June 30, 2010
HP's direct attention can evade Unix buyers
Buying a Unix server from HP is a direct transaction that's gone indirect. We recently got a message from HP 3000 software vendor STR Software, who also sells solutions for HP's Unix line. One of Ben Bruno's customers is struggling to get to the correct part of HP and upgrade an HP 9000 system.
"We are looking to replace our HP 9000 with a new HP system," the customer told Bruno. "Unfortunately, we can’t seem to get anyone at HP to talk with us on this upgrade. I’m looking for a sales team to evaluate our current solution and make a recommendation for the replacement/upgrade. I’m really hoping to stay with HP, but I just can’t seem to get to the 'right' person."
From the looks of this request, Bruno's customer isn't a large enough enterprise to merit HP's direct attention. This is one area where moving away from the HP 3000, and into HP's Unix systems, lands a customer in the realm of resellers. The resellers were moving most of the 3000 systems by the time HP stopped selling them in 2003. Direct vendor sales teams stick to large accounts.
The evaluation and recommendation is the dicey part of a Unix customer's needs, if they want the analysis from HP. The trick is for a customer to be of a size that HP wants direct contact with today. We've heard the current figure is a $4 million purchase to get direct HP attention -- otherwise, you're purchasing through the reseller channel reps. There are several HP hardware resellers, dealing in both old and new systems, 3000 and HP Unix, who can evaluate and quote. They provide the direct attention you would expect from a sales team of the old HP.
As to sources, lots of our sponsors sell HP's new-generation Unix servers. Genisys and Bay Pointe Technology come to mind right away. The Support Group is another hardware resource that also offers Unix sales, plus software and application support. These kinds of suppliers are eager for the direct contact with customers, regardless of the size of the deal.
HP is doing more direct sales of its ProLiant servers (Windows/Linux) these days, but the eval and recommendations are still not easy to get out of the HP reps -- unless you're a customer of a certain size. There's HP's Factory Express to get a custom build of an HP server, whether it's a ProLiant or the Integrity line. (800.282-6672, www.hp.com/go/factory-express.) But you won't mistake Factory Express for the classic HP 3000 experience of an in-person visit to evaluate, followed by a recommendation from a familiar sales rep.
One of the reasons that the ProLiant line has more HP reps is because of a lower price range. HP has to hold down its direct sales entry bar on systems like the new G7 servers announced last week at the HP Technology Forum. Bladed systems -- which is where the migrating midsize customer should be headed, using the C3000 or C7000 chassis -- start at $3,079 for the BL465c G7, and a 4-processor BL685 G7 blade is $10,268. You make the initial investment in the chassis, but you'd be buying a rack for the ProLiant DL servers, anyway. (A fully-stocked blade chassis is shown above.)
Of course, the above prices are for Windows/Linux servers, plus something HP has started to call UNIX x86. That last item is not the HP Unix which Bruno's customer is using, or something HP will sell. Sun sells UNIX x86 (or Oracle does, by now) and it's been called Solaris.
But there are also bladed HP-UX solutions in the Integrity line. You know that it's been awhile since a customer has bought a Unix server from HP if they're asking for an upgraded HP 9000. That's a product number associated with HP's PA-RISC chips, a design HP stopped selling for both MPE/iX and HP-UX environments. ProLiants run on Opteron, and the Integrity units use Itanium2.
HP is protective of the HP 9000 business which remains, however -- the companies who use HP for hardware and software support, or purchase storage and networking to integrate with their HP 9000s. A hardware virtualization solution is on the way for the Sun Unix and HP 3000 MPE/iX environments. But none at all for the HP 9000s, or IBM's Series i/AS400. In HP's case, the vendor doesn't sell the hardware anymore, but doesn't want to see a competitor rise up. In IBM's case, it's still building Series i systems.
Bruno's customer said they want to stick with HP while buying a Unix upgrade, a choice that will lead them to stick to the reseller market. IBM might be more directly responsive. But then there would be a migration from one Unix to another to manage, too.
June 29, 2010
Archived 3000 site donates $18K printer
Mike Cheng wants to give away a printer that was once as busy as General Chemical's HP 3000. Both components are in semi-retirement today, but the Printronix L5520 has been waiting for a new home for four years. At its height of duty, this $25,000 device output reports for a range of in-house applications that managed the manufacturing at General Chemical. This can be a tough item to find; one online resource listed it available at $18,000; it's no longer made by Printronix.
Cheng, who can be contacted at 973-599-5536, will donate the printer for the cost of shipping a unit that weighs 300 pounds by his estimate. "Or if they're in New Jersey, they could just come by and pick it up" at his East Hanover IT shop, he added.
General Chemical's unit, where Cheng manages IT, manufactures soda ash and nothing else, a substance that's key to creating glass and detergents, among many other things. The HP 3000 at his shop is in archival mode now, generating reports for ERP operations up through 2005. The system will be in service for several more years at the instruction of the General Chemical legal department. This is another HP 3000 that will continue to run longer than HP's support for the 3000. It just doesn't need a printer, or the boxes of 12 x 8.5-inch, left-right feed 3-hole paper needed for the Printronix unit. The paper's part of the donation.
Cheng called us at the behest of VEsoft's Vladimir Volokh, who visited the General Chemical site this week. The indefatigable founder of the software company that sells MPEX continues to travel the country to visit customers, offering a day's support consulting -- and staying in touch with the community that includes such archive outposts.
3000 expertise isn't needed any longer at General Chemical's New Jersey operations. The only 3000-savvy IT manager works at the company's Wyoming plant, where Phillip Kinkead calls on 30 years of 3000 experience, according to Cheng. The New Jersey operations split 11 years ago into two units, and Chang's group purchased a 3000 to stay in step with the ERP software already in use.
Now SAP runs the original 3000 site's day-to-day operations, and Oracle's Enterprise One ERP manages the ops at Chang's shop. The cutover for production computing took place in 2006, he said. But Beechglen is still supporting the 3000 at his shop, which remains "a really solid machine."
This is an HP 3000 that won't show up easily on searches for customers using the system. It's operating in a shop without a 3000 manager, implemented as an archival system, and doesn't have much connection to HP 3000 vendor's shops. Except for Vladimir's, of course, a company that knows about many HP 3000s that could be overlooked in a 3000 customer census.
June 28, 2010
HP keynote splash goes on Web from Forum
Complete with comedian Jake Johannsen as host, HP's splashy keynote show from last week's Technology Forum is online for lunchtime viewing at its website. The show's comedy starts at the 10:00 mark, while HP's czar of servers, storage and networks, Dave Donatelli, outlines the Converged Infrastructure product lineup starting at 20:00.
For a taste of the technical drill-down into the details behind the BladeSystem Matrix technology included in HP's new server products, HP engineer Nigel Cook -- inventor of the magic that configures extra processing power in a flash -- demonstrated the interface staring around the 1 hour, 3-minute mark. Cook is among a new cast of HP technical experts that Hewlett-Packard is pressing into customer videos and presentations. He shows how ProLiant and Integrity blades can be assembled through virtual machines to meet unexpected needs.
"It looks very simple here," Cook says, "but if you're doing this without Matrix, it can take months."
Converged Infrastructure sounds, from Donatelli's overview, like more than a way to "reduce IT sprawl." Reconfiguring blade servers on the fly is a way to cover your needs "in SuperBowl time, when a marketing department ran an ad." Given that such an ad runs more than a half-million dollars, you can get an idea at where HP wants to sell the Infrastructure with its own salespeople.But HP's also got a reseller channel to push these products out to the smaller customers who have made migration their new path into the IT future. If your marketing department doesn't have half-million ad buys on its radar, you still might need to accommodate sudden surges in sales. Donatelli described the ideal (from HP's perspective) moment of need, expressed by business units to IT managers.
"We didn't know it would be so good," he imagines the conversation. "I need a lot of Infrastructure, and I need it quickly." In plainer terms, IT needs might surface overnight, or inside a week, or after a budget has been set for operations. You're not supposed to need to scramble to respond to this, if you carry that Convergent Infrastructure at the ready.
HP had capacity on demand offerings more than seven years ago, just like IBM. The concept has evolved to include storage and leans on virtualization, technology that has improved (faster, more granular) as well as dropped in price. So it's probably within the reach of the midsize HP 3000 customer, but in the sub-category of "build-out" IT. That's the kind that starts in a small configuration you might increase, with supplemental purchases.
Johannsen might have put the audience at ease by pricking the HP balloon of rocket-launch adjectives. "It's good to know that we're the shape-shifters," he cracks after he strides through the dry-ice fog of the musical intro. He thanks the band and then goes into the comedy that pokes at HP's hyperdrive summary of Converged Infrastructure.
Jake: The beauty of the Superdome is...
HP booth guy: That it is a scalable, grow-as-you-go architecture that allows customers to grow applications as they need, right? That's it in one sentence.
