACCT for all groups (leaving PUB.SYS until last)
11) Shut down the box.
June 25, 2013
Hiring developers who are old is new again
Migration is the same as legacy modernization when it comes to its end result. That's change, even if the applications in the 3000 world still look and act just as they did on an HP 3000. Migration sounds more drastic because it describes the transition of apps from one platform to another. Modernization -- especially in the hands of services companies -- takes smaller steps but still wants to shift operations toward something more popular, current, and easier to hire for.
However, that ease can become a disappointment if the only goal is to hire newer and younger programmers who work cheaper. A recent study showed that the old programmer is not only a better value, but now in shorter supply.
Bruce Hobbs, a veteran 3000 developer, pointed out the article in IT World which said, "Like a fine wine, programmers get better with age."
Researchers from the computer science department at North Carolina State University have released a study in which they examined whether programming knowledge gets better with age. Specifically, they used data on over 84,000 members of the Stack Overflow website community: the questions they ask and answer in that forum, and the site reputations for each user as proxies for the general population of programmers and their level of programming knowledge.
Does age have a positive effect on programming knowledge?
Do older programmers possess a wider variety of technologies and skills?
To what degree do older programmers learn new technologies?
3000 managers who are planning for the future know it's not easy to find a senior programmer. "I'll be looking for a couple of experienced HP 3000 MPE resources very soon," said one IT director recently, "and I know they won't be easy to find. Been there and done that."
At the Stack Overflow site, younger programmers demonstrated a shorter range of knowledge, asked and answered questions about a narrower set of topics, and even scored lower than programmers in their 30s about nouveau topics such as iOS and Windows Phone 7.
Based on all this, one can conclude that as programmers get older, they get better; they know more about more programming topics, and they learn new technologies just as well if not better, than their younger counterparts. Take that, whippersnappers!
This is a development, so to speak, that runs counter to one of the driving mantras of migration and modernization: older technical choices, and the human resources that understand them, are more costly, because these programmers are harder to find. As it turns out, the value in a programmer is correlated with knowledge rather than age. But the gurus at places like Gartner are delivering a different message.In a briefing on how IT changed after the economic downturn of 2008, VP Dale Vecchio advised IT managers to control costs by looking at a calendar of birthdays.
Organizations are dependent on an aging workforce to deliver their applications. It’s become one of the single biggest drivers we’ve seen. We recommend that you ask HR to provide a chart of retirement dates for specific job titles.Tell them, ”I don’t care who the person is; just let me understand when these retirements are likely to occur." It’s about managing this skills challenge, managing the retirements of Baby Boomers.
Understand, Vecchio isn't crazy enough to presume that the technologies running modern business -- tech that's anything but nouveau -- should be replaced. No, COBOL will always lead business tech choices, it seems, at least in enteprise settings. But local schools should be enouraged to train young programmers in these elder skills.
We believe organizations must engage with the secondary education institutions, to help support these declining skills where necessary. You need to tell those institutions, “If you train ‘em, we’ll hire ‘em.”
It might be easier to hire younger IT pros, but that won't make them as productive or as experienced as the older programmer who's becoming harder to find. All programmers seem to become elusive after 50, not just those schooled in MPE skills. Vecchio suggested that the solution is to procure for the younger programmer a better development toolkit. "There are development environments to help improve the productivity of your existing workforce — to help you manage more change with potentially fewer people as they inevitably retire."
But the applications are not retiring soon enough to make a difference. Even HP-branded 3000 hardware is being purchased to keep the apps running. At one marketing company in the Northeast, "we are moving to an HP 3000 N4000-400-750 box, which is being built with a XP12000 disk array subsystem. Our backup HP 3000 will be the N4000-400-500 with a XP12000 disk array subsytem -- which is our current production machine."
The applications in place on systems which are modernization targets are best understood by older programmers. Not because they were on hand to document the building of the apps. The wisdom of interviewing users and developing to needs is difficult to replicate without the years of experience. If you see an older programmer available, sieze on the chance to employ them. The US Bureau of Labor Statistics shows the median age of a programmer at about 42 years old.wrote the articles for IT World, sums it up thusly: "Just because a guy finds himself getting up more in the middle of the night to go to the bathroom doesn’t mean he can’t still knock out a killer iPhone app for you. He just made need to take a few naps along the way."
June 24, 2013
Open source resource: Secure FTP on 3000
Even though FTP won't help much in transferring databases on an HP 3000, a lot of other data can be moved using File Transfer Protocol. The question of how to do this securely using SFTP just came up last week. We've covered the topic before, but a new contributor, Brian Edminster of Applied Technologies, chipped in with some advice and a new resource, built from open source.
The initial question:
I'm trying to use ftp.arpa.sys to FTP a file to a SFTP server and it just hangs. Is there a way to do a secure FTP from the HP 3000?
Brian Edminster replies:
The reason that using MPE's FTP client (ftp.arpa.sys) fails is because as similar as they sound, FTP and SFTP are VERY different animals. Fortunately, there is a SFTP client available for the 3000 -- the byproduct of work by Ken Hirsh and others.
It used to be hosted on Ken's account on Invent3K, but when that server was taken out of service, so was Ken's account. As you've no doubt already noticed, it's available from a number of sources (such as Allegro). I'd like to highlight another source: www.MPE-OpenSource.org
Edminster goes on to explain he administers that site, as well as puts together the 'pre-packaged' install available there. It's in a single store-to-disc file in Reflection 'WRQ' format, making it easy for the majority of sites to retrieve and use. Here's the URL:
I have a customer that's been using SFTP daily as part of their PCI compliance solution for several years. They push and pull data hourly from dozens of Point-of-Sale systems all over the country, and have moved lots of data this way.
The biggest caveat from that customer's implementation is that if you're moving data over a WAN, SFTP seems to be more sensitive to jitter and latency issues than conventional FTP. We ended up having to upgrade a couple of their more anemic 'last mile' circuits to accomodate that.
In all other respects, it’s quite a robust solution, and can be tightly integrated with existing legacy apps. I know; I've done it.
If you have any questions about how to use the pre-packaged install -- or how to get around any limitations you might run into,-- don't hesitate to contact me. I've used this on dozens of systems over the last decade, and have transferred many, many gigabytes of data with it.
June 21, 2013
How to Transfer Data Files on a 3000
It's a Friday and perhaps time to do some deeper maintenance of your HP 3000. A weekend offers an opportunity to move data from an older system to a newer replacement for homesteaders. Here's some advice on using FTP to make a data migration.
You can use FTP to ensure a safe backup of 3000 data. In one case a manager said he needed to backup KSAM XL files, but the manager didn't know that his FAK files were HP's special Keyed Sequential Access Method database files. "What appears to be program files are moved over," he said, "but database files get left behind. How do I get these files over to our Windows server?"
Please note, IMAGE has priviledge protections which don't exist for KSAM. Purging is controlled, for example. KSAM XL files are really just files combining indexes and data, and can be purged without regard to priviledges. Nonetheless, this 3000 uses KSAM XL is running MPE/iX 6.0 -- so there's more than just some management experience missing from this server. 6.0 is more than 13 years old.
MPEX, Vladimir Volokh reminded us, performs more powerful operations to go beyond HP's security. For example, an SM user can do a COPY @.DATA.PROD, @.DATA.BACKUP. Copying only KSAM files: COPY @.DATA.PROD (CODE=""KSAM""). MPE cannot give you such ability to copy only KSAM files.
One rule of 3000 operations is that these data-containing files act differently than all others while being transfered. So FTPing them to a Windows 2003 server won't be a successful way to ensure a safe data recovery. (There are third-party tools to do this, updated and supported in a way that STORE or HP's TurboStore subsystems will never be supported again.) But if a 3000 is stuck on 6.0, it's probably going to have only enough budget to tap included HP software for file backups and transfers.
Donna Hofmeister, whose resume helping 3000 users occupies a vast chunk of the 3000 newsgroup archives, suggests starting with mystd to store the files to disk -- then transfer the STD file."First try this little experiment," Hofmeister began in her answer.
if it works, you just stored/archived all the files that begin with 'a' in the .pub group of the .sys account into a file called 'mystd' (my store-to-disc). You can expand the number of files being stored into your STD file by modifying your store command to:
@.pub.sys -- all files in .pub.sys
@.@.sys -- all files in .sys
@.@.acct -- all files in .acct (for example)
@.@.@ -- all files on the system (and it's actually better to say '/' instead of '@.@.@')
Keep in mind that as you increase the scope of what your storing, so does the size of your STD file. In other words, to store the whole system you need 50 percent or more free space , which you probably don't have. So, break what you're storing into chunks (do one account at a time) and things should go smoothly.
