Oracle Harm to 3000s & HP, Past and Present
November 21, 2011
Oracle has become a loud competitor to HP in the migration alternative game. HP took an hour's webcast this month to assert that the Oracle-Sun-Solaris action is mostly heat without much light to lead the way. But the database has certainly gained HP's attention now that ousted HP CEO Mark Hurd is leading the Sun attack. If some stories are to be believed, however, the current fracas is a long way removed from Oracle's blows against the HP 3000's futures.
At one point HP was eager to keep its 3000 customers buying Hewlett-Packard enterprise solutions. And by 2001, the story goes, Oracle wanted a clear path to selling Oracle hosted on HP's Unix servers. HP was going to cut something out of its merged product line with Compaq. The 9000s were never on the block, but there was the HP 3000, sporting tens of thousands of systems, nearly all of them running an IMAGE SQL database.
So here goes: HP didn't kill the HP 3000, Oracle did. Oracle made HP a deal they couldn't refuse. Stop selling competing database software, and Oracle would partner to sell HP-UX systems as Oracle servers. Since MPE/iX is tightly coupled to IMAGE/SQL, this translated to the end of the HP 3000. The smoking gun was supposedly this: As part of HP's announcement on discontinuing the HP 3000, it included the end-of-life for Allbase/SQL on both MPE and HP-UX.
The theory has some credibility. In 2001 there was a lot of growth potential for Oracle-plus-HP-UX. Oracle had grown up plenty in the decade before that fateful date (when its CEO Larry Ellison, left, was a tender 47 years old.) In 1991 his juggernaut was pretty much out of the game, even with 20,000 customers, because it was scraping the bottom of its cash barrel. Some reports said Oracle creditors were ready to call in their 1991 notes. Those 20,000 sites arrived in the fold because Oracle got ruthless about sales and promises and vaporware delivery. $80 million in cash from Nippon Steel in exchange for a piece of the Oracle Japan arm helped Oracle back from the brink, too. Then there was a forced restatement of sales all the way back to the company's first quarter as a public company. Booking nonexistent products and stretching future contracts as current revenues, it was all the kind of behavior old-school sales reps would chuckle about today -- a day when Oracle wants to rock the 2011 boat, charging that HP is paying Intel to prop up the Itanium product line.
Over at InformationWeek, the "hungry to profits" company was struggling to deliver Oracle 7 (the latest version is 11, with 12 in the wings). A list of that year's Oracle Business Alliance Program members didn't include HP, even though by then Hewlett-Packard's strategy was in total thrall with Unix. Oracle still was the leader in Unix databases in that era, selling just a bit more than half of Unix RDBMS installations. Sybase was jousting at Oracle with a preannoucement of Version 10 of its database, which one newsletter said was "an announcement, if it's possible, with even more hot air than Oracle's Version 7."
Meanwhile, there was no hot air in the IMAGE of that period. HP was being led by the nose by its customers to keep IMAGE an integrated part of the 3000 solution. Hewlett-Packard wanted to separate IMAGE from the 3000 and make the database an add-on. A revolt at the Interex user group conference of 1990 ensued. Even though the users carried the day -- and Jim Sartain became a business-savvy IMAGE R&D manager afterward -- the IMAGE standard bearers look see this as the start of the decline of HP's attention to IMAGE, business-wise.
Oracle couldn't be bothered with current 3000 R&D from that point onward. Jennie Hou was yoked to the HP's thankless task of getting Oracle resources devoted to MPE/iX -- because HP hoped Oracle would attract customers in the 3000 arena. Oracle just didn't eager to attract any off of the 3000 platform, maybe because developing for MPE wasn't the Unix business Oracle had used to get off the mat. Or maybe Oracle saw the folly -- which it first told me in 1985 -- of competing with a bundled solution like IMAGE. While the rest of the industry was deploying Oracle 8, that release was out of reach for the 3000 user. We wrote in 1997:
It only took one Oracle sales rep in California to get the 3000 customer base worried about the future of the database on MPE/iX. One rep's comment to a 3000 customer circulated through the Internet, asserting that Oracle was only going to support its database through version 7.2.3 for the HP 3000. This led to a fair bit of piling on, as people wondered what the purported pullout meant for the HP 3000 and why anybody would want to get serious about using Oracle instead of IMAGE anyway.