Jake: Awesome. All I can say is, "Who dat?"
HP is selling this message to the sales force and pre-sales tech staff as much as to the customers at the Tech Forum. Some of the HP folks might hear the equivalent of "Who dat?" at the smaller sites. Spending some time with the HP smoke and joke show online, along with tech specifics from experts like Bruce, could at least carry you beyond Who Dat -- which is not the latest backup solution, by the way.
June 25, 2010
Connect confirms IT recession is receding
Connect, the HP enterprise computer user group, released results of a survey at the HP Technology Forum that show a majority of members and HP users believe the IT economy is already rebounding.
The survey of 103 community members, conducted through the Connect myCommunity website, reports that 60 percent of the respondents said they are "moving from recession to recovery, increasing IT spending." Slightly fewer members surveyed, 56 percent, reported they have plans to invest in new servers or have already begun to upgrade server farms.
The Connect membership is centered in the small- to midsize company, those with annual revenues of under $500 million, according to immediate past president Nina Buik. She said the survey was completed on June 15, in preparation for this week's Technology Forum. But another set of questions on the survey were designed to test the waters for HP's Converged Infrastructure offers of this week. Product releases are built around this concept of reducing IT sprawl through creating efficiencies.
The user group prides itself as being the voice of an important part of the enterprise community. Buik said this focused survey provided HP the input on how accepted the Converged Infrastructure might be.
"It was really important for us to understand where are members are," she said. "We asked them a few questions in terms of Converged Infrastructure," including how many are embracing the architecture or want to know more about it.
The survey reports that 67 percent of respondents have either moved to such an architecture or are just beginning a move. And 56 percent are "either totally on board with HP's Converged Infrastructure or would like to learn more about it," according to a Connect backgrounder on the results.
Nearly all respondents to the survey, 91 percent, think that the industry is moving toward such an architecture, or it is likely to do so. Benefits -- which Buik said the user group translates "from marketing adjectives to member verbs -- were listed as flexibility, followed by lower costs of acquisition and implementation.
The survey even asked about barriers to adopting Converged Infrastructure, in part to give HP insight on how to push the solution across the threshold of customers' analysis. Barriers they listed that would prevent them from implementing:
• Fear of high costs (37 percent)
• Worries regarding risks to current applications/operations (35 percent)
• The belief that current processes would be too difficult to change (32 percent)
Cost issues, as well as the difficulties in change of business processes, are among the chief reasons the HP 3000 community is adopting homesteading as a longer term plan. But even the storage aspects of Converged Infrastructure could benefit a 3000 shop with a mixed environment. HP was gathering its own responses at this week's conference, which Buik was estimated would attract 5,000 to 5,500 total attendees -- including all HP employees, conference exhibitor staff, presenters and customers. That's an increase over the 3,500 that attended the 2009 show, she added.
Even community members who are not attending will be engaging with professional help to maximize IT potential, some sparked by the HP messages of this week. According to the survey, 82 percent of the respondents "felt that it is really important to have trained professionals help with installation, configuration, integration and testing."
IT budgets are rebounding, recovering from two years of tight spending, Buik added. "“It's encouraging to see organizations increasing their IT spending after what has been two tumultuous years. immediate past president, Connect. “One of the best things about Connect is the unique opportunity for our members to share best practices and participate in on-going trainings so they can make the most informed decisions when investing or implementing new technologies.”
Connect's leaders also want to point to the group's website as a key in staying -- well, connected with user needs. 80 webinars are scheduled for this year. The heartland of its membership uses the OpenVMS and NonStop environments, staying involved with Connect because there's no other resource that fits as well. There are plans for Connect events to serve these two markets in 2010, including the OpenVMS Boot Camp that's been held for more than a decade. HP-UX users have a Special Interest Group. But storage, a large part of the Converged Infrastructure message, is also key to the group's relationship with the market.
"Connect has a strong focus on the OpenVMS, Non-Stop and storage communities," said president Chris Koppe. "We are proud to bring our members a place where they can come together to discuss specific technologies as well as network with peers who have similar interests. As the community continues to grow, we hope to strengthen these conferences as well as introduce new events around various technologies of interest to our community."
June 24, 2010
New ProLiant blade servers get IO flexibility
HP announced the new G7 generation of ProLiant rack servers and BladeServers at the HP Technology Forum, Linux/Windows systems that give IT managers immediate configuration flexibility.
The hardware, expected to ship early in the third quarter, adds new Virtual Connect FlexFabric IO adapters. The feature lets a manager change the type of host bus adapter on the fly, from Ethernet to iSCSI to Fibre Channel.
HP Blade Architect Gary Thome said the BladeSystem advancements were the most significant in the four years since the blades' C3000 chassis was introducted. HP recently sold its 2 millionth blade server, he added. The uplift on this option for 3000 sites migrating to Windows comes from features to reduce IT sprawl with last-minute configuration.
"It allows customers to decide at the last minute what the server configuration will be," Thome said. HP called the rack-mounted systems "scale-up" servers, nouveau terminology that might be better known as "distributed" in the language of the HP 3000 system manager.
The DL980 is an 8-scoket server that supports up to 2TB of RAM, while the BL680C is a BladeSystem supporting up to 1TB. The new systems have 20-80 Gbits of FlexFabric IO on the motherboard. They also use a new architecture that lets a system administrator keep virtual machines online even while one goes down with a non-correctible error.
Virtual Connect FlexFabric is HP's module to connect servers to any Fibre Channel, Ethernet or iSCSI network. The last-second capability requires a pair of Virtual Connect modules. Administrators can modify the number of adapters from between 2 to 8, plus allocate different, unequal bandwidths for each adapter. Thome called Virtual Connect "wire-once" technology, where the blade systems arrive physically wired and can be reconfigured "in minutes instead of month." The modules will start at $18,500 and ship in the third quarter.
A new Prima Architecture isolates virtual machines that have developed the non-correctable faults, lets admins isolate and take them down and then restart them while the remainder of the virtual machines continue to operate. Prior models of the ProLiant line, including last year's G6, required a restart of all VMs if one developed such an error.
The three new rack-mount ProLiant scale-up servers deliver “self-healing” memory capabilities to maximize application uptime. HP said the DL980 is 200 percent more available than the G5 generation DL785.
Seven new ProLiant G7 server blades offer the industry's first blade with both 1 TB of memory and integrated 10Gb Virtual Connect FlexFabric technology for IO scalability. Thome said the architecture must do more than deliver CPU speed for balanced performance, extending beyond processor advances to expanded memory and IO capacity. HP says these systems can support up to four times more virtual machines than competitive blades, while requiring 66 percent less hardware.
A new Intelligent Power Discovery feature is part of the G7 line, creating an "energy-aware network" between the ProLiant rack-mount units, third-party tools and datacenter power grids. HP said that a company can save up to $5,000 per server over a year's time in energy costs through the network's provisioning. HP's calculations are based on a 1,000-server "midsize" datacenter, capping and recovering 200 watts per server on average.
HP also introduced a solution that lets datacenter managers provision resources on an application level. BladeSystem Matrix, working in conjunction with the HP Server Automation software, lets customers "snap their fingers and it happens" while provisioning, based on a server template an administrator draws. HP calls the technology an essential part of "private clouds" comprised of servers and storage. BladeSystem Matrix's latest version starts at $150,000 and ships in mid-July.
June 23, 2010
HP works its storage options wider
Riding the thoroughbred of its own innovation, HP has announced new enterprise-grade products from its StorageWorks unit, unveiling the news at the HP Technology Forum in Las Vegas. New software, StoreOnce, is being rolled out to roll back the need for more petabytes of storage capacity. VP of Marketing, Strategy and Operations Tom Joyce said HP's "engine of innovation" with StoreOnce is HP Labs, where the software was developed.
Joyce, who joined HP from EMC over the past year, invoked the name of HP's stoutest storage rival several times while describing storage solutions that apply integrated de-duplication practices to arrays and blade-based storage devices. Joyce said that the competition's point products for dedupe "aren't designed to together," selling the new integration of network, server and storage mantra of HP. Moving stored data to another datacenter introduces extra steps to "unduplicate, move, and then deduplicate it again," Joyce said. "It's complex and not easy to manage."
StoreOnce operates with HP-UX as well as Windows and Linux -- but thinking of the software as environment-based misses the point. It's built for storage appliances to use, processing data that streams over a network. (HP's Unix needed to be checked by a press rep before Joyce could affirm the environment is supported. Linux and Windows needed no such checking.) StoreOnce is a migrator's benefit, if the site can justify backup service to hundreds of desktops. But its benefit soars if a site is moving from traditional tape and offsite archival practices, like so many 3000 shops have used over the years.