If STD doesn't work, you might be able to get tar to work. The same space precautions apply. One advantage of using tar is you should be able to verify the tar file on the destination system -- something you can't do with STORE without a 3000 in the mix.
Chris Bartram, who still operates an information treasure-house in 3k.com, added explicit advice about FTP and the 3000's files.
If you package all the MPE files up in either a store-to-disk (aka std) or a tar "wrapper" (disc file) you can transport that file around at will -- as a BINARY file -- don't try to transfer it as ASCII or CR/LF translations will trash it.
Once you get it back on the 3000, a simple file equation directing your source (tape) to the new (std) file name (and add the ;DEV=DISC to the file equation) will allow you to restore the files onto the 3000 preserving all the MPE specific file attributes they started with. Tar will work similarly for almost all MPE files, but can't handle database/PRIV files and probably not MSG (message) files and a few other very MPE-specific files.
tar works the same way on MPE as on *nix boxes. But is much more "familiar" if you run it from the Posix shell (sh.hpbin.sys), though that's not necessary. Treat the tar file just like you would on a *nix box.
For store-to-disc files, you use the same MPE syntax for storing files as you do normally; the only difference is that the output device is file-equated to ;dev=disc. As mentioned, be aware of the disc space required to store another copy of your backed-up files online.
Likewise, when you restore instead of pointing to tape, you point to the disc file -- and don't forget to add ;dev=disc to that file equation as well. If the store-to-disc files are going to be very large (several gigabytes) you can use some additional syntax to break them into chunks - but hopefully you needn't worry about that for now.
Treat a store-to-disc file just like a tar file. Record size and most other attributes aren't so critical, but if you move it around do NOT let FTP transfer it in ASCII mode or it will corrupt the file.
As for examples; I back my primary HP 3000 up to a disc file then transfer (FTP) it to a Linux server. Here's the gist of my JCL:
!STORE / - /BACKUP/ ;*T &
user userid password
put FULLB.PUB.BACKUP hp3000-full
June 20, 2013
How to Back Up On An Emulated 3000
Stromasys product manager Paul Taffel is reading our reports closely (always a good idea, to keep up with the latest in your community.) Upon reviewing yesterday's story -- which included an account of the operations a prospective user said he still needs to investigate -- Taffel quickly offered answers to a key question about backups using the CHARON HPA/3000 product.
The 3000 manager, who's investigating emulation en route to a migration, asked
How do we do backups and restores with the emulator? Its architecture is that each HP 3000 LDEV is a separate Linux file, so identifying where MPE files are for backup and restore looks more difficult. If I need to restore a file to one of our production accounts, how do I identify which Linux backup it is on, and how do I then mount that virtual disk to do an MPE restore?
Copying entire CHARON disk image files is only useful as a way to back up an entire disk (or collection of disks) for Disaster Recovery purposes. To create a backup that is useful for restoring individual files, use the MPE STORE command (or a third-party tool) to write the backup, either to a physical tape (internal SCSI, or external USB drive), or to a virtual tape image. Virtual tape images are stored as files on the Linux server; they can hold the contents of any Store (or SLT) tape, but take far less time to write.
The technique of copying an entire CHARON disk image file is useful when moving an entire virtual HP 3000 to another server, or returning the local server to a prior state. You can, for example, back up your system, conduct a test of your month-end processing, and then return your system to its original state, quickly and easily, by copying the Linux folder containing your disk images. One important note: the virtual CHARON HP 3000 must be shut down before copying the virtual disk image files.
June 19, 2013
Operations and applications get watched and tracked in emulation efforts
While explaining what a virtualized 3000 does with its MPE bootup volume disk image, questions come to mind. A systems manager will be asking about the following, since they're probably unfamiliar with tapping an MPE system instance which is part of a Linux environment. Here's a set of queries from a prospect who was working though proof of concept this spring. He is preparing to use the Charon emulator as a migration stopgap.
How do we do backups and restores with the emulator? Its architecture is that each HP 3000 LDEV is a separate Linux file, so identifying where MPE files are for backup and restore looks more difficult. For example, I have configured an 18-GB virtual disk drive as LDEV 32, so in the Linux directory where the emulator resides is an 18-gig file named 'LDEV32.DSK'. All of the MPE files stored on LDEV 32 are in that file. If I need to restore a file to pub.admin (one of our production accounts), how do I identify which Linux backup it is on, and how do I then mount that virtual disk to do an MPE restore?
This is an HP 3000 administrator with some applications which have already been moved to other host environments. Not a pro who's unfamiliar with Unix or Linux. He allows that there are "lots of questions that I'll have to work through, operationally." It's such operational questions that define the legend of building a datacenter around a general-purpose computer like the HP 3000 -- one designed to operate as if it had to be reliable enough to be installed in a satellite.
Applications, languages and utilities are coming on board in such emulated environments for the 3000. Some of these vendors must be contacted directly by the customer. For example, Nobix's Transpooler that manages jobstream operations will be part of one manager's emulation configuration. Could that manager do without the Nobix software?
The answer is "probably not." They are jobstream related -- scheduling, after execution error examination, and so forth. The CSL Sleeper utility [from Boeing] can handle some of the scheduling, but it's not as flexible as the Nobix Transpooler product. Also,the product is better at sending spoolfiles to printers than plain MPE.
The ability to re-send spoolfiles, delete them and otherwise manage them, without the use of MPE spooler commands, is very useful to us. We would probably not be able to go forward without it -- at least not without dedicating a lot more resources (personnel and time) toward developing a workaround.
Stromasys is promoting the idea that companies like Nobix would rather transfer a license and keep a support contract than see a customer disappear. This is all up to the emulator customer to arrange. But the truth of it is, some vendors might believe they are certain to be part of an emulator setup, and they might hold out for an upgrade fee. People suspect Cognos will be in this group, but the reports from customers are surprising. Cognos/IBM has made a tidy living doing that sort of re-license over the last 20 years. Powerhouse for MPE is on "Vintage Support" by now, but the real money is in a license upgrade fee.
"They have been very gracious in this," one manager said. "As of PowerHouse 8.49F, IBM removed licensing requirements on the MPE version of the product."
Unfortunately, we allowed our PowerHouse license to lapse when we were on PowerHouse 8.49E, so we missed that feature. We let it lapse to save on the annual maintenance fee, which for the N-4000 box with unlimited users was several thousand dollars annually. For testing, IBM gave us a 'universal' license, which will work on any HP 3000 box. I haven't asked, yet, if there will be a charge when we purchase the emulator.
Our optimism has mostly to do with the way Stromasys has implemented the emulation. It's an elegant solution because they've emulated the HPPA chip in software, so MPE thinks it's running on regular HP 3000 hardware. I am very impressed with that. It behaves just like a physical HP 3000, in terms of booting and system management. All of the tools are there and work: sysgen, ioconfig, nmmgr, and more. I was able pretty much to have the emulator in a Reflection window side-by-side with our HP 3000, so that I could make sure nmmgr values were similar for network config.
June 18, 2013
How infrastructure survives heated times
Over the past 24 hours I feel like I've been living the work life of a 3000 IT manager. We've had telecomm outages here, the kind that can mean lost business if it were not for backup strategies. Unlike the best of you, we don't have a formal plan to pass along in a disaster. Today's not really a disaster, unless you count the after-hours pleasure we hope to savor from Spurs basketball.
In a lock-down IT design, writing captures what to do when a telecomm service winks out dark. Our broadband provider is ATT, with an 800-number repair line to call. We poked at that twice today for one of our landlines, now without a dial tone since yesterday afternoon. There's a different repair number for the Uverse Internet service -- and also the world of IP everything else, since our downed data line means not only no fast Web, but no San Antonio Spurs NBA Finals basketball in about 2 hours or so.
Consolidation to a single provider promises savings, but also a single point of failure. Coordinating service between two arms of the same company? Well, that's not an automatic thing anymore. Meanwhile, the cloud-based IT promised by HP and others just pulls all of this recovery farther away from your affected IT shop.
About 10 days ago, MB Foster gave a thorough primer on the issues any company faces in keeping its disaster recovery process up to date. There's old tech (phone trees to spread the word on outages) as well as new elements like measuring the Mean Time To Recovery of Operations. MRRTO can help you decide where to put the effort first in a downtime event. Foster can help you ready for the calamity with a thorough inventory of what's running, something that CEO Birket Foster says too many companies just don't have up to date.
