HP claims that Oracle is abusing its software strength to freeze it out of the hardware market, according to Reuters. The wire service said the disclosure came in a hearing Tuesday concerning HP’s suit against Oracle for pulling its support.

Oracle maintains that the Itanium is at end-of-life. HP claims Oracle is trying to migrate Itanium users to its Sun Sparc machines. Oracle says HP is perpetrating a fraud on the marketplace by artificially propping up Itanium.

The court is reportedly pressing HP and Oracle to settle now that Meg Whitman is running HP, not ex-SAP exec and Oracle enemy Leo Apotheker. Apparently there’s little chance of a settlement.

Meanwhile, Oracle claimed in a court filing that it’s confirmed from discovery gotten from HP that what it’s maintained for months is true – that there’s a “top-secret” pact between Intel and HP to keep the Itanium chip alive for two more generations and that HP is paying Intel to keep the part going despite Intel’s better judgment.

If Intel had its way, Oracle claims, it would kill the thing off for lack of sales. Instead Intel is forced to pretend that the “dead microprocessor is still alive” and “has a future” so HP can keep cleaning up on HP-UX service and support fees, fees it would lose if Itanium customers went to Linux-based x86 boxes.

“HP understands,” Oracle says, “that the future prospects of IT products drive customer purchasing decisions. A buyer who knew that Intel saw no future for Itanium, and was only continuing to invest in the line pursuant to a contractual obligation would devalue the future prospects of Itanium servers and be less inclined to buy.”

Unhappily the juicier bits of what Oracle told a court are redacted.

HP is suing Oracle for refusing in March to keep porting its software to the chip. HP claims it’s because Oracle bought Sun and Sun’s Sparc machines compete against its high-end Itanium systems. It says Oracle made its decision to force customers onto Sparc. Meantime it’s paralyzing the marketplace with uncertainty.

Oracle claims it’s the victim.

HP and Intel co-developed the processor back in the 90s on the supposition that the 64-bit part would go mainstream. Instead HP is practically the only company that uses it and it has billions riding on it although sales are down 23%.

HP claims Oracle is contractually obligated to support the chip. Oracle says it’s not.

Oracle wants the February 27 trial delayed claiming it hasn’t been able to get all the discovery it’s demanded and depositions haven’t even been scheduled. It wants documents from Intel and they’re proving hard to get, it said.

Both HP and Intel have denied Oracle’s contentions. Intel CEO Paul Otellini has said the Itanium roadmap goes beyond more than two generations.

The company said Thursday afternoon right after Wall Street closed that it would keep the business, expecting it to “deliver greater customer and shareholder value.”

The new decision has the “full support of the board,” a statement that only gives HP’s much vilified board another black eye.

Ostensibly HP did an “objective” evaluation of strategic alternatives for its Personal Systems Group (PSG), an exercise that reportedly involved about 100 “subject matter experts across the businesses and functions,” 18 teams that supposedly looked at all the angles of spinning the unit off.

Under the circumstances one might suspect that the findings were pre-ordained but the investigation reportedly calculated that a spin-out would run about $1.5 billion in start-up costs.

In a statement Apotheker’s replacement Meg Whitman – who, remember, signed off on the August plan – said, “It’s clear after our analysis that keeping PSG within HP is right for customers and partners, right for shareholders, and right for employees. HP is committed to PSG, and together we are stronger.” A schoolgirl could have figured that out.

Sources have whispered that the only reason HP ever thought about dumping PSG was to cover the $10+ billion it cost to buy the equally unpopular and expensive British ISV Autonomy. There was no one to buy PSG for more than $10 billion and that led HP to fall back on the cover story that it was thinking of spinning the unit out as a standalone company despite the sacrifice of the synergies it gets from the thing.

It is now reckoned that the loss of purchasing power and joint brand opportunities would have cost another billion dollars.

The Wall Street Journal said the total cost of a spin-out was originally only supposed to be $300 million-$400 million. Where that figure dropped from is unclear.