HP and Oracle went to work on damage control almost immediately. The two allies whipped up a quick update on their plans for the HP 3000. The sale's rep's comments were based on a partial truth: Oracle is still not willing to commit to a list of supported platforms for Release 8. Despite what some might see in the tea leaves of whether the 3000 is mentioned on Oracle's Web page roll call, no one using any platform knows for certain when they're getting an Oracle 8 -- not just yet.
Oracle 8 on the HP 3000? One year later, HP had committed its own engineers to just getting a fresher Oracle 7 onto the platform.
CSY has engineers working on the port of Oracle 7.3.4 according to Jennie Hou, the manager of the HP 3000-Oracle relationship. “The porting resources are still engaged in 7.3.4,” she said, which HP expects to be available to 3000 customers within calendar 1998. Version 7.3.2.31 is currently shipping for both MPE/iX 5.0 and 5.5 HP 3000s. There has been no announcement of an Oracle 8 port from CSY yet, “because customer needs are being met by Oracle 7,” Hou said. Oracle-based applications are ready for the HP 3000 in manufacturing, financials and human resources, and Oracle plays a part in data warehousing solutions for MPE/iX.
An Allbase pullout as a smoking gun would be hard to point at HP's 3000 history. The database had its fans among some 3000 sites, but it seemed to be more of a "other option" item on the HP pricelist. More than 95 percent of the 3000's sites were running IMAGE/SQL and still do. That's one reason that Eloquence has done so well as a database migration replacement for 3000 sites: it behaves just as IMAGE does with 3000 programs that are moved to HP-UX or Windows platforms.
3000 users watched a lot of one-sided pursuit of Oracle affections during that decade leading to the pullout. By the fall of '98 it was obvious HP 3000 customers didn't want to fork over tens to hundreds of thousands of dollars for a database instead of IMAGE. Winston Prather announced the end of Oracle futures on MPE/iX at the HP World conference -- one year before taking over as 3000 GM which led to, well, the end of HP's futures altogether for the 3000.
“The current plan is that there is no plan to port Oracle8,” Prather said at the 1998 forums. “The real fundamental issue: Oracle’s position is that, ‘We’ve worked really hard with you to bring it to the platform and done joint marketing programs, and it’s not working.’ Oracle is looking at it and going, ‘Why am I doing this?’ They haven’t come close to recouping their investment. Their current feeling right now is that there’s not enough business on the 3000.”
Customers using Oracle say the database runs faster on HP 3000s than on HP 9000s, something that HP might want to use in making a case for additional development resources. Prather noted that customer adoption has been slight in the face of a more cost-effective 3000 database: IMAGE/SQL.
“It’s very expensive to implement an Oracle solution relative to a TurboIMAGE solution or even an Allbase solution,” Prather said. “The bottom line is that 3000 customers like TurboIMAGE. They don’t want to leave TurboIMAGE. Our application providers don’t want to move to Oracle, because they like TurboIMAGE."
It's not like HP wasn't trying to sell the database. In 1996 it offered Oracle's 7.2.3 version to 3000 sites at prices starting under $1,200 per seat with an eight-seat minimum. For the first time, you could get a small 3000 into Oracle for just under 10,000. The price per-seat increased based on HP's CPU tiers of the day -- that under-$1,200 price was only available for the lowest 3000 tier. Price was a barrier vs. an included database, and then there was migration.
The HP pricing doesn't alleviate the roadblocks of data migration and increased management demands. Taurus Software's Forklift migration tool, which uses a graphical interface to map IMAGE datasets to Oracle tables by way of the Taurus Warehouse utility, was the start of moving data as easily as any tool on other platforms. Forklift gave managers a visual aid to get IMAGE data into Oracle databases.
Within five years Oracle wouldn't have IMAGE to dodge around anymore while it sold HP's systems. Wht good did that do in the long run? By this month, Oracle doesn't even want those HP Unix systems to exist. Its charge of paying Intel to keep Itanium alive is pretty blustery with hot-air, even by Oracle's standards. HP doesn't make Itanium anymore, even though its engineers retain a role in processor design. Itanium is an Intel product by now, and if HP is chipping in extra dollars to keep development going, that's in HP's best interest to keep selling its Unix servers.
We find it interesting to see how Oracle has crept back from the "Itanium is dead, and Intel isn't saying" hot air of this spring. Intel came back with an opposing gust that might have knocked long-time yachtsman Larry Ellison off the "attack Intel" tack that's part of his warpath on HP. Oracle can't hurt the 3000 anymore, can it? That depends on how you think of the Integrity servers as 3000 migration replacements. This Oracle war is creating distractions for HP's Integrity sales.