"You look at the different processes in your company and figure what's critical to keeping the business alive," Foster said in a June 5 Wednesday Webinar. "It comes down to understanding if there's a cluster of applications which work together, so you have to bring them all up together at the same time," he said. A DR plan must identify key users -- old tech, like keeping up to date with user cell phone numbers, so they can be notified.
"Hardware is usually not the problem here," Foster said. "That said, there was a vendor in the HP 3000 community who had a board go bad on their 3000. It took them 13 days to get the other board in and back up, and then into recovery. It was mostly about sourcing the right part. They didn't have good connections in that area." Then there was also the matter of getting competent resources to install the board.
Tomorrow MB Foster offers another Webinar, since it's a Wednesday. Gods of Data Quality examines Master Data Management (register for free), the MDM that "ensures your company does not use multiple – or potentially inconsistent - versions of data in different parts of its operations; understanding the concept of 'one version of the truth.' "
Each one of these Webinars gives me plenty to think about and try to plan for.We're feeling some pain today in our little micro-sized shop, but it hasn't cost us business up to now. We're done what Foster advises: knowing what is running in your system lineup through an inventory. but that knowledge is in my head today, and if I swerved to avoid a texting driver and got myself the ER, my partner or a backupn helper wouldn't know how to deliver this news story to you, even if I'd written it in advance. What do you do when your broadband pipe goes down and stays down for awhile?
"This is a business problem, not an IT problem," Foster explained. The trenches-level repairs are on the IT lines, but the stakes are up at the boardroom level and in the finance officer's purview. That increases the pressure on IT, especially if the economies of curtailing support have been demanded from the CIO or CEO. In a personal example, just last week I toted up savings of dropping a hot-spot wireless feature on my mobile phone account. It's there when Wi-Fi can't do the job. It seemed costly at $25 monthly on a micro-business budget. Hot-spot I'd only used outside our offices on travel could be cut out, right? To pay more more crucial IT services, like website renovations. There's always something.
Except that for the last 24 hours, that hotspot off an iPhone 4GS has kept the Newswire's email and Web blog services online, right here in our offices. (It's not effective to have to go to a coffee shop to do secure Web work, but it's better than nothing.) Have you been forced to economize, debating over dropping a service contract or support agreement you rarely use? Or been told to drop? The finesse is in keeping these DR lifelines intact, ready for the day of disaster. The more you know in a formal plan, the more professional your respose looks to the executives in charge.ATT brings everything into our offices now. 25 percent of our email, and all through their lines. 100 percent of the bandwidth for everything on a wire, including the TV. Our landline numbers, the ones which rarely ring anymore in the era of email but always can open our door to new business. 512-331-0075 has been in the public eye so long that a transition to a cell-only number seems unthinkable. We pay for extra support and maintenance on these relics -- our headquarters is smack in the middle of some of the oldest and messiest copper in Northwest Austin.
As I write, the second ATT truck of the afternoon cruises our street. Matt (they all have names you should use) is unsnarling and fixing a network pedestal at the property next door. This hub controls our telecomm and that of a half-dozen other addresses in the area.
I'd call these residential issues -- our office is in the midst of a a stately 40-year-old neighborhood in one of Austin's oldest high tech corridors. But when I register our trouble ticket for the phone llne, ATT says in its recording we are a Major Business Account. I don't question that designation, because it gets us to the head of the line with a human being. Broadband service, sadly, doesn't enjoy this distinction. ATT considers us consumer-grade customers, even as we work with an 18GBit download speed.
Take this checklist and answer honestly to see how much you must do to survive calamity.
- Did you recently cancel support for software still crucial to the business, but now on a "declining" platform of the 3000?
- Is your support provider working within a Service Level Agreement -- so you know how much the "increasing impact of a system costs" after an outage of one hour, or four, or 8 or 12 or a day or a week? What's the pain and cost of each of these downtime periods?
- When you place a support call, how soon to talk with an agent, human being or expert on your system?
- Do you have redundant hardware in place for when a computer does offline -- and is it hot-standby, or not?
Perhaps most importantly, how long has it been since your DR plan has been tested? By a test, I don't mean the last time you needed it to work. Those reports are costly. This is a controlled event that yields a lot of documentation on the success of your DR-MTTRO plans. Foster pointed this out
Here at the Newswire we're light on our docuementation. I could write out for my partner how we recover from calamity internally -- the locations of our backups, the process to restore, the way to transfer a full backup onto reserve hardware. Who we call when we cannot resolve it ourselves. How the telecomm is supposed to work. We have religion to do that today, but you can't just drop that kind of information into the hands of your best sales person, chief muse and dreamer, or even a veteran office manager who's unfamiliar with the fundamentals of problem resolution.
This can happen inside a 3000 shop, one with other environments like Linux and Windows at work. Our partner and friend Alan Yeo had a UPS calamity with his power last month, and it was five days before the affected 3000 went back into service. This is an organization with more than 30 years of 3000 and IT background that presumed a UPS could keep a system online -- instead of permit the server to be fried, while other computers all around escaped that fate.
And so, Alan is preparing an article entitled, "Do you want fries with that?" in his set of cautions. Electricity is about the only essential service that hasn't rolled over on us over the last week. Without it there's the coffee shop, alternative business allies nearby (like our friend Candace's personal coaching service). We called her as a backup to the Spurs game tonight, too -- just before ATT's broadband repair succeeded after six hours of heroic effort.
June 17, 2013
Emulator: how far it goes, and what's next
Even among the potential allies for the Stromasys emulator, uncertainty is afoot. I had a conversation with a reseller last week about the product, and he was not sure that IMAGE was a part of the solution. People approach the Charon emulator from their best-known persepective, and in most cases that’s MPE/iX and its database. Good news: Charon doesn’t emulate any of that software. It simply uses what Hewlett-Packard created and installed on everyone's 3000.
Instead of fooling with the 3000's software, the Charon product provides a pre-configured MPE/iX disk image. This is a system disk (your LDEV1), but it’s not a physical device. It’s an virtualized disk file, running on a Linux server, which the emulator then reads when it boots up MPE/iX. Once you have this LDEV1, you populate it with the software on your 3000 system -- specialized databases, configurations for IO, the works. The wizardry comes in making an Intel server which runs Linux -- the host OS of the emulator package -- behave the same as an HP 3000 server. MPE/iX is changed in no way. This is why there've been no lingering reports of the emulator failing to run an MPE application or a utility.
Emulator technology has a reputation from more than a decade ago of being a horsepower hog. But the first two generations of emulation have blown past us all, and now Stromasys is beyond instruction-by-instruction interpretation. It’s well past dynamic instruction translation, which pre-fetched a platform’s CPU instructions, then translated them into target platform code. That translation might have been called dynamic, but it was only suitable for entry-level to midrange systems.
Stromasys has left all of MPE and IMAGE’s software stack alone, and patched nothing. The product’s task is to use the latest, multi-level translation technology. Stromasys perfected this technology — the third generation of emulation — while it served users of the VAX hardware who wanted to continue to run OpenVMS after HP-Digital stopped selling VAXes.The mission for Stromasys and its Charon products is behaving like the hardware abandoned by vendors, but not abandoned by customers with mission-critical requirements. In short, if your software is running on an HP 3000 today, it will run on Charon-HPA/3000 unchanged. Databases operate intact and as expected.
Stromasys GM Bill Driest explained that this third generation emulation takes the 3000 source hardware instruction code, then moves it to a specially developed intermediate language code that’s optimized for cross-platform virtualization. Finally, it’s translated to target platform code to let Intel’s broad, standard family of Xeon processors do the HP PA-RISC work that happened inside 3000s.
What this means is that the software doesn’t have to evolve to increase its performance. It’s a good thing, because MPE/iX is not going to evolve beyond its current rock-solid release. HP won’t permit the source code holders to create new versions of MPE. IMAGE isn’t getting new functions. Believe or not, that’s a good thing. Nothing was broken with MPE or IMAGE except for HP’s model to sell the software at the heart of a 3000.
Instead, Driest says that Charon will rely on hardware improvements to get to the next level of performance.
I’ve been in the industry 30-plus years, and the industry has learned there’s one thing to do when there’s new technology that comes out. You chase it. You migrate, you port, you replace. We’ve turned that paradigm completely upside down. Why should we always modify the software to take advantage of hardware innovations? Why can’t we adapt the hardware to fit our applications? What you’ve done up to now is create the exact same system you had before, but on a new box. We think there are other options here.