In a press release ahead of a conference call HP said, “The data-driven evaluation revealed the depth of the integration that has occurred across key operations such as supply chain, IT and procurement. It also detailed the significant extent to which PSG contributes to HP’s solutions portfolio and overall brand value. Finally, it also showed that the cost to recreate these in a standalone company outweighed any benefits of separation.

“The outcome of this exercise reaffirms HP’s model and the value for its customers and shareholders. PSG is a key component of HP’s strategy to deliver higher value, lasting relationships with consumers, small and medium-sized businesses and enterprise customers. The HP board of directors is confident that PSG can drive profitable growth as part of the larger entity and accelerate solutions from other parts of HP’s business.”

Todd Bradley, who was on the conference call, will apparently continue to run PSG. Although PSG miraculously lost no market share in Q3, he still has to deal with the August 18 hangover. In answer to a question, Meg said she was confident HP would get its “fair share or more” of disk drives out of poor waterlogged Thailand because of its size and long-term relationships.

Whitman is giving the company another couple of months to figure out what to do with webOS. Apparently it’s seen as some cross-organization something or another but unlikely to return inside a tablet. HP wants Windows 8 for tablets.

HP made back a pittance of its squandered wealth Thursday. Its stock rose to $26.99, up $1.24 or 4.82%, on a day when the market surged. Before the August 18 announcement HP was at $31, down more than a third from its glory days under Mark Hurd.

Meanwhile, Meg – or her people – are trying to reduce the “what is HP” question to a bumper sticker. She said it had to be right because the company will have to execute against it. Before that she’s got to make Q4 guidance. The numbers are due out next month. By then she needs guidance for 2012. She said, “HP tries to do a lot of things.” She, on the other hand, is “a big believer in doing a small set of things really, really well.”

The paper heard that “recent analyses by Hewlett-Packard and its top advisers indicate that the costs of spinning off HP’s personal-computer business might outweigh the benefits,” if for no other reason than “separating the PC division would significantly diminish HP’s buying power with component makers, complicate its supply chain and reduce product margins on some products.”

We heard – well, everybody’s got leaky HP sources – that HP was “surprised at the synergies,” like nobody’s been watching the store.

The company is doubtless spending a boatload coming to a conclusion that people who can’t balance their checkbooks could do intuitively.

HP’s new CEO Meg Whitman reportedly hasn’t come to a decision yet. We distinctly remember saying weeks ago that HP would pull in the decision – Whitman then said she wants it made by the end of the month – and would decide to keep the thing.

It’s suspected the decision has already been made and the delay is meant to give the HP board some measure of deniability for sanctioning the idea however implausible. The reason HP backed off its first impulse, which was to sell the thing, was the realization it couldn’t get more than $10 billion for it.

In an unexpected move, Ray Lane, who put Whitman on the HP board in January, was named the company’s executive chairman. As non-executive chairman, Lane was very much the power behind Apotheker’s throne. Now apparently he will be Meg’s Gray Eminence. He said he was there “to support Meg,” same thing he said about Apotheker.

The two are close because Lane helped Whitman out when she ran into problems at eBay according to what some Yale guy said on CNBC. At least they know each other as customer and vendor from the days when she was at eBay and he was at Oracle.

The change in Lane’s status, which is likely to be remarked on at some point by Lane’s old boss, Oracle CEO Larry Ellison, more closely resembles the speculation last week that Lane would get the CEO job when Léo fell and in fact Lane admitted during the conference call that followed the announcement that he had considered stepping into the slot on an interim basis.

Lane said he had found “weaknesses in parts of the business” and that Léo was let go because of poor execution, failed leadership, lack of understanding HP’s various businesses except on a cursory level, and poor communications skills, which culminated in the disastrous August 18 announcement that HP would spin out or sell off its PC unit, abort its nascent drive into webOS-based tablets and smartphones, and buy Autonomy for a king’s ransom.