Joyce said that eliminating IT sprawl was the goal of the HP storage releases, but that sprawl is defined in terms that might overwhelm a modest sized 3000 shop. A new StorageWorks P4800 BladeSystem SAN can start configured small and grow, he said, but small in this definition is $20,000 per storage node. StoreOnce and the new BladeSystem SANs are part of HP's Converged Infrastructure, "hardware and software architecture that scaled to support thousands of virtual desktops. HP's release noted that the storage solution supports 1,600 users at 50 percent less cost and 60 percent less space than traditional client virtualization implementations."
Joyce said the new releases offer de-duplication "baked in" to the products, rather than added on with virtualization libraries. De-duplication is a concept as
old as the initial HP deadline for leaving the 3000 market, starting in
2006. The concept reduces large amounts of backed up data which can chew
up disk space. Software algorithms use hashing to tell whether backed
up data is unique. HP says its dedupe reduces storage needs by at least a
factor of 20:1. HP says its solution is twice as fast as a comparable
EMC product, too.
HP introduced its first dedupe solution in 2008. At that time, dedupe was being offered integrated with a HP StorageWorks disk-to-disk D2D system, or in software. But every such analysis of what not to back up can slow down backups, something that Joyce said HP needed to improve upon. HP's StoreOnce is built to solve the problem while it works with both the D2D and EVA storage units. What's more, HP will be able to integrate storage arrays from EMC, NetApp, Sun and IBM in its newest EVA Clusters. The vendor is not about to let a competitor's storage create what it calls "islands of storage."
Storage at HP has become a dominant business in the company's enterprise sector, riding the kick-start of advanced Compaq solutions integrated as part of the 2002 merger. Joyce said that dedupe solutions are one of the top two sparks for the business, the other being iSCSI devices. EVA is probably overkill for the majority of the 3000 market looking at migration storage solutions, starting at $63,600 for a base unit without drives; the cluster can be as big as 2 petabytes and contain 2,000 drives.
The and D2D2500 D2D4000 systems are more in line with the average storage needs for the SMB customers the 3000 has served, priced from $4,600 for 2 TB to $17,000 for 4TB of capacity. A new $93,000 D2D4312 extends the capacity to 48 TB. The integration means the arrival of StoreOnce, suited for the D2D devices, delivers more immediate benefit for small shops.
HP's building storage solutions that scale far larger than the needs of even a half-dozen servers, its hardware and software built to serve multiple operating environments and systems through CommandView browser interfaces. Channel partners have these dedupe solutions in house for testing this month, Joyce said, establishing what he called "proof-points" on the new technology. With StoreOnce, "this is one where we think we can lead the market."
June 22, 2010
Emulator boots with sponsor partner push
Stromasys has announced a new timeline for the only emulator project in the 3000 community, but it's more than a beta-to-release schedule. The vendor that sells "cross-platform virtualization" is starting the Zelus product's outside life with an outreach to 3000 vendors and first-wave customers. Stromasys needs partner input on how to manage the third parties' checking for valid licenses using a software-only emulator.
In an interview a few days before this week's announcement in Las Vegas, Stromasys CEO John Pritchard and CTO Robert Boers said that HP has released a pair of system ID strings (HPCPUNAMEs) that the software can use to mimic a 3000 model. One name is a larger system than the other, which gives Stromasys two levels of virtualization performance to sell.
The most crucial hurdle to achieving this first call for sponsor partners has been cleared. HP gave the company the secret code that permits PA-RISC hardware to continue to boot up MPE/iX. This processor dependent code was a year delayed coming out of HP, 12 months that Boers wishes he had back in the product's lifecycle. But that loss of time won't eliminate the extension of life that the product called Zelus will provide. Once the Stromasys team demonstrated its software would boot Linux, HP figured out the paperwork it needed Boers to sign.
Support for the software will come from centers in both the US and in Europe, he added. Stromasys, which was founded 12 years ago when managers of the European Digital Migration Center bought out their HP owners, has operated an office in North Carolina for years. Pricing for the product has not been set; the Stromasys executive team is taking imput on how much the 3000 customers would be willing to pay. At the moment the Digital product CHARON sells for between $5,000 and $200,000. Even the latter price represents a big savings over a code move or migration.
The company's history and credentials with HP have been important in gaining access to the technology which HP has held close since the 9000-to3000 lawsuits of the 1990s. As for speed potential, Zelus is going to bypass all the slowdown code HP inserted in MPE/iX to hamstring the HP-built PA-RISC CPUs. HP was interested in the choice, but didn't insist. Speed is going to increase as efficiency and power of the hosting hardware grows. Even HP's Converged Infrastructure products, many of which are unavailable on PA-RISC hardware, can help this solution grow faster.
Earlier in the emulator story, HP inserted a clause that an emulator product needed to run on HP hardware. Since there's no technical means to check this requirement, it's a moot point on an operational level. But writing it into HP's agreement doesn't make it a legal or moral point, either. Boers noted that HP cannot insist on this, since the 1973 court ruling gave Amdahl the right to run IBM's OS on non-IBM hardware. It's nice to have a Department of Justice precedent that's almost as old as the 3000 to rely upon. HP's not coming after anybody who uses Zelus on a Dell server.
The most immediate task is gain the support of the third party software providers. They maintain and sell tools and apps which customer will require to operate on a non-HP PA-RISC implementation of the virtualization. "In the past this licensing was linked to a hardware model," Boers said. "Now that linkage is gone, and we need to understand what issues the software providers might have with that. We don't want to deprive them of revenue."
Boers mentioned an option of an honor system, where vendors rely on the integrity of their customers, or something that verifies a match between the HPSUSAN number from the 3000 and the software's ID. Every license of the Zelus product must be matched to an existing 3000, after all. HP didn't allow any emulator to create new places for MPE/iX to run.
"We can emulate a system serial number as a standard ID," Boers said. HP wants no part of managing this portion of the transfer of an MPE license to Zelus, he added. There's the $500 fee for the transfer HP wants to collect, probably the last fee a Stromasys customer will ever pay to Hewlett-Packard for MPE operations. There may be another path, but it's an easy guess that the honor system is as unlikely a choice as a Zelus price below $2,000.
Stromasys says it has sold more than 3,000 licenses to the Digital customers who need to remain on Alpha and PDP architectures, and it also announced a Sun Microsystems SPARC virtualizer ready for shipping this summer. The timeline above calls for six months of recruitment and work with customers and software suppliers before Zelus enters a beta-test process. Sales to customers are scheduled for the second half of 2011.
One of the first places Stromasys will look for sponsor partners is in the independent support community for the 3000 homestead sites. These companies will be first-line support for Zelus, since the product will make Intel-based systems look identical to PA-RISC hardware, right down to accepting all internals calls from MPE/iX. In practice, this virtualization needs fine-tuning, so the support companies like Pivital, the Support Group, Allegro and Beechglen and GSA, the MPE Support Group and more will have to learn about Zelus -- enough to handle first calls. Stromasys will support these companies, so having an MPE/iX expert like the firms listed will be essential. Self-maintainers of 3000s may not be the best prospects for deploying Zelus.
Here's how the Stromasys announcement describes the community supporter-Stromasys relationship.
Software running on Stromasys’ virtual systems remains easy to maintain, modify and extend, as the virtual system provides equivalent functionality to the original hardware platform. For maintenance of their MPE/iX operating system or applications on the virtual HP 3000, users can continue to work with their existing support organizations, for which Stromasys will provide training, to be announced at a later date. The ability of legacy software to be maintained in the usual way is particularly important as these applications typically embed substantial business practices that are very costly to redefine and implement on a new platform.
With a new timeline in place that involves the 3000 community's established resources, and the HP technology in hand after a disappointing delay, it's possible that this solution will extend MPE/iX software use for many years to come after 2011. Boers even sees a potential for HP's own services organizations to deploy Zelus when it's the best choice for a 3000 site. HP already does this for CHARON in the Digital market, a way for HP to retain the VMS support contracts even as a customer leaves HP's hardware.
These kinds of relationship issues have become the crucial elements in the success of the Stromasys product. Technology choices used to create Zelus are in place in other Stromasys products, so even sales of the product can be given time to grow. Homesteaders are still looking for a way to keep 3000 program business logic running without an environmental overhaul -- and gain access to new technology for storage like iSCSI. Storage is big business at HP's enterprise unit today, so the tacit support of Hewlett-Packard could play a part in the potential of Zelus, too. Enterprise sales opportunity can make for unexpected bedfellows.
June 21, 2010
Vast 3000 rollout echoed in coming week
More than nine years ago the HP 3000 world experienced the sort of system news that's is due to emerge from this week's HP Technology Forum and Expo. When announcements of new HP Converged Infrastructure elements, covering blades and storage and more spill out in press conferences scheduled for the end of today, 3000 owners might do well to recall how sweeping HP made its upgrades to the line back in the spring of 2001.