Namely, that’s to use the increasing horsepower of Intel’s designs — the ones driven by commodity markets — to employ additional cores in processors and lift up MPE/iX performance. Soon enough there will be Charon models to match performance of the biggest 3000 HP ever sold. Eventually the rise of hardware power will take this OS faster than HP ever could.
But Driest recognizes that Charon itself has evolution in front of it. “Some of this emulator technology should become self-aware, so the emulator decides, ‘I know what I need from this hardware I’m hosted upon. Why don’t I carve out the amount of memory I need from the new host platform, and give it to MPE. No need for having the level of expertise to do that level of maintenance. And where are monitoring and reporting tools? They’re all around the place, but they’re not inside our product.
Stromasys has plans, Driest said, to enhance its virtualization-emulation products with all of that. In the meantime, however, the company could use some introductions to customers. Stromasys and the community can benefit from having the same people entrusted with MPE/iX system support to guide a 3000 site into the world of virtualization.
Those are pros like the reseller who asked about IMAGE being a part of Charon HPA/3000. They know where the prospects are who use use 3000s. They’ve become trusted advisors in this independent era of transition. While Charon still requires expertise for its implementation, these resellers and support companies are the next place the solution needs to go. Technical leaps are important, but virtualization needs to cross a threshold of trust.
June 14, 2013
How 8 Years of Web Reports Changed Lives
This week the Newswire celebrates the 8-year-mark on our blog reporting. Starting with a eulogy for fallen 3000 savant Bruce Toback -- taken too early, by a heart attack -- we wrote about the nascent and uncertain era of transition in June of 2005. The Interex HP conference was still a possibility, HP was still creating some patches for MPE/iX -- many things that had gone on for years continued to roll along.
IMAGE jumbo datasets were supposed to get eclipsed by LargeFile datasets. HP was fixing a critical bug in LFDS and needed beta testers, something that was harder to come by for HP. LargeFiles remain less robust than jumbos for most customers. LFDS repairs consumed precious resoures in the database lab, all while HP tried to fix a data corruption problem.
HP sold off more than 400 acres in South Texas as layoffs started to mount up. CEO Mark Hurd set aside $236 million in severance pay. Sun offered up a open source program for Solaris, begging the question about when open source practices could be applied to MPE/iX. This week OpenVMS managers examined what stood in the way of VMS becoming open source.
Even though parts of MPE/iX are well outside of HP's labs, the whole wooly bunch of source, millions of lines, isn't a candidate for open source like the Sun project. But it might be, someday.
We looked at whether a transition era demanded the same rigorous HP testing of beta enhancements and patches. "We heard HP say they'd be satisfied with one site's beta test report, a comment offered when HP engineers discussed the lack of beta-test sites last summer at HP World." we reported. "When the labs closed in 2008, software that languished in Patch Jail was bailed out. HP was seeking beta testers "who want to try out the new networked printing enhancements for the HP 3000."June of 2005 began the period when HP said it would decide about making MPE source available outside of HP. "No source means no more patches. Is that a problem? Steve Suraci of Pivital Solutions, a third party supporting 3000s, talked about this in 2004. "Can we find workarounds?," Suraci said. "Almost always. We haven’t run into a situation yet where we haven’t been able to get a customer back up and running."
More than three years later, Suraci's shop licensed MPE/iX source code to produce patches and workarounds. Six other licensees got what the Sun customers seemed to want: source. Gilles Schipper of GSA, one of the longest-tenured support providers, said HP code was not key.
"I don't think [access to HP's source] is a necessary thing for the 3000 to maintain its reliability," Schipper told us. "I'd like to see it happen, because it may allay the concerns of some customers out there."
That's a lesson that the OpenVMS customer might embrace, with all the direct talk of parts of the OS built by third parties or created in HP Labs.
We've been so grateful for the support of the community, especially our sponsors, in keeping the blog the leading source of HP 3000 community news and experience. Thanks for making us a pick to click in the 3000 world since before the days of a user group bankruptcy.
June 13, 2013
The Last 5 Percent, and Toughest Surprises
For more than 10 weeks now, the offices and headquarters of the 3000 Newswire have been under siege. We're doing battle with a project, the largest one that Abby and I have ever faced together -- since launching the Newswire, at least. We're re-engineering our home with a whole-house remodel. This afternoon, our general contractor said "Well, you're basically rebuilding your home, except for two rooms."
It started more simply, of course. A little patch of the house, which serves as a yoga studio and writing classroom, had its older wood floor ruined in a flooding rain nearly two years ago. The rot crept in and so it became time to replace it. This might be the kind of "it's broken" event that triggered a migration or two -- back in the days when living in the comfy house of MPE built by HP was still an option. When Hewlett-Packard tossed its pail of water onto the floor of your community, the vendor started off its campaign with success stories about transitions.
Sadly, these were as disingenuous as the ones HP offered to VMS customers this week. At the same time it was tolling the bell for its VMS business, HP's added that OpenVMS is great enough to power Accuweather and the Singapore Stock Exchange. Except the vendor doesn't want Accuweather to run on OpenVMS more than seven years longer -- because the OS support is not scheduled to survive beyond 2020, according to HP's decree. Perhaps just long enough to collect support money for another seven years. After one VMS user noted that Accuweather was an old story placed in odd context, said Neil Rieck,
We heard similar stories in 2002 about the likes of Ceridian and even Summit Technologies' Spectrum credit union software. Neither company was on the timeline of the 3000 community, but yes, they had done migrations. One company started six years before HP's announcement, and the other began more than two years earlier and then didn't finish for another three years. Meanwhile, 3000 server sales continued apace. People bought new HP 3000s even after HP's announcement, because their floors of IT with antique servers were rotted. They wanted to stop at that level of their project, however.
Yes, that Accuweather blurb in the middle of the announcement was very much like a corporate version of the Jedi-hand-wave ("These are not the droids you are looking for.") In HP's case "we have no intentions of spending another cent on OpenVMS, but continue feeling good while running a has-been OS." What I'd be more interested in finding out is how the Singaporean stock exchange --- which only a few months ago moved to VMS --- is feeling right now.
So here at Newswire Headquarters we're weathering the last five percent of a project that will require more than three months of displacement, losses and unexpected expense. Unlike the efforts that migrators are making, all we're doing is working to keep open three businesses' offices -- the Newswire and our Something Elses, Heartfelt Yoga and the Writer's Workshop. The project was something that we asked for, which makes it easier to bear than any migration that might have been needed, but was never desired. As anyone who's done a migration or a Y2K project will concur, that last 5 percent of something large takes four times as much energy. You're worn down after weeks, or months, or even years.
Our salvation -- and perhaps yours and one for the VMS faithful -- is to act in phases.Our project came in at an astonishing initial budget. We could never hope to succeed with that kind of beginning, because anything this big only expands its expenses. There's the unforseen and under-estimated in improving anything beautiful, elegant and old. We loved the praise for our house's pre-transition look. Unique and purposeful and a statement about us as artists and creators: those are the things we rolled around in, when we thought of the value of our house.
So with a colossal budget at our start, we divided up the dreams. Like perhaps a lift-and-shift, and then a revitalization, and finally a re-engineering, we had Phases 1, 2 and 3. Master Suite, Kitchen, and All Else, they became known as, concurrent with a Phase independent of all, pool and landscape. Connecting all the interior parts was the flooring. Yes, the floor that started it all.
The VMS community, like yours that precedes it, will probably look out over the next 7 years and divide the time similarly. The period to stock up with the latest hardware. That's a time to start the analysis and inventory of what's running their companies, toting up every VMS system that's far less obvious than a weather forecasting website. Then the necessary effort of replacement, pushing their Itanium and PA-RISC servers into archival and emulated modes, respectively.
For some in the 3000 community, who are now homesteading as their best business choice, there will come a day when the pail of water hits their wood floors. They have the advantage of expertise from migration services companies, better tools than a decade ago, plus the sympathies and cautions of everyone who's already done it. It won't make the last 5 percent easy to accomplish. But knowing that others have survived a calculated calamity helps you sleep on the nights like this evening, when there are no doors hung on the jambs of our house, and the scent of oil paint is heavy in the air.