Since the board blessed the plans, it’s not backing off them, which is why HP’s stock was down again after-hours to $22.60 after Wall Street plunged a very nasty 4% during the session.

Investors don’t want to hear that the board is indecisive about what to do with the company’s giant $41 billion PC arm. Delay is losing business and HP CFO Cathie Lesjak got on the horn long enough to say revenues this quarter would be worse than the $32.1 billion-$32.5 billion Léo predicted in August when estimates were for $34 billion but by some financial gymnastics HP will still make his predicted ESP of $1.12 to $1.16 compared to estimates of $1.31.

Although Lane has now decided that HP is a hardware company – something else he said Léo didn’t realize – the best he and Meg could do was promise a decision on the PC unit by the end of the calendar year. Betcha they decide to keep it. Betcha they say so in a few weeks.

In another slap at Apotheker, Lane said he had banished the word “transform” from the HP lexicon. “We will have more services and software but we’re a $120 billion hardware company.”

Insiders claim the only reason HP is sticking by the fatal August 18 strategy is that Lane was part of the troika that included Léo and chief strategy and technology officer Shane Robison that reportedly persuaded the board to rubberstamp the plan.

Meanwhile, it’s still quite possible that dissonant shareholders who have lost half their money since HP tossed out Mark Hurd may make a bid to dump the whole board and substitute their own slate. That slate may then make some management changes. They’ll have to live with the Autonomy acquisition. HP can’t get out of it and it should close by the end of the calendar year.

Whitman pledged to turn things around and mend fences with the investment community, but said it would take time to rebuild confidence. She has met with HP’s Executive Council to discuss how they will work together and boost employee morale. “The organization’s been through a lot,” she said.

There are of course widespread doubts about Meg’s skills and her purely consumer background, her inability to grow eBay passed the $7 billion mark, and her pricey acquisition of Skype for $2.5 billion which ultimately forced eBay into a $1.4 billion write-off. Skype now belongs to Microsoft.

Lane bushed aside the issue of Whitman’s hasty appointment and said he was acquainted with all the candidates and there’s none better. Possible insiders aren’t ready and Meg apparently presented him with a plan for the next six month and the next year.

In answer to questions about the board itself he said it’s not the pretexting board or the board that fired Hurd or the board that hired Apotheker. He saw to that in January by bringing in five new people and getting rid of four troublemakers. He never dealt with the fact that it is the board that sanctioned the August 18 plan and okayed the Autonomy acquisition.

Bloomberg claimed that Apotheker was blithely unaware that he was about to lose his job until Wednesday when the news broke that he was about to be ousted. The notion that he hadn’t a clue seems a bit farfetched given that he was gagged last week and replaced by Lane at two outside meetings where he was expected to discuss HP’s strategic direction, but the men did same to have a tin ear.

Having left its party manners at home it told everybody who would listen that Cisco is the master of complexity, high prices and that dirtiest word of all, lock-in.

“Single-vendor, proprietary approaches such as Cisco’s lock in customers while driving up cost and complexity with different architectures required at each point in the network, including data center, campus and branch. This lack of convergence and increased complexity make it difficult to roll out new applications and services.”

Naturally it’s got an alternative – a just-announced FlexNetwork architecture that it expects to eat Cisco’s lunch.

It’s an ideal time for HP to move while Cisco, its market leadership under pressure and its investor base fleeing, is distracted by restructuring and cost cutting.

Anyway, HP’s single architecture approach is supposed to unify network silos across the data center, campus and branch bringing consistency and performance to what is otherwise “disjointed” and making it easier for users to leverage chi-chi technologies like cloud computing, virtualization, mobility and media-rich content.

It’s supposed to do wonders future-proofing legacy networks, where it claims there’s been little innovation in the last 10 years.