"At no time in recent memory has the lineup for 3000 ownership been reset so," I wrote. "It is now composed entirely of systems just announced with a new architecture, or computers whose end of support date is already known. The HP 3000 division expressed enough confidence in the new offerings to sweep everything else in the 3000 product line aside by the year 2006."
The architecture transformation came in the IO and network subsystems, not in the main CPU. And there may have been a coded message HP sent to the installed base by making 2006 its end of support date for the Series 9x7 through 9x9 3000s. Within nine months this entire HP 3000 line, including the carpet of brand-new systems, was to experience that same end of life date. Customers were only starting to assemble budget to buy the new systems when HP pulled the plug on the 3000's future.
But for a sweet period of late spring through late fall, the outlook for a hardware renaissance was bright, and the 3000 had attained parity with its Unix brethren. Viewing the feeds and speeds of that last decade's hardware, it all seems built of another era of technology. However, it would be hasty to assume the '01 revamp was the last of the 3000's architecture. There's another path to evolution, building in the wings.
The moment you hear about fresh HP blades, whenever you hear it, you might be like me and recall the scene from The Empire Strikes Back. (Forgive this sci-fi nerd his devotion to ancient cinema, released when the 3000 was in its prime.) Luke jets off in his fighter to save his friends, before he completes his Jedi training. Obi-Wan despairs in the wake of the engine roar.
Obi-Wan: That boy is our last hope.
Yoda: No. There is another.
Another, you scoff. How could there be another generation of HP 3000s, when HP ceased making them in 2003 and is shutting 3000 operations entirely this December 31? There's the emulator project, or as Stromasys calls it, a "cross-platform virtualization" software product.
With the right combination of 3000 community support and development advice, along with elastic pricing to suit budget-challenged homesteading sites, new hardware as sweeping as the 2001 release could provide another. We'll have more on that project's timeline tomorrow, one of the few pieces of 3000 homestead news we expect to emerge from the Tech Forum.
June 18, 2010
MB Foster hosts Eloquence training
The creator of the Eloquence database and language is coming to North America for a thorough session on deploying and using the product that understands IMAGE structures better than anybody's. MB Foster will host Michael Marxmeier's class on Eloquence 8.0 and 8.1 July 29-30 at the MB Foster HQ in Chesterville, Ontario, just a half-hour outside of Ottawa.
The training comes at a significant time for the migrating 3000 community. The 8.1 release of Eloquence has been rolled into the market for more than six months by now, so questions on the database's new features will be on the minds of its users. Of course, the countdown to the post-HP era of 3000 ownership will be under six months when the training takes place. Companies motivated to learn new technology to get off the 3000 are polishing their migrations this summer.
The training is $500 for the two days, but MB Foster is offering a discount for early registrations. Even at the full price, the class is likely to be one of the least expensive elements of the training trip. But it's a rare thing to learn a product at the hands of its creator.
Attendees who sign up by July 9 pay only $400 for the two days of education. Registration is through the MB Foster offices, 800-ANSWERS.
The agenda for the class:
- What is new in Eloquence 8.0
- What is new in version 8.1
- How to get more from your Eloquence environment
- Maintaining an Eloquence DB
- Using Forward log files
- How to recover after a disk failure
- Using Encryption in Eloquence
- How security works in Eloquence
- Using Eloquence with MB Foster products
- Support capabilities and options
June 17, 2010
PostgreSQL vs. MySQL: Research today
When you're migrating in-house apps away from an HP 3000, replacing a database is a crucial choice. Much of the 3000 community is moving such apps to Eloquence for good reasons: the database is affordable, it understands IMAGE structures and programming, and the solution is loaded with lots of extras. Eloquence is full-featured enough to be considered a language.
However, your migration choices might be dictated by your IT personnel. If your company is losing its 3000 expertise and you're exploring the world of open source databases, eventually you'll look over SQL options. They don't all have to be spelled Oracle or SQL Server. You can get a comparison of two open source tools in an online seminar mid-day US time.
EnterpriseDB, advocates of using Postgres, is sponsoring a GoToMeeting webinar starting at 1PM EDT today. The invitation says in part:
For years, the common industry perception has been that MySQL is faster and easier to use than PostgreSQL. PostgreSQL is perceived as more powerful, more focused on data integrity, and stricter at complying with SQL specifications, but correspondingly slower and more complicated to use. Like many perceptions formed in the past, these things aren't as true with the current generation of releases as they used to be.
If you register, the company will email you a link to download a recording of the seminar, even if you can't attend.
• Pros and cons of using PostgreSQL or MySQL
• Ongoing trends towards using open source in the enterprise
That last item will be of interest to the migrating HP 3000 site that's trying to contain costs and control its own future. One well-regarded vendor of 3000 applications, QSS, has been migrating its customer base slowly onto such open source tools as Ruby on Rails. You need relationships in technology to stay stable and vital, but the old sort of arrangement where a vendor can cancel your platform grows more passe with every day. Vendor lock-in is something that open source is relegating to a "perception formed in the past."
June 16, 2010
Go for the Gold Book
An HP 3000 staple went on sale at eBay this week. The tool contains no toxic elements, is mobile and can be easily customized. HP 3000 experts say that keeping it maintained leads to stable system environments. The tool dates from the 1980s, but it's never going to go out of date.
The staple is the HP 3000 Gold Book, known as the Hewlett-Packard System Support Log. HP 3000s were shipped with a Gold Book, where the diligent and professional IT manager kept log-on information, system configurations, serial numbers, support handles and more. Paul Edwards, a treasure of the 3000 community who devoted a career to teaching and developing HP 3000 skills and software, has preached the lesson of keeping a good Gold Book.
The Support Log on eBay is just a tool to organize key information about HP 3000 documentation, but it's unusual to see such an operational resource for sale. Only $9.95, but worth every penny if your homesteading shop doesn't already have one that's up to date. Edwards wrote a white paper on homesteading practices, free for the downloading from his website, that features Gold Book use for documentation. Docs, of course, are the story of the 3000 you leave as your legacy of your work at a company. The Gold Book even has a useful function when your homesteading site goes into migration mode.
Edwards explains in his paper:
The Systems Manager Notebook is vital to the proper management of any HP 3000 site and consists of many parts. Every site should have one because it contains critical hardcopy information to back up the information contained on the system. It is part of the Disaster Recovery Plan that should be in place and is used to manually recreate your environment.You can’t have too much information in it.The pages in the Gold Book supplied by HP with each system should be used to record ongoing details about each system. All parts of the notebook have to be kept current at all times. This information will be valuable when migrations are planned in the future.
Details in the Gold Book can save a system that has endured a critical failure. Edwards says that the book should include "BULDACCT.PUB.SYS, a listing of BULDJOB1 for MPE commands to rebuild the accounting system structure and settings and a listing of BULDJOB2 for the MPE commands to rebuild the COMMAND.PUB.SYS file settings."
Mervis Equipment and Electronics Equipment Exchange of Kokomo, Ind. posted the eBay notice that expired yesterday. (Alas, without a single bid; you can contact Mervis about the book at firstname.lastname@example.org, or phone them at 765-454-5833.) But even if you're not ready to spend a pittance on this tool, you can create your own, working with the details that Edwards defines in his homesteading white paper. Documentation, after all, is the easiest way to establish a path toward a stable future, whether that's on the 3000 or in another operating environment. You must first sustain your IT environment, documented well enough for others to understand, before you can make a transition.
June 15, 2010
Red-blooded sites shape new schedulerThe new Windows-based MBF Scheduler grew up in MB Foster’s labs, nourished by the experience of engagements with several sites migrating from the 3000. The migrating sites across several business sectors all needed the scheduling and jobstream power built into the HP 3000. They didn’t need, or want, any Unix in the scheduling loop.
The software is another step toward new business for MB Foster. Over the past several years the company has transformed itself into a services provider which employs software. MBF Scheduler was built, tested and deployed by the company’s engineers, a team that’s been intact since 2001, all based in North America and employing 3000 experience.That’s the same lab that served HP when the vendor made a basic-level ODBC product available as part of the OS. MB Foster licensed its ODBCLink in an SE version to HP, then supported it for 3000 sites using the bundled SE.
Although it came out of a seasoned lab, the new software’s pedigree grew out of requirements of customers in sales, manufacturing and the Canadian government, according to Chief Technical Officer George Marcinek.
“Given that we started from MPE, it gave us a very good starting point,” Marcinek said. “But the availability of true customer requirements was equally important. Sophisticated customers could tell us exactly what was important to them, versus what would be nice, and what was just window-dressing. We could vet all of our design decisions based on that.”
The 3000’s depth of scheduling was integrated into the environment from the early days of system delivery. MB Foster’s cloned feature set reminds migrators of what they’ve learned to rely upon.A master job queue, input priorities, job fence, job limits, and other MPE-specific job scheduling features are built in. Windows users can automate all of these features by using the MBF Scheduler command line interface. Command line utilities gives managers the ability to submit, control, and monitor MS Windows Server jobs. Batch grows more 3000-like with a feature that includes MPE-style job cards in Windows batch files.