June 12, 2013
Newest HP song of server exits same as old
Now that there's another homesteading-migration movement afoot in the HP enterprise community, it's worth studying. What's different about the shutdown of the OpenVMS operations at Hewlett-Packard, versus the tale of the last decade from the 3000? Many moments and passions are similar. Slides not even six months old like the one below foretold of nothing but clear sailing. But with HP's 11 years of extra embrace for VMS, beyond the 3000 sayonara, things may be kinder for the VMS acolytes, those whose faith HP praised in an exit letter.
Within a day of posting the letter, the VMS community was trying to organize an effort to get the operating system source code from HP, re-licensed as open source. Perhaps they didn't take much heed of the 7-year quest by OpenMPE to win the rights to MPE/iX. First there was a set of legal proposals, followed by the logical proposals that the OS couldn't be worth anything to an HP which was casting it aside. I'm talking here about both the 3000 community, as well as those wounded in the world of OpenVMS.
"Is there no one who can free VMS from HP?" asked one member on the comp.os.vms newsgroup. Another member replied with an update from the group devoted to Rdb, the Digital database as vital to VMS as IMAGE is to MPE. He wanted to deal with Digital people in place before a controversial CEO served up the first sale, to Compaq, before HP.
Up on the Rdb list, Keith Parris raised the possibility of HP open-sourcing VMS. While I would prefer VMS to come from DEC before [former CEO Robert] Palmer, that is no longer an option. If done correctly, an open-source VMS might be better than no VMS. Perhaps HP should pay a peanuts-scale salary of, say, $150,000 so that someone can coordinate this full time.
Unless a revolt has pulled down the walls of HP's IP legal group, such license freedom sought by customers won't be forthcoming. HP got badgered into releasing MPE/iX source to a select group of licensees, who cannot improve upon the 7.5 release but use their code to create workarounds and patches. However, the VMS people do have the advantage of a thriving emulator company for any Digital VMS implementations which run on older, non-Itanium servers. The tech issues have been long-solved for Charon for VMS, but there are licensing issues that the Digital user will need to manage for themselves.
Here's where the HP 3000 community is a decade ahead of the drop-kicked Digital group. Stromasys reports that licensing hasn't been an issue in getting Charon HPA/3000 up and running in the early days of sales. HP's provided the MPE/iX license, and that just leaves the third party software.Stromasys reported last month that the license arrangements for the emulator have to be left to each customer who will transition to a virtualized 3000 server. You make your own deals.
But product manager Paul Taffel said that "There have been no problems with vendors. We finally figured out who you have to call in IBM to get the Cognos license, for example." That would be Charlie Maloney, at 978-399-7341.
What the Digital faithful do not see in place yet is a license arrangement from HP for OpenVMS on every platform -- including some that may not yet exist, like an Itanium emulator. In these earliest days, they at least can point to the emulator company that's arranged for such a thing in the past. But there are doubts and uncertainty to go along with fears.
"Are these emulators a serious option?" said one customer on the newsgroup. "The emulators could be a serious option, but what of them, if HP clams up and refuses to license VMS on them?"
The reply from another customer echoed right back to the earliest days of outrage over the 3000 transition. "This is why prying VMS from HP's clammy hands would be the first priority, and nothing else matters if that cannot be done."
Your community marshalled its forces in late 2001 and into 2002 to try to wrest the entire 3000 business from HP, at a price. Hewlett-Packard was not interested, but these are more interesting times. HP just won a lawsuit with Oracle, fighting over the future of Itanium. Oracle didn't want its software to run on Itanium anymore. Neither does HP want OpenVMS to run on Itanium. The wounded customers in the VMS world suggest that Oracle ought to sue to get back its judgement from the prior suit.
To demonstrate there's still value in working with Itanium, HP might be induced or coerced to smooth the OpenVMS path from HP product to community asset. Just like the 3000 odyssey of the previous decade, HP was assuring the VMS user in slide decks dated as recently as December.
Despite Oracle’s announcement to discontinue all software development on the Intel Itanium microprocessor, we remain committed to supporting you and your IT environment. We will continue to support OpenVMS on Tukwila-based and Poulson-based Integrity systems beyond the next decade.
As if that were not enough, another message came down from the man recently promoted to head HP's Labs. Martin Fink was formerly the head of the Business Critical Systems group where OpenVMS remains for sale until the end of 2015. In 2011, while HP battled Oracle in that suit, Fink found the moxie to make a rallying statement that will sound familiar to the 3000 customer. At least any who recall the mid-summer assurances of 2001 that preceded the November shutdown notice.
Fink told OpenVMS customers
Let me reassure you. HP plans to continue the development and innovation of Itanium-based Integrity NonStop and Integrity server platforms with our HP-UX and OpenVMS operating systems for more than 10 years.
At the bottom of each and every slide in these decks is the standard HP disclaimer that anything can change at any time. It's just this: until the song of departure is sung for you, it's hard to believe it HP would sing it to anybody as faithful as you've been.
June 11, 2013
HP tolls bell for penultimate enterprise OS
It took more than 11 extra years, but HP is finally swinging the axe on OpenVMS, the next to last HP-crafted OS for business. Customers in the DEC world got a pass for their OS onto the Itanium architecture in 2001, a route that HP blocked as it started to end its MPE business. But the OpenVMS customer base will die the death at HP from dozens of cuts, beginning with an end of Integrity i2 server sales for the OS at the close of 2015.
Server upgrades for the OS will end one year later, if HP keeps to its plan. The strategy was announced in a letter that began
For over 35 years, the HP OpenVMS operating environment has served as a mission-critical platform upon which you have built your IT infrastructure. We deeply appreciate our long partnership and also the loyalty you have shown HP during this time.
In WW II these were called Dear John letters, received at the front from a back-home sweetheart who was stepping out of a relationship. HP couched its news in the cloak of a "Mission-critical Roadmap Update," (click for detail) and the vendor used phrases like "at least" in front of dates for ending server sales. But an 8.5 version of OpenVMS is not on HP's map, just like an MPE/iX 8.0 evaporated from Hewlett-Packard futures slides in 2001. The equivalent of OpenVMS 8.5 would be needed to support the Poulson class of Itanium chips, processors HP will use in its newest Integrity i4 boxes.
For the most loyal and patient OpenVMS customers -- who view migrating their proprietary systems as kindly as 3000 folk did -- HP will continue supporting Integrity i2 server hardware through the end of 2020. That year aligns with the one picked by HP's expert witness when calculating how long Itanium would be an HP revenue generator. HP learned something from the 3000 market while ending a business line. OpenVMS users will get more than six years of continued HP support -- longer than the five that HP first imagined when it curtailed its MPE business.
The move leaves just one HP-created general purpose OS running on Itanium boxes, HP-UX. (NonStop, from Tandem in an acquisition, is aimed at a much narrower purpose.) Like HP's Unix, OpenVMS servers come from the embattled Business Critical Systems group, where the HP 3000 lived out its remaining HP days. HP promised more remaining sales cycles for VMS servers than the 3000 servers got, but only by a few months. VMS on Integrity will serve out "at least" a remaining 30 months on HP price lists; the 3000 got 24 extra monthly reports.HP's nod to customer loyalty in a letter from Ric Lewis, GM for the Enterprise Servers Business, was an echo of the letter HP sent to its MPE sites in 2001. It's never a good thing when HP starts off by saying your enterprise system "has served," as in the past tense.
The OpenVMS user has an advantage in its homesteading era which starts this week. The OS already has the benefit of an emulator company that's making a nice living in the DEC marketplace. Stromasys has 5,000 installations in 50 countries running its Charon products for VMS.
The road isn't fully cleared for any company to offer an emulator which will put Integrity onto Intel chips like Charon does with PA-RISC. A couple of redoubts remain. HP-UX and NonStop communities haven't gotten their letters yet. The Unix customers might slide sideways into Linux installations, but only with a level of pain and expense lower than the MPE community weathered.
The NonStop customer -- a group small enough to be unable to prop up Itanium design and improvements -- won't escape transition pain. This week would be the time to ask about HP's futures in NonStop, of course. One major difference from the 3000 wind-down: HP broke the news in the middle of its annual HP Discover show. Plenty of damage control would be on hand in Vegas this week.
HP's reaching out to support its less-independent OpenVMS customers with a "high level" of support as long as a company wants to pay Hewlett-Packard. The vendor is also equating transition with homestead support, the same kind of misunderstanding it held for 3000 customers.
We will continue to provide a high level of support to you through the lifetime of your OpenVMS environment. We have a full portfolio of servers, software, and solutions, including support for transitions to NonStop, HP-UX, Linux, and Windows environments.