A core component of HP’s vaunted Converged Infrastructure, the widgetry is based on modular building blocks – so it can be incrementally introduced – and open industry standards like OSPF v2 and v3, HP preens, as opposed to Cisco’s stash of patented gotchas like its EIGRP protocol.

FlexNetwork consists of:

* FlexFabric, which converges network, compute and storage across virtual and physical environments in the name of hybrid cloud computing;
* FlexCampus, which converges wired and wireless networks to deliver identity-based access to multimedia content;
* FlexBranch, which converges network functionality, security and services at the branch;
* And riding herd on the others FlexManagement, which converges network infrastructure management and orchestration.

HP toted up the number and says that by itself its single-pane-of-glass management tool does what it takes Cisco 30 different tools to do. Its latest Intelligent Management Center (IMC) 5 widgetry manages HP’s entire networking portfolio as well as more than 2,600 network devices from over 35 vendors, more than 1,000 of which are Cisco’s.

It claims it manages Cisco better than Cisco manages itself.

The stuff automatically discovers virtual machines and virtual switches and their relationship to the physical network. HP means to add automatic synchronization of network connectivity information with its Virtual Connect technology for server blades and automate the process of creating a server profile further, moving a step closer to one-button cloud provisioning.

And since the exercise is all about dumping on Cisco, HP trotted out a new A-series 10500 campus core switch for delivering media-rich applications – HP paints campus architectures as the forgotten man in other people’s schema – that are supposed to offer 75% lower latency, 250% more switching capacity and 270% more 10GbE density than Cisco’s Catalyst 6509.

The thing can be virtualized into a super core with 208 wire-speed 10GbE ports for large campuses.

Mind you the campus switch segment is supposed to be the largest segment of the worldwide Ethernet switch market, forecast to reach $13.7 billion this year up from $12 billion last year.

For the access layer of the FlexCampus HP’s got new line cards for HP E5400 and E8200 switches that are supposed deliver up to 90% lower latency, 600% higher throughput, 128% more port density and 35% less energy consumption than the Cisco’s Catalyst 4506.

On the security side – where legacy solutions sometimes only protect physical environments supporting single workloads – HP’s got a new Tipping Point S6100N Intrusion Protection appliance that propagates standard security policies across a data center, campus, branch and WAN and is supposed to automatically build in protection as virtual machines are created or moved across an enterprise.

It’s a single solution for physical, virtual and cloud environments that reportedly offers 60% higher performance than the previous generation and can inspect up to 16 Gbps of high-bandwidth application traffic in real-time to improve the availability of mission-critical services.

It’s supposed to have 100% greater performance than Cisco’s 4270 at 33% less power and reportedly discovered 309 vulnerabilities to Cisco’s one. (Ouch!)

The appliance is available now for $209,995.The A10500 switch should be available in the second half at prices starting at $38,000. IMC 5.0 is due next month for $6,995; IMC 5.1 should be available by the end of the year.

By the way, HP’s waving around a study that says that 75% of 90 Cisco resellers that account for $5 billion in revenue acknowledge that HP is part of the sales discussion and that 29% of them find HP influencing the terms of the deals.

He’s the first member of HP’s top echelon to bolt since the new administration took over in November.

Léo’s answer to this apparent dilemma was to name Jan Zadak, a Swiss who’s been managing director of HP EMEA, to replace Hogan effective May 1. Hogan’s last day will be May 31 “to ensure a smooth transition” but that doesn’t mean he’s showing up at the office.

Zadak may have something of a steep learning curve since his heritage is Compaq PCs and he doesn’t know the American market, where HP does more direct business than it does in EMEA. EMEA taps more of the channel for the cheaper selling cost.

Gossip says that Hogan, an ex-IBMer, was frustrated with Apotheker’s inability to understand that HP is primarily a people-intensive hardware company with a supply chain, not a software company.

Hogan was supposed to be “transforming” HP’s enterprise sales to make it easier to do business with a company that has a bunch of units pitching accounts.