Those who mean to take the Windows environment into the datacenter are among the target customers for MBF Scheduler. The company also wants to engage Windows enterprise users who have little to do with a 3000.
“We were looking at the people who are in the Windows workspace,” Marcinek said, “not just because they are coming from MPE. The people in the Windows space who come from MPE may have some additional, very specific requirements. But we tried to generalize the requirements in such a way that they will be equally applicable to the native Windows environment.”
Up to now, Marcinek added, the Windows users who wanted to cover 90 percent of the 3000’s scheduling features did it themselves, with advanced knowledge of scripting in-house or from outsourced services. “They can build arbitrarily complex solution to scripting needs,” he said. “We’ll be competing with existing products as well as internal R&D resources. We can compete for the customers who are looking at two factors: price and feature set.”
The product has grown up from the production needs of “true, red-blooded sites,” Marcinek added.
June 14, 2010
Delete empties, checks on batch, and moreI have a group that contains flat files that get FTPed to our network by a job on the HP 3000. Before FTPing the files, I would like the job to delete files that are empty, but the problem is that the file name is never the same. How can I use MPE/iX or MPEX to determine which files in the group are empty and delete them?
Robert Schlosser replies:
In MPEX you can say PURGE @(EOF=0) and purge all files with no records
Dave Powell adds, for those who have only standard MPE/iX at hand:
If you don’t have MPEX (gasp), you can still :listfile into a msg file, and read it back, call finfo to check the eof, and you are in business. Knowing how to do this can be handy if you want to do something to a fileset that isn’t a one-line MPEX command.Here’s an example slightly dumbed-down from a routine we use to do things only to files with more than seven records. You could adapt it by changing “> 7” to “= 0” on the third finfo. It’s a bit fancier that you really need, because I insisted on echoing the file name and eof to the job listing, padded with blanks so the columns add up.
!PURGE MG, TEMP > $NULL
!FILE MG; REC=-72,,F,ASCII;MSG;DISC=5000,32,1;TEMP
!LISTFILE fileset.group,6 >> *MG
!SETVAR _CNT 0
!IF FINFO("MG",'EXISTS') = TRUE
! WHILE FINFO("MG","EOF") > 0 DO
! INPUT _LONGNAME < MG
! SETVAR _NAME_X lft('!_LONGNAME',26)
! IF FINFO(_NAME_X,"exists")
! SETVAR _EOF FINFO(_NAME_X,'EOF')
! SETVAR _EOF_X RHT(' !_EOF',6)
! IF FINFO(_NAME_X,"eof") > 7
! ECHO File-Name = !_NAME_X EOF = !_EOF_X
! your commands go here, like maybe "purge !_NAME_X" or "print
! SETVAR _CNT _CNT + 1
!IF _CNT = 0
! your "nothing found" routine goes here, if any
Is there a clean way to determine, either in a command file or programmatically, whether I am running in batch? I don't want to simply check the variable HPINTERACTIVE or HPJOBTYPE. Those don’t do what I want if we are in a session that was created by a REMOTE HELLO from a batch job on a remote machine. I want a way to know that such a “session” is really “batch.”
Cathlene Mc Rae of HP's Response Center replies:
When a job is streamed on a system the hp variable set for the job are as follows:
HPSTREAMEDBY = CATHLENE,MGR.MCRAE (#S102) HPJOBTYPE = J
HPINTERACTIVE = FALSE
When the job does a dsline system and a remote hello the following variables are set for the remote session:
HPSTDIN_ACCESS_TYPE = NS/VT
HPVT_CLIENT_LDEV_NUM = JOB INPUT SPOOLER HPVT_CLIENT_JOB_NUM = #J14
HPVT_CLIENT_JOB_NAME = MGR.MCRAE
HPINTERACTIVE = TRUE
HPJOBTYPE = S
HPSTREAMEDBY = CATHLENE,MGR.MCRAE (#S102)
The variable set when a session does a dsline system and a remote hello are:
HPSTDIN_ACCESS_TYPE = NS/VT
HPVT_CLIENT_LDEV_NUM = 3
HPVT_CLIENT_JOB_NUM = #S102
HPVT_CLIENT_JOB_NAME = MGR.MCRAE
So, how can you tell if a job or session started the remote session?
The hpvt_client_ldev_num will be may be set to “job input spooler “ and hpvt_client_job_num will be “#J”.
Tony Summers adds:
Please note, the variables won't exist if it's a local session. In the past I've used the BOUND command to prevent command scripts from failing.
IF (HPVT_etc = "whatever")
Walter Murray, who posed the original question, commented:
The most clever suggestion was from Jeff Vance, who pointed out that you can try setting HPTYPEAHEAD to TRUE, and then check HPCIERR. Sure enough, that works!
In the end, I decided to go with a technique that several others suggested, and Cathlene amplified. There are a number of HPVT_xxx variables that are sometimes set. I decided to check HPVT_CLIENT_JOB_NUM.
Here's what I came up with:
COMMENT * Are we running in batch or interactively?
SETVAR RUNNING_IN_BATCH FALSE
IF HPJOBTYPE = "J"
SETVAR RUNNING_IN_BATCH TRUE
COMMENT * A real session, or REMOTE HELLO from a Job?
IF TYPEOF(HPVT_CLIENT_JOB_NUM) = 2
IF LEN(HPVT_CLIENT_JOB_NUM) > 2
IF STR(HPVT_CLIENT_JOB_NUM,2,1) = "J"
SETVAR RUNNING_IN_BATCH TRUE
I face a serious situation on our HP 3000. A batch job which used to run for about 30-45 minutes (stores data in an IMAGE detail with about 3.2 million entries) runs now 4 hours and longer to handle the same amount of data. What should I look for?
Brian Donaldson replies:
Did anyone make program changes? Or change locking strategies in the program -- locking data sets and/or data items and not DBUNLOCKing until the program completes, instead of DBUNLOCKing immediately after an update (DBPUT, DBDELETE, DBUPDATE) has completed?
If TurboIMAGE is the problem, then you can investigate the possibilities of synonym entries in your automatic and manual masters.
If you just use HOWMESSY, however, it doesn’t give you info regarding the offending entries causing the problems. It will just give you a percentage of synonyms in a particular data set, "SET-A has 23% synonyms." Not very helpful if you want to find out the offending entries that caused the collision in the first place.
data-set-a (master), key item = data-item-a
data-type= X4, key-value="ABCD"
hashes to IMAGE record 1234. Record 1234 already occupied by key value “XXYY”. So then IMAGE has to go find a slot to place key value “ABCD”.
With each DBPUT this hashing/collision nightmare could slow you down immensely.
If TurboIMAGE is your problem you can try changing the capacities on your offending automatic and manual master data sets.
Integer keys on master sets work fine if the key value is between 1 and the capacity of the data set.
When a key value is greater than the capacity then watch out -- the synonym problem will bite you. Generally speaking, “X” type key items (search items in detail sets) work best.
If your app uses integer keys, you cannot just change the data types of these keys without having to modify every piece of source code that manipulates these data sets. Then you have to recompile the entire application. Not exactly the optimum solution.
If you determine that TurboIMAGE is your problem I suggest you go straight to the big guns for assistance and call the gurus at Adager. Those guys will be able to help you best, I’m sure of it.
June 11, 2010
Updates on Extensions to the 3000
There are a few stories budding in the community this week about extending the 3000's future, but they're not quite ripe enough to pluck for all the details. You ought to know anyway what's in the works to make a 3000 work harder.
The stories cover software for the 3000, something a lot easier to change than the hardware in the eighth year since HP built a 3000. The long-awaited and much-debated 3000 emulator is making progress, according to the Chief Technical Officer at Stromasys. By the week of the coming HP Technology Forum, the company will update the community about how long it will take to move the "cross-platform virtualization" software from testing to supported product for sale.
CTO Robert Boers extended a few details in advance of the announcement. HP's technology to permit a boot-up of MPE/iX is being shared with Stomasys. Boers calls it a "technology transfer," adding that he's signed several HP non-disclosure agreements to gain access to the knowledge.
For more than a decade, the technology that turns a PA-RISC server into an HP 3000 has been discussed in public by the community. Most reports concur that Processor Dependent Code (PDC) is written into stable storage on a PA-RISC system -- and MPE/iX looks for that code during the cold-boot process. Any virtualization solution is going to need the magic code to feed to MPE/iX so a boot could succeed. It says something about HP's goodwill to the community that any transfer of technology is going forward in 2010, the final year of HP's 3000 operations. We'll have more details on this story soon.
Then there's still activity coming up for open source software users managing 3000s. Way back in 2004, HP advised homesteaders to get savvy on using open source, if homesteading post-HP was to succeed.