None of those environments have much to offer that will make OpenVMS transitions less painful, but it's possible that Hewlett-Packard took some HP 3000 lessons to heart. That would mean that vendor-supplied transition tools will lead the way in the marketplace. But who might move from one HP proprietary OS to another? Just ask the companies which found replacement HP-UX apps for their MPE/iX applications over the last 11 years.
Hewlett-Packard doesn't consider these products proprietary, even though customer ecosystems have grown deep software roots into the operating systems' remarkable engineering. There's no significant profit that HP can make in selling servers for OpenVMS, HP-UX and NonStop. There's plenty of support profit, as there was for MPE/iX. So for at least five years after the last Integrity OpenVMS server is sold from HP, the vendor will collect support payments. This would align on the five-year mark HP that settled upon for the 3000. Sales ended in 2003, and HP's formal and full support -- including the MPE/iX lab -- shut down in December of 2008.
So even without letters to its last two Integrity environments, Hewlett-Packard demonstrates that the only future remaining for its most loyal Itanium customers will be 18 months of server sales, and 30 for upgrades.
HP called the update "a rolling (up to three-year) roadmap and is subject to change without notice." HP promised -- in the roadmap's fine print -- that what it's now calling Standard Support for OpenVMS won't be ended unless a customer is given 24 months notice.
A storm of indignity and dismay might arise in the OpenVMS community like the one we all watched here in ours. The revolt and rally might begin with the OpenVMS Boot Camp this fall, but that was never a meeting with vast exhibits and thousands of customers on hand.
VMS customers have already seen HP scuttle a business that was producing profits -- modest ones -- when Hewlett-Packard started its clock on the 3000, however. VMS managers might have made their plans accordingly. There's nothing like tolling the bell on hardware sales to spark migrations, as the 3000 community learned.
HP's making no promises about anything in the roadmap, either. "Nothing herein should be construed as constituting an additional representation or warranty or binding commitment upon HP. This is not a commitment to deliver any material, code or functionality, and should not be relied upon in making purchasing decisions." That's language HP learned to use after its 3000-exit adventures more than a decade ago.
June 10, 2013
How shooting off Moonshot can hit your IT
Some HP 3000 customers are making a migration from a small installation. But for others, their systems are as big as MPE will let them become -- and those sites need even greater computing power. Power is a crucial element in the HP Moonshot server, whose 1500 chassis is a hot topic at this week's HP Discover conference.
HP is making the technology behind smartphones -- one IT manager calls those toys -- shoulder the load of serving up massive websites, or perform financial analysis. Any application with a growing base of users and the need for horsepower that will scale, independent of power needs, might be a good fit for Moonshot. Or as HP calls the product on the server's website, the HP ProLiant Moonshot Server. Look a little harder at this server and you'll see an x86 architecture that's driven by Intel's Atom processors. Atom has a life inside mobile devices like Lenovo and Motorola Android phones.
Not exactly the top tier of phone makers. Apple makes its own A6x. Many other phone makers use ARM chips. In a way, the Atom processor in the Moonshot is a repurposing of the CPU. Atom was built to burn less than 10 watts of power at a peak. HP says the Moonshot chews up 89 percent less energy than the same compute power driven by Intel's Xeon family, or the Intel Itanium. You know, the traditional servers.
HP's not aiming Moonshot at small to medium businesses. When its website says "Shop for Moonshot" you don't go to a Build To Order menu like you can for other ProLiant servers. "Find a reseller," it says underneath.
HP started building the Moonshot line in its labs four years ago. That was an era when R&D got no love at HP, but Moonshot stayed on target anyway. This was HP's entry into ultra-dense computing. For many customers relying on MPE, that's just a buzzword. But ultra-dense computers address a common 3000 need: reduced use of energy, in a small footprint, and cheaper than tradition. You have to go back to the 3000's Mighty Mouse Series 37, or the Series 918 PA-RISC server, to find something comparable in impact.HP is not entering this orbit early. IBM has been in the ultra-dense and software-defined server market far longer. But there's a need for something to drive big sites like web hosting companies, social networking providers and video warehouses without burning countless kilowatt hours. HP CEO Meg Whitman said that the equivalent in traditional servers could use as much power as 20,000 homes to keep up when applications scale up.
But that's at the high end of concerns. For someone who manages 3000s and thinks, as one IT manager wrote us,
If I could put a Moonshot in a workgroup of 8 or 10 people, and connect them to it with thin clients or Windows terminals, and call it their workgroup server, then I could significantly shrink the datacenter.
Not really what the server is built for, sadly. We're talking thousands of people in a workgroup. Running something that wants to be Facebook or Twitter, but is not yet. Or serving video like YouTube. But the sweet promise of Moonshot is that being software-defined, its something that enterprises can optimize based on specific workload needs.
As it has for a long time, HP is showing off its large-scale system offerings to medium-scale customers. Because everybody wants to be bigger. One way to do that is to use something smaller, like an ARM processor or even Intel's Atom, the latter being IA-32 and x86-64 compatible. The processors from those toys are gathering in racks of 1,000 to make high horsepower computing -- perhaps on the horizon for a migrator -- cost less and draw less power and footprint.
June 07, 2013
Find invention at Discover's web, or garages
The 3000 community recently took note of the MPE departures from HP's ranks. The revelation about what else everybody's doing elsewhere has triggered chatter about HP's ability to invent. This subject can be important to the migrating sites that are sticking with HP's platforms, whether those systems run HP's Unix, Linux, or Windows. Customer Delight can result from inventions, so the matter of whether HP invents anymore can be a part of an evaluation.
To nobody's surprise, the chatter judged HP harshly on its invention score since dropping the 3000 as well as those MPE experts. But HP's Jim Hawkins, one of the select employees still holding MPE skills in his toolbelt, came to his employer's defense. If you'd like to see HP's most recent inventions, he said, going to next week's HP Discover conference would be a good start.
Few 3000 migrators are going to Discover, however. HP recognizes that the majority of its customers can't swing a trip to Vegas and an $1,800 entry fee. So Hewlett-Packard, a company which for a time had the word "invent" bolted onto its logo, will put selected Discover talks on the Web starting June 11. Whether you'll consider those presentations inventive depends on a customer's definition.
"My understanding is that CNBC will be broadcasting live from the [Discover] floor, Hawkins offered up on 3000-L. "If you want to learn what HP is really up to, there's a good place." The engineer whose expertise lies in MPE's IO added that he'd been to a "poster fair" at Hewlett-Packard recently, a gathering of IT inventors who presented their HP Labs concepts for things like network innovation.
If only the customers who've been in IT for decades were an easier crowd to convince. They remember an HP that prized inventing which flowed from the Spirit of the Garage. Plenty of things look like iterations of concepts, rather than inventions. Even Apple and Android-ians are facing that comparison next week. To seek out garage inventing you need to go farther online than vendor websites and conferences, whether Apple's or HP's.While HP's hawking its advances at Discover, Apple will be presenting the best of its newest mobile software at the World Wide Developer Conference. WWDC sold out in under two minutes this year, at a cost of $1,700 per seat. Last year it took two hours. Even as Apple earned $9 million in 90 seconds or so, it still had to extend a promise of inventions applied to a white-hot mobile market. Registered Apple Developers can watch WWDC presentaations online, too.
"Apple is 'winning' because they're getting what they want in the (former) HP Way," said 3000 consultant Glenn Cole in the discussion. "They make a contribution, with millions of people able to take advantage of the attention to detail that Apple puts into their (still imperfect) products, and they get the cash to be able to invest in pushing the industry forward -- and also be everyone's personal computing design lab."
Computing design used to be paramount at that Hewlett-Packard of the HP Way. Everyone acknowledges that generation of HP is gone. But Hawkins mentioned inventions that won't be a part of Discover. These came from the HP Labs' poster fair. Whether HP's got the spark to market what follows is an answer to discover.
There were some really cool things like a no-glasses 3D display, and a method to determine co-location of devices without GPS or WiFi location information or any communication between the two devices. They use audio from device microphones to establish a pattern of -- background silence -- which is apparently better than the background noise [think two people listening to the same radio station in different locations]. That pattern information would be shared on a server and help you meet-up with people close by.
At the same time there were lots of other things which looked vaguely familiar to someone who's been in commercial computing for a while. There were ways to handle the change in the amount of data (tera-, exa-, pentabyte data sets), to structure unstructured data, to reduce the impact of the mismatch between CPU-to-memory-to-secondary storage.