Reportedly he doesn’t have another job, but it shouldn’t take that long.

Hogan joined HP in 2006 after Mark Hurd got there and according to the company “drove a significant extension of the company’s software portfolio, tripling revenues and expanding operating results.” Between IBM and HP, Hogan was CEO of Vignette.

Apotheker’s “evolutionary” strategy so-called was encased in a thicket of buzzwords under the banner “Everyone On,” heralding HP as the platform for connectivity and the cloud. Details were pretty thin on the ground and although he said, “It’s the first time HP is trying to put all of the elements of what it’s doing together,” the rhetoric seemed to be missing the cohesive glue as well as the how it happens. Makes one recall Lou Gerstner refusing to saddle IBM with a vision.

Anyway, although late to the party – and picking a fight with any number of folk – HP is going to go into the IaaS and PaaS public, private and hybrid cloud business with its own cloud platform – based on technology and tools whose exact origin is unclear – reportedly overriding internal suggestions that it use a budding open source platform like OpenStack.

Aside from offering storage as a service and compute as a service, the tad obvious scheme, whose timing is this year or next, will include some kind of App Market selling software HP builds or acquires or allows in that’s supposed to serve consumers, SMBs and the enterprise.

Naturally the new cloud stack will be integrated with HP’s servers, storage and networking – presumably making its hardware less desirable as the basis of other people’s clouds, which is where HP has been selling. Apotheker is depending on gentlemanly “coopetition” to pull that particular piece of fat from the fire. “We will continue to be a good partner to all our existing partners.” Like Cisco?

Oh, yeah, there’s a personal cloud in the mix somewhere too but it wasn’t teased out.

HP believes the cloud opportunity will be about $143 billion by 2013.

The connectivity is supposed to be supplied by HP’s Palm-derived webOS, which it means to install on all its PCs on top of Windows, on its printers, on sensors and set-top boxes, and on what few mobile devices it has. It also appears fated to be installed on HP servers although that factoid was more implicit than explicit.

The company figures this outrageously scaled stateless phone operating system will be on a 100 million widgets a year in no time. First, you gotta wonder whether this is gonna work and then you have to consider the lock-in if it does.

Although Léo thinks it won’t, installing webOS everywhere should tick off Microsoft, if not now eventually, but for the time being HP’s still missing the crucial smartphone and tablets needed for any ecosystem to appeal to developers. Still, it’s looking for 30,000 apps by the end of the year.

Apotheker has evidently taken a big bite of the “consumerization of the enterprise” brownie, and figures that’s the way innovation works these days. He’s counting on business buying his webOS TouchPad, which is rumored to start at $499, same as the iPad 2, and should appear on the market in June.

He also means to compete with IBM on analytics. HP is in the process of buying Vertica; IBM has bought practically everything else. According to Apotheker “Analytics is a huge space that is poorly served today.” It’s doubtful IBM is losing any sleep over HP’s intentions to deliver analytics in a year or so on-premise, in appliances and as an SaaS service. Among other things, Vertica will be made into a Big Data appliance.

IBM probably had a good chuckle over Apotheker’s claim that “We are not playing catch-up to anyone, particularly IBM” in analytics.

Disappointing a lot of the Wall Street consolidation players, Apotheker said HP will avoid big transactional applications and won’t buy any legacy software. It’s only interested in stuff that’s distributed and cloud-ified and will be “disciplined” albeit “aggressive” in making acquisitions (no more overpriced 3PARs?). Presumably it’ll be in the market for analytics. Apotheker also indicated he’s interested in security and management.

One might expect the relationship between HP and SAP, where Apotheker was briefly CEO, to tighten up (think ERP and the App Market) but no wedding bells are likely.

Lip service was paid to R&D (more spending) and services (more salesmen).

After downgrading HP’s Q2 and full-year guidance last quarter, Apotheker said this time through that the company could do $7 a share by fiscal 2014, up from $4.58 last year. It’s probably doable with some margin expansion but that doesn’t seem to cover the cost of Léo’s vision.