Open source opening soon: Brian Edminster of Applied Technologies checked in a little while back to let us know that his outstanding project to corral open source for the 3000 is still forthcoming.
I enjoyed your recent Community Counts Homestead Resources post, although it's a painful reminder of my lack of recent progress with getting MPE-OpenSource.org online. If nothing else, it was a swift kick to get me moving faster on it, and a not so gentle reminder that time is a precious commodity too.
Edminster, who's been managing the 3000 IT operations for Host HMS duty-free shops as part of his business, promises to update us on the open source site.
We'll get back to you with details as they surface on these efforts to extend the lifespan of HP 3000s.
June 10, 2010
Forum summons Woz, plus Who's founder
From the groupies of '60s rock to the tech geeks of the '70s, this month's HP Technology Forum is summoning celebrity to attract IT attendees. Steve Wozniak, Apple computer founder, and Roger Daltrey, former front-man for The Who, will be on hand at the conference that shows off HP's latest for those customers sticking with their vendor after migration. It's a historic event, judging by the age of the headliners.
Wozniak is the guest star at a rollout party for Fusion IO's solid-state storage-accelerator product, dubbed the HP IO Accelerators. Fusion IO got serious about storage solutions last year with its ioDrives, deployed at MySpace's datacenter. Fusion likes to point out that an ioDrive is not a Solid State Drive. The Fusion hardware, which integrates with a server, is supposed to run rings around SSDs in performance. But that 2009 device is sold in capacities that will remind you of a drive (80, 160 and 320GB). In addition to performing with Linux, Fusion's product runs with Intel-based servers using Windows, the most popular alternative for the migrating HP 3000 shop.
You can shake the hand of the man who plucked technology's first Apple at a beach party at the Forum. A beach party in the desert of Las Vegas -- nothing unusual for Vegas, where an Interex show once included custom beach-towels passed out by Robelle at a Wet 'N Wild splash party -- is scheduled for June 22 at 8:30 PM. We've been invited to the Woz-fest, and were also told to pass along registration details to our customers (that's you):
Go to www.hpandwoz.com/register and use a password of wozinvegas.
As for Roger, he's having his first Las Vegas appearance since the Who's Reunion Tour of 2006. The Connect user group, organizing much of the content and entertainment for the week, has promised its most loyal members a chance at access to the rock star who was nominated for 1978 Golden Globe playing Tommy.
For Tech Forum attendees who add a Club Connect VIP Pass to their registration, the user group will enter their names into a drawing to meet Daltrey, or the show's other music headliner, the Goo Goo Dolls. Rolling out on another tour this year, The Who's frontman has been performing and writing since before the HP 3000's predecessor, the HP 2100, booted up in the '60s.
The closer access to the music comes as a bonus for registering for introductory training that Connect calls You-Conn. Hands-on labs on June 21 in EVA management tools and techniques, or Network Forensics and Analysis: Inside the Filth of Network Communications, include a VIP Pass and a $100 Home Depot gift card. Such perks tacked on for one-day sessions in tech wizardry like HP-UX system admin or virtualization.
Registration for the conference is still underway at pre-show pricing at the HP-Connect website. Despite Daltrey's lyric from his song My Generation, no one left at the top of this bill will die before they get old. Keith Moon and John Entwistle have already departed the band and headed into clouds of a non-technical kind.
June 09, 2010
New Channels and Tactics for 3000 Growth
MB Foster gained the insights and experience of a veteran this year when David Greer joined the company as Director of Marketing and Sales. Greer posted more than 20 years of accomplishment while developing and managing business at Robelle. He then took about eight years away from the HP 3000 marketplace, working elsewhere in the IT industry, so he's returned with a fresh outlook. We asked him in a Q&A interview what non-3000 experience brings to his return to the community.
What experiences during your 3000 hiatus can contribute to your work now, on your return to the community?I spent three years doing early stage work with companies. For many cases for me it was helping out in marketing and sales, working out the messaging. In working with startups, you have really young, really bright people. The speed they operate at, their total lack of fear in trying new things like social media, or recording a video of the CEO and putting it up on their Web site, that's all quite outside what we're used to in the HP 3000 world. It's refreshing and gives you a new view. Some of those ideas we're working with actively at MB Foster.
What are the prospects for sending messages to the older generation of IT managers about experimenting with new technologies or social networks, based on your start-up experiences?
Enterprise computing is going to be conservative because they have to keep their businesses running. They have to make big investments on technologies that have lasting power. Most of these can't afford to be flash in the pan around their core application. If you want to do e-mail marketing, or take something off to the side to see if it works, most organizations can take a risk on that. Eventually, if it works, you have to integrate it - and that's when it really gets interesting.
Social networking is an issue that's independent of IT. I think we have a generational difference. While you and I both have blogs, a Twitter account, and are up on Facebook, we're the exceptions. People of the baby boomer generation are more likely to want to print things out and read it on paper.At MB Foster we're thinking about what our audience wants. If they want to read from a blog, I'm happy to publish one. I'm still looking at what we'll want to publish in that medium for them. There's transition and homesteading to deal with now in the 3000 market. Does it seem to have gotten more complex when you make plans for providing services?
When you go to a deeper engagement with a customer, it's no longer about the infrastructure or about the technology. It's about the business. That's a different message, a different group of people and a different way of thinking about it.
Over the past quarter MB Foster took on the EasyIQ product line, software created in Europe, and will be supporting and providing it to North American customers. Is this a practice you'll want to pursue again soon, or absorb the lessons from it first?
We see this as a strategic future for the business. It's a long-term play, one we're committed to, and a major new initiative this year for the education market. For the EasyIQ piece, only half of it is the product's technology. Half of it is MB Foster's integration and data expertise. The real value we will provide with Easy IQ is its ability to integrate with many different applications in education.
One of your first missions after you returned from the Med was ramping up a solution in the scheduling field. With 3000 companies migrating, is there an opportunity to improve the range of scheduling and bring 3000 features to other platforms?
There is very little out there that gives you MPE-like job queues, job fences and other features that MPE people are used to for job scheduling. The products I know of so far almost always require at least a Linux server somewhere in the mix, because most of them got written for a Unix environment at first. The Linux server coordinates in the background. For those companies that want to go to a pure Windows solution, I'm not certain there's a pure one out there that emulates most of the capabilities of MPE. That's why we are releasing MBF Scheduler, MPE job scheduling for Microsoft Windows Server.
June 08, 2010
First 5 Years Delivers News Built to Last
It began with an ending. Five years ago on this date the 3000 NewsWire's blog came to life, celebrating a notable death. Bruce Toback, a man of deep technical prowess and great humor, had passed away in that week of 2005, claimed by a heart attack that cut his life short too soon. He once noted a study which reported 10 percent of all tech gifts would be damaged after the year-end holidays by enraged low-tech users, then added, "Go team!" You might say the same about the cutting short of HP's 3000, except that only its HP life was cut short by 2005. We like to think we've lead the cheer over these years from the community of "Go team!" even as many have gone away.
Five summers ago we started our first week of workday articles writing about Quest Software's tools and a claim they'd already migrated 100 HP 3000 sites; about HP hiring this Todd Bradley fellow from Palm to run its PC group (Bradley bought up his old company for HP this spring); and a popular 3000 community topic of the day, open source code. Our story of June 15 was about Sun's new sharing of source for its Solaris flavor of Unix.
Some 203,000 page views later, after 1,320 articles, we know for certain that open source isn't a good answer to propel the 3000's future. Even while Sun opened up its OS internals (not that it did much good for Sun) the 3000 didn't have the same vast populace to enhance and maintain its OS. We said five years ago that a better quest for extension was in order.
Although MPE/iX's future development will have to take place in the third-party developer community, open source wouldn't work for the 3000 -- something most customers realize when they get honest about the size of the 3000 development base. You can't count up customers to measure the potential of open source resources; you have to look for people capable of doing their own builds of software such as perl, sendmail and the like. HP's Mark Bixby has warned 3000 customers who want to homestead they better get fluent in such development, or get to know a consultant who knows his way around the make command.
The summer of 2005 offered some once-in-a-career moments, like the overnight meltdown of Interex and the 30-year-old user group's conference; the last Systems Improvement Ballot to enhance MPE/iX, a document that didn't get a hearing at a conference because of said meltdown; and an HP conference postponed by a Category 4 hurricane. Interex lost millions, the SIB was reduced to wishes, but that HP conference roars to life once more this month in Vegas -- where the hurricane threats are few.
None of our many reports would be possible, however, without the steady support of our sponsors and avid readership of the community. Expertise, bandwidth, dedication and persistence come at a cost, one that our advertisers have believed is important to bear on behalf of 3000 users worldwide. This blog became the first step around the world, using system boots built to last like the 3000 itself.