As this HP-MPE engineer, still working at Hewlett-Packard, went on to examine the prospects of things like a system with memristor memory which could be permanent and as fast as your CPU, 3000 veterans grew quieter. One former reseller even expressed hope HP could return to its invention roots.
"I hope that HP can get back to blazing new trails to the future," said John Lee. The moving-off-MPE customer will likely hear a lot about HP Moonshot servers at Discover, as well as the newest, award-wining Software Defined Network products. "A conversation with HP chief technology officers." But Moonshot looks like a slimmed-down blade server to discerning eyes in the 3000 community
"It was difficult to tell what was innovative, unique or patentable about Moonshot," Lee said in reply to Hawkins' offering of Moonshot and SDN. "I'm not being critical, I'm giving a layman's opinion. HP looks the same as everyone else. Same with the SDN solution. I get product literature from IBM and they make the same claims."
Apple's invention looks just iterative to those who have no more affection for that vendor than the 3000 base does for HP. The Android-driven Samsung has the same problem with its clever Galaxy 4 phones. Waving at a screen to move items looks sexy, but might have its greatest value in a commercial. Skeptics don't like identifying things as inventive anymore.
If you're wondering what invention looks like in 2013, you'd be well served to go elsewhere online. You can find it at sites that serve up news coming from garages. Sites like hackaday.com or armageddon.com show off more than poster presentations. Inventions like an atmospheric water generator that extracts water from humid ambient air, or the Xprotolab portable oscilloscope, logic analyzer, and function generator. Or the biohacking work of Steve Mann, "a professor of electrical and computer engineering who has dedicated his career to inventing, implementing, and researching cyborg technologies, in particular, wearable computing technologies. Mann has been critical to advancing biohacking through his self-implementation of his inventions."
Much this technology got built in a garage, even the likes of the Rostock delta 3D robot printer. The passionate, rogue inventor might feel little comfort working in a corporation. But HP rose up out of a garage. Its first product was an oscilloscope innovation. There is hope for invention from a company so sophisticated that it hosts its own IT conference every June. Tim O'Neil, another 3000 manager, considers Discover the rebirth of the Interex conference. He sees hope in HP's pursuit of eyeballs for its latest products, though.
"After selling the HP chip design team to Intel, they are now going to lead the way towards flash storage and universal memory?" O'Neill said. "Nevertheless, it is encouraging that they actually want customers to attend."
June 06, 2013
Today's Limits on Emulation Speed
The long-term future of 3000 virtualization looks sunny, in part because there's the remainder of our lifetimes for the Stromasys engine to get faster. Using this month's offerings from the company shows there's plenty of performance to make up, simply to get to HP's 3000 benchmarks from 2003. Migrators won't care like homesteaders will.
When you look at the now-shipping Stromasys Charon product line, it's easy to see the product will run as fast as nearly all of the A-Class and N-Class servers. But HP sold three models of N-Class that are still out of the reach of today's virtualization engine speed. Those models represent the threshold Charon must still break to operate as fast as the fastest, hamstrung HP iron.
Hamstrung is a word from horses and people meaning to cripple. It's an acknowledged practice that the 3000's CPUs were down-clocked by MPE/iX. In some cases, like the lowest-end A-Class, the operating system dialed back the processor by 75 percent. A 440MHz CPU was forced to run at a genuine 110 in the entry-level A-Class. This is one reason why a fully-revamped, 1-processor A-Class HP 3000, with a faster bus, still only ran 70 percent faster than a Series 918. Even at the dawn of a new generation of 3000s, HP was keeping the servers in check.
If a company is considering an emulation scenario on the way to a migration, these limits might not matter. At one Dallas-area e-commerce company, consultant Doug Smith reports the 3000 was moving to archive-system status. A migration was in the wings. But for other companies, hoping to match those three biggest-sized 3000s, June's Charon product line will leave them short of a match.
Migrators might not make up much of the Charon customer base. If they've concluded that a midrange 3000 will do the interim job, however, even companies leaving the platform will have enough horsepower. One of the reasons for this involves adding extra CPUs for the unreleased Charon versions. Unlike HP, Stromasys will support a 6-way or 8-way N-Class.The virtualization lineup won't support these today, unless you count the ability to provide model numbers in the lineup Stromasys that showed at its recent Training Day. The N4060 and N4080 offer 50 percent and 100 percent more CPUs than HP ever would for a 3000. At the moment, Charon will need those extra CPUs to run as fast as HP's fastest N-Class configurations.
When it was still a product for sale from the vendor, this kind of top-end N-Class would cost more than $100,000. At that, it was a complex hardware-based server subject to the indignities of iron: CPU boards, specialized disks, proprietary memory -- all might fail in the way physical components do. But you could get very fast, with enough memory in an N-Class 500MHz or 750MHz server.
AICS Research, which sold QueryCalc to hundreds of 3000 sites for more than 25 years, still hosts the comprehensive 3000 Relative Performance Matrix on all MPE servers. The listing of benchmarked systems goes all the way back to the Series 30, as underpowered at the time as the early Series 930 was at the start of the First Generation of PA-RISC CPUs. HP wrapped up its benchmarking with a muddled picture. In 1998 it changed its processor comparison rating to the "HP Performance Unit." Since it was near the time of the e3000 renaming, this became known as the HP EPU.
"In effect," wrote the late, great Wirt Atmar in his notes to his matrix at AICS, "what they did was to begin to use a new set of test suites which they felt were more appropriate to the way that HP 3000s were being used."
The numbers aren't absolute measurements of anything other than the time the various systems take to process one of the several HP test suites, but they do allow you to compare with some accuracy the relative performances that you should expect when upgrading to higher performing system.
You must divide by 10 to get a set of EPUs that make sense in HP's final 3000 product lineup. The Series 918 became a 1.0 when HP recalculated its benchmarks. That means the 4-way 500 and both 750 MHz models at right are coming in at 49.9, 60.6 and 76.8 EPUs, respectively. The Stromasys lineup doesn't go that high today. As Atmar said, those are just numbers from a test suite. Your mileage will certainly vary.
Migration prospects these days are most likely to be smaller customers who didn't see a good business case for leaving the 3000 from 2003 to 2013. They need help from third parties to make their move, and useful advice might be to start the migration while using more reliable iron than HP's 3000 "kit," as the British would say. These people have got what they need from Stromasys today.
However, as the economy rises, some of the larger 3000 homesteaders may be replacing HP hardware that makes their boardroom directors nervous. Emulation can happen even at the same time the big homesteaders make a move -- like Amisys healthcare hosted sites are doing -- to something non-MPE.
The biggest homesteaders don't have a place to go yet to match their top-end performance on today's virtualization engines. However, using extra CPU cores will let Stromasys keep adding CPUs.
Charon-HPA/3000 systems use two cores to emulate each PA-RISC CPU – one core fetches instructions and interprets them on-the-fly into equivalent Intel instruction sequences, while the other core works in parallel, pre- fetching code pages and applying various analysis and optimization algorithms to identify higher-level optimizations.
This More Cores strategy puts some of emulation's limits in the hands of Intel. That's a company that has not been shy about adding cores to its Xeon-x86 line of CPUs, the line that drives Charon. The level of innovation on that line is moving at many times the evolution speed of Intel's other enterprise chip, Itanium.
It's a gamble to guess at future performance needs, so selecting a hardware-based 3000 always involved predicting the lifespan of use -- the headroom for a site's application growth. Virtualization isn't much different in the forecasting, but moving up is a new experience. What makes virtualization attractive is the ease of upgrading to a faster model. HP might have called this a board swap back in the days that it sold hardware. This kind of swap was still on the order of $20,000 for the leap from HP. And that was after you turned back your old processor to HP, so it could be resold. Nobody's going to have to turn back a model of Charon software in order to turn up their performance.
June 05, 2013
Legacy hardware evolution looks limitless
At the recent Stromasys Training Day and HP 3000 Social, the company's GM Bill Driest asked a question about the future of the HP 3000. But he may as well have been asking the same thing about HP's Integrity servers, too. What's to become of these vendor-specific systems, once the vendor leaves the system behind?
GM Bill Driest suggests the sky's the limit for futures in hardware that's been curtailed by the vendor. At right, Stromasys CEO Ling Chang talks over the possibilities at the recent HP 3000 Social with Eric Sand of Sandsoft.
“People like Gartner are talking to us, and there’s been a fundamental sea change,” Driest said. “They’re saying this: isn’t it conceivable that the end state of all legacy hardware is some kind of emulation or virtualization?”