Goldman Sachs observed that “To shift its portfolio towards higher-growth and higher-margin segments, HP will have to take share from category leaders such as Apple, IBM, Oracle and Cisco. This will take opex investment, and gross margins are likely to be volatile as well. As such, HP’s expectation for persistent operating margin expansion and growth appears overly optimistic.”

He’s also got to pay for the dividend increases, his anti-Mark Hurd no-cost-cutting policy and the raises he restored when he arrived in an attempt to be popular with the staff.

It says it will deliver “private cloud as a service” from its state-of-the-art data centers.

The company expects to start in North America and EMEA in February and expand to Asia-Pacific and Japan mid-year, according to Sandeep Johri, president, corporate accounts & industries. Apparently it’s already been practicing with France Telecom and some of its non-French customers.

It’s got what it calls, in its flat, inimitably pedestrian way, HP Enterprise Cloud Services-Compute, a hybrid private cloud that bundles SLA-based pay-as-you-go server, storage, network and security resources for running core applications and processes. It promises enterprise- and government-grade security, availability and ease of integration.

CloudSystem, based on HP’s popular blade servers and its notion of a “Converged Infrastructure,” uses HP’s cloud automation and management software for delivering unified security, governance and compliance across applications, as well as physical and virtual infrastructure.

It’s advertised as “the most complete, integrated system to build, manage and consume services across private, public and hybrid cloud environments.”

CloudSystem is supposed to enable new cloud services to be up and running in minutes and supports HP Cloud Maps, which provide preconfigured catalog objects to automatically provision optimized application and infrastructure resources. It’s got Maps for Citrix, Microsoft, Oracle, Red Hat, SAP, SAS, Symantec, Tibco and VMware.

HP sees the cloud as part of its two-month-old notion of “Instant-On Enterprises,” where everything and everyone is connected and users can quickly adjust to changing demand.

It figures the cloud is going mainstream and claims to have “the enterprise experience, breadth of portfolio and global service delivery organization to lead our clients through this transformation.”

Naturally it will sell services into the endeavor and offer financing for clients building private cloud deployments who want to minimize capital expenditures.

HP never got back to us about the rumor that initially it would have only storage.

The move finally came Thursday after the stock market closed only to be stepped on by Google’s news that it was moving one of its young founders into the CEO’s chair.

According to HP’s non-executive chairman Ray Lane four directors when asked volunteered to withdraw: Joel Hyatt, John Joyce, Robert Ryan (the board’s former lead independent director) and Lucille Salhany – that looks like at least two maybe three vocal Hurd supporters and an anti-Hurd agitator – people who, by the way, also decided to bring in Léo Apotheker as Hurd’s replacement, an unpopular, unproven dark horse choice to say the least.

In their place HP – apparently Lane who appears to be running the board – recruited five replacements: former eBay CEO and last season’s failed gubernatorial candidate Meg Whitman; former Alcatel-Lucent CEO Pat Russo; former GE CIO Gary Reiner, now also a special advisor to private equity house General Atlantic; Booz & Company CEO Shumeet Banerji; and AXA Private Equity founder and CEO Dominique Senequier.

That brings the board to 13 effective January 21 ahead of election – and heavy on VCs – and their job Lane told CNBC is to support Apotheker (which sounds like the kind of yes-man thinking that gets companies into trouble). Banerji and Senequier, he said, bring international experience; Reiner can represent the enterprise customer; Whitman supposedly knows the consumer like eBay is the consurmer, and Russo networking.

So much for rearranging the furniture at that end of the building.