In five years of reporting we've only gone offline once, when a six-block, five-alarm fire in San Francisco took down the servers of the NewsWire, LiveJournal, Second Life, Craigslist, Facebook and Yelp. We were restored along with the big boys within a day. Our 15-year-old gateway address, 3000newswire.com, has been offline from time to time thanks to the unpredictability of communication carriers, but the alternative 3000newswire.blogs.com address serves day and night. Moveable Type powers the NewsWire blog engine, but we admire the technology of WordPress, too. The latter is as open a source as any we're seen.
So Abby and I toasted today to five years of lifting our up-time, with glasses of icewater lifted at breakfast at The Frisco, the last remaining diner in the NightHawk chain here in Austin. The Frisco is a classic that's been refurbished and relocated to bigger digs, but still retains its sweeping breakfast counter, banana creme pie and chili over eggs charms. We still print ink on paper and size up headlines, buy staples and envelopes and fill post office tubs. But it's been fun to post and report quickly on this blog, even more fun to tweet on Twitter. (Follow us @3000newswire for the latest.)
By the time we started this online community rally point in 2005, we had no idea the 3000 world would never again see a major conference to pay heed to its needs. We now tend to gather around screens and smartphones and even tablets to stoke our community's fires. For us, the first steps started in the hot months of five years ago, when even the cooling embers of HP's dwindling activity tossed sparks that we turned into stories.
Thanks for reading and supporting us. We look to our next five years with relish and curiosity, as we push toward 20 years of HP 3000 storytelling.
June 07, 2010
Apple developers aim with 3000 experience
Today's announcement of a new operating environment for Apple mobile products recalls the days of HP's MPE/iX rollouts. Some of the same developers who studied those HP 3000 OS releases are poring over the internals of iOS 4, Apple's new name for its latest mobile environment that was once called iPhone OS.
Michael Casteel retired from the active 3000 community more than a decade ago, but he still remembers his time developing the scheduling app Maestro from Unison to serve enterprises using HP's mission-critical server.
"I miss the trenches now and then, and the folks, too," he said. "Some of the best times I ever had were tracking down obscure bugs in Maestro after everybody else had given up. I do like good puzzles.
"That’s probably part of why I keep hacking on computers. No more 3000s, though I miss them a bit, too. They were a lot of fun, for a lot of years. From the Series II to the 9xx’s. Now, it’s programming and a little web site maintenance using my MacBook."
Casteel is the author of Klondike, the best Solitaire program ever built for the Mac -- through three generations of OS -- now sold for the iPhone and iPad. Casteel is attending this week's WorldWide Developers' Conference for Apple developers. He's not the only programmer with MPE in his blood camped out in the Bay Area this week.
Others who are working toward a release on iOS 4 include Neal Kazmi, the lead developer for Minisoft and its connectivity software. Kazmi is creating an iPad version of Javelin for release this summer. Today the developers at the WWDC received a Gold Master release of the new OS. The move reminds us of the early releases of MPE/iX shared with 3000 developers, often using the organization of the SIG-SOFTVEND special interest group chaired by MB Foster's Birket Foster.
The HP 3000, born in the early 1970s, is scheduled to getting its first iPad app this summer.
The distance between the two solutions’ release is close to 40 years, but 3000 managers and developers are thinking ahead to a time when staying attached to a keyboard, wired to a backplane or even a PC, will seem antique. And even though the 3000 is looked upon like a relic by some IT architects, users and software gurus want to bring Apple’s latest mobile marvel in step with their systems.
Kazmi is working on an iPad Javelin, software that combines HP700/92 terminal connections, as well as Digital VT320, IBM 5250 and IBM 3270 connectivity in a single client. “For the HP 3000 it will support Secure Telnet and NS/VT network connectivity,” said Minisoft’s CEO Doug Greenup. “It will be an app that can be downloaded from the App Store, and probably sold for $9.95.” Greenup said the software will be available in July.
Other developers aiming at iOS 4 -- ones who can claim enterprise experience with 3000 -- include Bruce Hobbs and Michael Watson, long-time engineers and programmers serving the COBOL and 3000 scheduler communities. The two said they “took a two-day iPhone development course back in November,” Hobbs reported. “I’ve also been attempting, with limited success, to lure a couple of other HP 3000 COBOL developers into a joint effort. Not sure yet exactly what we’ll put together, but I’m still hoping to have something in the App Store before WWDC.”
Apple is reaping the benefits of quantum leap power in a company-first in-house CPU for the iPhone and the iPad. HP once coordinated the hardware advances with software innovation for the 3000. The practice is still active for HP's Integrity Unix servers. New products are expected at this month's HP Technology Forum, introduced by Dave Donatelli, executive vice president and general manager of HP Enterprise Servers Storage and Networking.
Meanwhile, the 3000's hardware is frozen in the year 2003 and its OS hasn't had any enhancement since 2008. It's not the perspective that draws 20,000 developers to the biggest meeting space in San Fransico. But the size of the community is not as crucial as the depth of the 3000 knowledge base. Apple's community benefits from the experience of programmers like Kazmi, Casteel, Hobbs and Watson. They bring decades more development skills, an asset the 3000 community continues to rely upon.
June 04, 2010
One year later, HP resolved one 3000 issue
In May of last year we tracked a list of HP 3000 issues that remained unresolved by the vendor. HP had announced it was closing its 3000 group operations, including its lab. It expressed the belief that it had addressed all outstanding issues concerning 3000 ownership.
One year later, nearly all of the list of unresolved issues remains in limbo. The lone exception is the identity of the parts of the MPE/iX source code HP licensed for outside use. But only the eight licensees have learned what's included in the millions of lines of 3000 OS code. At least someone outside of Hewlett-Packard has discovered something of the 3000's secrets.
What remains secret, or undiscovered this year, are issues around support in 2011, as well as the fate of dozens of 3000 enhancements and fixes finished by the lab in 2008. There's also the matter of HP's assistance to creating an emulator -- in the event that an emulator emerges for cross-hardware virtualization.
Among the support questions without a response are those related to HP's post-2011 services as well as independent support assistance.
1. The process to unlock the HP 3000 diagnostic software hasn't been outlined. Some of these programs were built for the HP Unix servers, both HP 9000 as well as Integrity systems. As of the last report from the independent support community, some diagnostics require HP-supplied passwords.
2. The HP 3000 knowledge base from the HP Response Center has no plans announced to become available to the independent support providers. HP has said that the database includes customer-specific comments throughout, a privacy issue. But you'd think that the world's top enterprise computing company knows how to remove such comments, so known problems can be identified quickly.
3. HP's services to restore an HPSUSAN number for a system board seem to be landing in the exclusive realm of Client Systems in 2011. This kind of disaster recovery will need to continue to be available to all 3000 owners by next year, and 2011 has no HP support offers for the community. Client Systems competes with several HP 3000 resellers who also offer support. That means Client Systems will best serve the community by responding to these disaster requests as if they don't compete with some customers.
For example, Cerro Wire's HP 3000 support, for both hardware and software, comes from Genisys. Client Systems is in competition with Genisys for hardware upgrade sales. HP divided up this kind of conflict by separating system sales from support in two divisions, but that kind of separation is hard to maintain for a small company like Client Systems.
4. Throughout the past year, HP's charges to restore HPSUSANs on a per-call basis have been inconsistent, according to resellers. Published flat rate pricing seems fair, but it's still a dream.
The above unanswered questions can be resolved by HP's Worldwide Support team. An announcement of the post-2010 policies from HP Support could clear up major issues related to homesteading on a 3000. Places like the 80-year-old Cerro, which has been using 3000s since 1980, will homestead for years beyond HP's exit date.
Then there's the development and enhancement issues. Dozens of HP patches to several versions of MPE/iX remain in limbo today. This engineering passed alpha testing but didn't get enough testing from the customer community to pass beta. Some of the problem lay in HP's methods: only HP Support customers could test and report.
HP's release policy remains unchanged about patches it's created. The beta-test limbo since 2007 has seen a lot of patches check in to be built, but far fewer checked out for public release. HP was supposed to be considering reducing the test requirements. But the vendor closed its lab early in 2009 without altering the policy.
Finally, there are requests that went unresolved for years on the list, wrapped more around HP's operating policies than strategic HP exit plans. The community asked for the test suites HP used while developing MPE/iX. Denied. The community has asked for years that Series 9x7 3000s be allowed to boot up on the more modern MPE/iX 7.0. Denied. The community asked that the CPU throttling that cripples processors on 3000s be removed. This is a simple software command sent during boot-up to the 3000 system-specific stable storage. Denied.
We don't even need to get into assistance on bringing SSH security for the 3000 up to Secure Copy Protocol standards. HP started the work but didn't finish in the labs. Risk is a part of every IT solution, even HP's 3000 replacements. Remaining on the 3000 is an alternative HP understands but doesn't recommend. Resolving issues like this list from last year would show some support of homesteading. In 2009 we could still say, "There's always next year." In 2010 that's not true anymore.