Driest admitted that five years that belief was “so much of an early adopter message. There’s a fundamental pause as we ask, ‘On what platform do you believe we’ll run the last MPE production environment?’ Do you really think that it’s going to be on some refurb HP hardware?”
The company was introducing a strategy of “Rebuild, or Revitalize?” as the driver towards virtualization of the MPE-ready hardware. It exhorted the customers and resellers, along with support providers and consultants in the Computer History Museum's meeting room, to “Join the Revitalization Movement.”Asking about the future of legacy hardware was once a moot question. Of course it would be decommissioned and drop-kicked to the curb. A few gun enthusiasts even executed a 3000 and captured the gunplay on video. But systems continue to serve despite such schoolyard jeering, and in spite of the age of the hardware. It's the age of the software that matters -- something that can be updated more easily once the box is virtualized.
This might seem to disrupt HP's plans to step away from platforms like the MPE engine of the 3000, or the VMS hosting on Alpha or VAX -- or perhaps the PA-RISC HP-UX servers, and dare we say it, the Itanium-based boxes. HP's own expert has said he figures Itanium production to be good only for another seven years. After that, the Integrity box might become legacy itself. Why, we wonder, has HP added the Charon products to its Worldwide Reseller Agreement?
Stromasys has never claimed to be creating a virtualization engine for Itanium processors. But given the size of the 3000 market versus the efforts to create Charon HPA/3000, I'd speculate that some of that HPA engineering could be re-used for another HP project.
According to Driest, analyst Andrew Butler of Gartner published a report this year that identifies Stromays as a disruptive technology in the server market specifically. “Butler said that ‘Functions commonly thought to be part of the underlying part of the OS are being distributed to the server, to the hypervisor, to storage.’ People are virtualizing everything. We’re almost re-inventing what it means to deploy an application,” Driest said.
The meeting presented a new and more extensive group of Stromasys executives to the 3000 community. CEO Ling Chang said the company “is proud to become a part of the 3000 ecosystem.” More than a decade after HP announced the 3000 would drop off Hewlett-Packard’s price list because of a declining ecosystem, that group has gained a member worthy of citation from Gartner.
When Gartner first started tracking the Stromasys offerings three years ago, they called it processor emulation, Driest said. “We fit in on this little bubble called Processor Emulation. We are fitting in at the peak of their curve called Inflated Expectations. It says there’s a promise for this technology.”
Hard questions came from one reseller who reported that he serves 135 companies using 3000s and had clients in the room from several California school districts. “Think about how you can help us help them,” he said. The reseller has been in the reseller business to service customers who manage K-12 schools using MPE/iX solutions.
The product lineup from Stromasys (click for detail) showed the top-end emulator being an N4040 virtualized system, running four processors at 250MHz — not the 750MHz rated for the 4-CPU HP hardware. The N4060 and N4080 models are forthcoming. The latter runs at an estimated 61 HP EPUs — four times faster than the A520 Charon model.
June 04, 2013
Stromasys opens HP's way to Charon gates
The maker of the emulator solution for the HP 3000 community demonstrated the natural resting state for MPE applications during its recent training and brewhouse social. Dedicated community veterans, as well as some customers looking for a way to extend those applications, took note of a new alliance. HP's got the 3000 version of Charon on board.
Stromasys also announced it’s just been named a Gartner Group Cool Server Vendor for 2013, the freshest part of the news, plans and futures the company unspooled in its first North American Training and Social event for 3000 customers and allies. The room of the Computer History Museum on May 10 was full for the day-long briefing on company strategy, as well as Paul Taffel's extensive demonstration of the HPA/3000 model of Charon in action.
Stromasys is one of only seven vendors who’ve made the server technology cool list, just published by Gartner. The company showed off a product lineup that includes a pair of implementations that are designed to out-perform some N-Class HP 3000 hardware. General Manager Bill Driest said he’s seen his company's software run on a cutting edge HP DL380 server with a 4.4Ghz processor installed, a pre-release from Intel.
But the power promises may extend beyond hopes of matching high-end N-Class performance. HP's taken on the software as a potential solution for its customers. Stromasys hopes the 3000's will share the view that hardware is only a waystation to a virtualized platform.
Work is underway in the Stromasys labs to utilize extra cores on the DL380's processor for such servers. With each 4-core set available in Intel chips, HPA/3000 could emulate another HP 3000 processor. The 32-core limits of today could yield an 8-CPU MPE/iX machine. This is a configuration HP could never ship or officially support while it built and sold its 3000 servers. The HP top-end was 4 processors for its 750-Mhz.
Driest made his debut in front of a HP 3000 crowd during a morning session that outlined where Stromays is heading from its current position as the only virtualization solution in the PA-RISC space. One new wrinkle was the announcement that Charon HPA/3000 has made the cut onto HP’s Worldwide Reseller Agreement. Stromasys already has its product that emulate VAX and Alpha systems on that list.
A Worldwide Reseller Agreement gives HP the right to resell a product from a software supplier. Companies as large as security supplier McAfee have entered into such a deal. HP now has the mechanism to sell Charon HPA to customers who might want to remain as MPE/iX application users.
The Gartner announcement was a sneak peek at what Driest was describing as a way to earn its solution onto the Hype Cycle of Virtualization. Processor emulation is in the Expectations part of the curve, but Stromasys hopes to be securing a spot in the Trigger, a rising wave of the lifecycle.
“People like Gartner are talking to us, and there’s been a fundamental sea change,” Driest said. “They’re saying this: isn’t it conceivable that the end state of all legacy hardware is some kind of emulation or virtualization?”
Driest admitted that five years that was “so much of an early adopter message. There’s a fundamental pause as we ask, ‘On what platform do you believe we’ll run the last MPE production environment?’ Do you really think that it’s going to be on some refurb HP hardware?”
June 03, 2013
Touting Meg, HP says Discover's not too late
An email in today's in-box reminds me that "It's not too late to attend HP Discover." The 2013 edition of the HP conference, wrapped around all things Hewlett-Packard for enterprise, cloud and mobile computing, plays out next week, June 11-13 in Las Vegas.
HP's email promises a chance for direct interaction with its CEO Meg Whitman:
Engage directly with Meg Whitman and top HP executives to see how HP innovations shape the IT industry and help IT leaders like you succeed. Talk one-to-one with HP industry leaders via the HP Meeting Center and Partner Meeting Center. Book a tailored discussion of your choice.
Converged Cloud, Information Optimization, Mobility Security and Risk Management, plus Business and Technical sessions are HP's main tent poles at the show. Signing an NDA gets you "the chance for a sneak-peek at what's next from HP during our extremely informative confidential disclosure sessions."
In one public session, HP will tell us how a full one-eighth of its website operation is being run from its latest servers.HP Moonshot drives 12 percent of the HP.com website today, according to one session description from the conference catalog. HP's "eating its own dogfood" with this massive cluster of CPUs to make up one computer, built out of Atom processors and an HP original computer system. An original enterprise system -- that's news.
"In this in-depth technical session you will gain a understanding of how to deploy HP Moonshot System in a production environment," the catalog reports.
Savvis, "a leader in managed computing and network infrastructure for IT applications, will share the latest results of their testing on the Moonshot System."
The meeting at the Venetian on the Vegas Strip will also include two hours of talks on the OpenVMS environment, the closest thing HP's got to a 3000-loyal customer base. Plus, there's a Special Interest Group meeting. Another 692 talks and demonstrations are on hand. HP's keen enough on OpenVMS to schedule one demonstration for three separate 90-minute demos.
Typing the word "Unix" into the session catalog's search engine yields one result; there are dozens more for HP-UX. But HP is launching into HP-UX saying that you can "Pave the path for your mission-critical future with HP-UX and Integrity server innovation."
It's an ongoing battle trying to get ahead -and stay ahead- in business. Continuous innovation to your mission-critical IT infrastructure is vital to beat your competition. In this session you will learn about the newest ways to make your UNIX environment excel. We will cover the latest developments in Integrity systems, including increased flexibility and virtualization, along with new Serviceguard high availability features. Accelerate your critical IT to achieve better business results when you need continuous business.
As if that were not enough, there's HP's "Tectonic Shifts, where the future of convergence is taking us,” and where Martin Fink — the former head of the embattled Business Critical Systems group, and now HP CTO and Director of HP Labs, “will outline HP Labs' research in flash storage and universal memory, with an eye on where future technology might take us.”
$1,795 will take you to Discover this year.