Earlier this week the Wall Street Journal said Apotheker has been working on a plan with the HP board – which we imagine really means Lane, the one-time president of Oracle and still a managing partner at Kleiner Perkins – with maybe Marc Andreessen kibitzing – that would see HP put more emphasis on software, networking and storage (no surprise there), focus on selling into the cloud (isn’t everybody), increase HP’s R&D spend (his predecessor was criticized for cutting back), downplay PCs and servers, kick three-time-CEO-hopeful Ann Livermore, who runs HP’s $57 billion enterprise business, upstairs as vice-chairman (apparently it was wrong there), and carve her unit in two giving equipment to EMC import David Donatelli and services to Enterprise sales chief Tom Hogan both reporting to Apotheker.

The paper said the plan could change and HP denied it was going to reshuffle senior staff.

What we hear is that senior management hasn’t a clue what going on.

The next senior executive to depart, the Journal predicts, will be CIO Randy Mott.

The company has – and this is confirmed – quietly hired Emil Sayegh out of Rackspace as marketing VP, cloud computing. He was general manager of Rackspace’s Cloud Computing Division.

The plan, whatever it is, is supposed to be made public in March. Reportedly Apotheker is importing a lot of SAP people.

Oracle needs to find market share somewhere because both Sun, and Unix in general, are losing out badly according to the latest server figures from Gartner.

So Oracle’s completely refreshed it Sparc server line – which everybody wondered whether it would actually do – and, as a result, based on a new TPC-C benchmark, it’s comparing its OLTP performance to a cheetah, IBM’s to a stallion and HP’s to a turtle.

Oracle was reportedly able to achieve a record 30 million transactions a minute on a newfangled soup-up Sparc Supercluster made up of 108 T3 Sparc processors with 1,728 cores, 13TB of main memory, 1.7PT of storage, 246TB of Flash memory and a 40 gigabit network running a standard Oracle database with a quadrillion rows.

The best HP has been able to muster is a real estate- and energy-hogging Superdome that scored four million transactions a minute a few months ago. IBM, which Larry allowed – my, my – has “good products,” can do 10 million with a special non-standard clustered version of DB2 running on a Power 7 machine, Ellison said. Oracle used a plain vanilla Oracle RAC database to set the new world’s record.

Of course this is all benchmark craft. Oracle doesn’t expect anybody to actually buy the benchmark machine that’s supposed to be capable of 43 trillion transactions a day but next year Oracle will be peddling three dumbed-down general-purpose commercial versions of these new Sparc Superclusters based on Sparc T3 and M5000 servers under a new so-called “Sun Rises” program.

The Superclusters, whose price is unclear, is a complete infrastructure solution for running Oracle RAC database environments. Besides servers, they include software like ZFS, InfiniBand networking, FlashFire storage and new Gold-level support.

Ellison claimed they have no single point of failure and are completely fault-tolerant. T3 chips can have up to 16 cores. T4 chips are reportedly running in Oracle’s labs.

Oracle also introduced a new Solaris 11 Express-running blades-based Exalogic Elastic Cloud box as an alternative to the x86 Exologic Elastic Cloud machine it wheeled a few weeks ago.

The widget, to be delivered in Q1 running with Solaris or Oracle Linux, is supposed to be good at running middleware like Oracle WebLogic Server and is advertised as the “fastest Java machine in the world.” (Larry quickly pointed out that all Oracle Fusions apps are pure Java.) It also runs non-Java apps and is supposed to be optimized for multi-threaded programs.

Ellison said it’s “not a born-again cloud” whatever that means.

It offers on-demand capacity and multi-tenancy, scales and when the job is done returns the resources to the pool. It’s targeted at enterprise-wide data center consolidation and consists of a rack of 1U machines with dual six-core processors, solid-state drives and an InfiniBand I/O fabric, as might be expected since Oracle is partial to InfiniBand.

Oracle also announced a new high-end Sparc Enterprise M-Server server line bearing a new 3GHz Sparc64 VII+ processor from Fujitsu with a maximum 12MB of L2 cache, double what’s been available, offering a reported 20% performance increase. The boxes, up to the 64-socket M9000, are jointly designed, manufactured and branded.