June 03, 2010
Brand-new Joins the Hearts of OldThis year I lead the life of a grandfather, nourished by the newest technology. My life grew larger in April with the arrival of my first granddaughter, Paige. Although she is not as close by as our grandson Noah, the digital reach of Facebook albums and iPad slideshows bring her Houston home much closer to ours in Austin.
Paige will never know a world without Facebook or Twitter, and will probably think of film-based pictures as relics of her parents’ time. She'll wonder why we say "tape" or "film" when we mean the verb "record." We’ve all moved on toward the better of newer. But some old school practices serve very well. On Mother’s Day weekend, Abby and I tended to this tiny baby just four weeks old, while parents Maribeth and Peter got a night off and trusted us to care for Paige. Walking the floor with that little girl nestled over my heartbeat, I felt the rush of love and memory from the first days with my son. In the dark of the night, Abby lay in bed next to the bassinete, rocking it with her foot. Sleep was a blessing we all pursued that evening. Walking and rocking remain fundamentals.
But we also experienced a comfort in our overnight of tending to Paige, perhaps the same calm you can muster while you face newer technology challenges. You’ve earned your stripes and embraced one new marvel after another by the time you log 15, 20 or 30 years doing your career’s work. Your 3000 is likely to be able to take commands from a 1.5 pound mobile iPad tablet.
Mobile goes beyond phones when an enterprise company takes on the sector. This week HP's CEO said that the company didn't purchase Palm to create and sell smartphones. Instead, said Mark Hurd, "it's all about the IP," intellectual property and patents around the WebOS operating environment. HP will move the technology into mobile devices it's developing. Old technology like enterprise servers will provide a playground for the brand-new WebOS tech, once HP gets done integrating.
The world’s networks lure us into better technology, providing contact we once struggled to acquire. HP rolled out new Integrity servers late in April, and I could learn tech advances from my office, watching streamed videos. The iPad doesn’t know how to make those videos appear, since the device is bereft of Flash. But HP’s answer in the tablet derby, the Slate, won’t show them either, because HP has scrapped the device to start over with its newest operating system, Palm’s WebOS. HP now owns both, a deep reach into new technology.
HP’s purchase of Palm shows off an aspect as old as many grandparents’ practiced walks. Hewlett-Packard has a new operating system to call its own, a play it hasn’t made since the 2002 purchase of Compaq. Non-industry-standard environments were supposed to be a part of HP’s past. But just as Abby and I revived skills of bottle-feeding, HP has again embraced the ideal of technical superiority over market acceptance. Old skills resurface. I didn’t imagine I would enjoy warming bottles beyond midnight again. But there I was in that Houston kitchen, my skills improved with new tech that heats a bottle precisely in less than two minutes.
We can enjoy our new little girl via Skype video calls, the same kind of fast bandwidth that can bring a 3000 console closer to your multi-touch interface. As a graybeard I can be accused of exploring Grandpa’s Computer for my career's work. But it’s also a comfort for all of us to know that our elders’ experience enriches us, cradling the brand-new against the hearts of the old.
June 02, 2010
Fresh Eyes for the 3000 Community
MB Foster gained the insights and experience of a veteran this year when David Greer joined the company as Director of Marketing and Sales. Greer posted more than 20 years of accomplishment while developing and managing business at Robelle. He then took about eight years away from the HP 3000 marketplace, working elsewhere in the IT industry and sailing the Med with his family, so he's returned with a fresh outlook. We asked him in a Q&A interview what he's seen so far that's changed in your community.
Have you had any conversations yet with customers worried about HP's departure at the end of this year?I wouldn't say I've heard any angst or concern about that. I've been asking people what they're doing for support, since HP's coming off at the end of this year. My gut says the majority have already moved on [to independent support]. Others are looking, but they're not particularly concerned. It's kind of business as usual. I don't see it as any driving factor, at least so far, that would make people leave the platform.
How are the changes in this market prompting you to change the MB Foster message?
The main things I'm trying to get out is reminding people of all the things MB Foster does. That's a pretty broad swath; a lot of people may not be aware that we can enable data replication for a data mart: or the depth of experience the company has in services with many organizations, working with senior management to produce effective reporting to drive their businesses.
We're also reminding people that we're here whether you're homesteading or you want to transition. We've got solutions for both. We're trying to help people leverage more out of what they have.The approach to a 3000 customer seems to have changed, with less focus on engaging through a product or meetings with technologists. How has your approach changed?
It's much more services-oriented, tailored. I don't think there's as much of a cookie-cutter solution. What we bring to the table in our intellectual property and experience lets us leverage a lot of things in a short amount of time.
A lot of people have been on the 3000 a long time. Lots of them started out as programmers, and they're now CIOs. For them, a more senior business message is what they want to hear. In other cases we're talking with a technologist, and we have to raise awareness within that organization, especially with people responsible for lines of business about what's possible.You've been away from this market since HP made its exit plans. What reasons have you seen why people need to leave the platform?
There's an aging workforce that has familiarity with the 3000. There's also an aging workforce that has familiarity with the applications. One of the reasons we see as a key driver to migrate is a company loses a key technical resource. Or they lose a key end user resource, someone who really knows the application inside out. We've had a couple of cases where we had to do a rush assessment, “because at the end of the month Joe is leaving.”
We also see some pent-up merger and acquisition activity as a key change event. You suddenly get new senior management, a team that has different ideas and wants to take IT in a different direction. They don't necessarily want to support the legacy platform.
June 01, 2010
HP reveals 9,000 enterprise service layoffs
Tucked inside of an SEC 8-K announcement that Hewlett-Packard will invest $1 billion in enterprise services lay a bitter nougat: 9,000 employee positions will be losing their jobs over the next three years. The cuts are coming from the same services business that HP says it's expanding. It's not clear if these jobs are part of the 24,600 layoffs HP announced last spring.
But it's a certainty that HP took on more than 140,000 consultants and engineers in 2008 when it acquired EDS, swelling the Hewlett-Packard headcount to more than 300,000 employees. The new business unit has supplied the biggest share of HP's 2009-10 profits. Enterprise computing expansion has been a series of acquisitions for years at HP, from software purchases like Mercury Interactive to the EDS deal. HP's services business overtook HP Printer and Imaging's sales and profit leadership once EDS's 140,000 employees were folded into the company.
HP reported that over the next three years it will "replace approximately 6,000 of the 9,000 positions" it will cut. The headcount move gives the vendor flexibility to curtail compensation while it swaps out personnel based on performance and relationship factors. In an HP which has boasted of cost-cutting ever since CEO Mark Hurd took over, layoffs have been a steady fact of life. Hurd and other executive officers mentioned no layoffs in the quarterly conference call with analysts two weeks ago. Two kinds of operations have come under the HP scalpel: those with high headcounts and those not pulling in expected profits.
The company operates units more profitable (Software) as well as larger in sales (Personal Systems) than Services, but none with a higher number of employees. HP said the layoffs will net "annualized net savings, after reinvestments in sales resources and other initiatives, of approximately $500 million to $700 million by the end of HP’s 2013 fiscal year."
Some of those reinvestments come in the form of getting HP customers to adopt "modernized infrastructure platforms" for applications. Mention platforms in the 3000 community and the customers will think of servers and operating environments. But these changes are most likely to impact the customer who is considering the use of HP's datacenters for cloud computing. HP has been following an industry trend to promote cloud services as a migration plan.
HP said in today's 10-K report that the reinvestments come in the form of
- Fully automated, standardized, state-of-the-art commercial data centers
- Investment to facilitate the migration of client applications to modernized infrastructure platforms, and
- Consolidating the enterprise services business’s commercial data centers, management platforms, networks, tools and applications.
The changes in headcount will cost HP a $1 billion charge to its earnings over the next three-plus fiscal years -- which amounts to about one month of profits for the $114 billion company. The replacement staff of 6,000 will arrive in a vague bundle of positions.
6,000 of these positions will increase global sales and delivery resources. The changes to HP’s workforce will be made over time and will vary by country, based on local legal requirements and consultations with employee works councils and other employee representatives, as appropriate.
The last language in that statement shows that HP will be laying off enterprise services employees worldwide, since the US operations involve no works councils. The layoffs will affect HP employees in services delivery services, the technical knowledge arm of its EDS business. HP has maintained the EDS brand since the acquisition.
Hewlett-Packard spent $1.2 billion acquiring Palm and the WebOS in April, a deal that will close sometime next month. While these transactions appeal to unrelated customer sectors, it would be easy to say that the approximately $600 billion in HP savings from layoffs will pay for about half of acquiring Palm. Employees who are dismissed are likely to look at it that way. HP will be eating its own dog food, as the use-your-own-solution saying goes, to make bigger and better datacenters replace headcount in Services.