by Tim Negris

Larry Ellison is one of the most misunderstood guys in Silicon Valley, and his extemporaneous pronouncements often only serve to make him more so. Sometimes his remarks are explosive and expository, like his now-famous Churchill Club cloud computing rant, and other times they are compact and cryptic, like his recent remarks about buying some chip companies. Either way, the pattern is familiar: he says something that prompts a lot of misguided chatter, and then he does something that only deepens the confusion.

After his Churchill Club diatribe, he was roundly, loudly, (and wrongly) tarred as a cloud-hating Luddite. Then, last week, when he showed off his new “cloud in a box” at OpenWorld, there was a lot of confounded mumbling damning the first as crafty equivocation or the latter as a cynical epiphany.

But, anybody who really listened to what he said and who understands what he did probably isn’t so confused. At the Churchill Club, he railed against marketing hype passing for technical innovation, and at OpenWorld, he touted tangible technology that is, in his view, the real deal.

Now, with this chip thing, it’s happening again. All he said was, “You’re going to see us buying chip companies. Silicon is very important.” But it was enough to get tongues wagging and the stock prices jumping for companies that are unlikely to be the ones he was talking about.

Chips is Chips – Not!
A lot of the chatter and (stock) speculation has focused on companies that make CPUs – AMD, ARM and Nvidia. But, to believe that Ellison would buy one of those you’d have to also believe that he’s interested in putting Oracle into a business other than enterprise technology, like supplying processor chips to makers of PCs, cell phones and embedded systems. Or, you’d have to believe that the Sparc CPU wasn’t a big part of why he bought Sun and that the new five-year Sparc road map is an elaborate prevarication.

To paraphrase Sigmund Freud, sometimes an enterprise technology company is just an enterprise technology company. And, in 30 years, Oracle has never tried to be anything else, so why start now? No good reason comes to mind.

Recognize that there are few Silicon Valley executives with a clearer, straighter, shorter path between thought, word and deed than Ellison. Now, think about his demonstration of the Exalogic Elastic Cloud at OpenWorld and recall what he has said this last year about vertical integration of hardware and software and about the folly of customers piecing together solutions from “best of breed” elements from different vendors. The answer to the question of which, or at least what kind of chip companies he may buy is as plain as the box on the stage.

The Exalogic Elastic Cloud comprises three main subsystems: processor, disk and interconnect, all of which, of course, incorporate an assortment of different kinds of chips. Some are standards-based commodity products with low prices and many sources, where ownership brings few rewards. Others, though, are powerful proprietary products that would bring Oracle both strong product differentiation and competitive leverage in relation to other server makers that must also use those chips.

Remember that right before his lob about buying chip companies he said, “Our focus is to build our [intellectual property] portfolio.” Forget standards and after-market boards. Larry’s cloud-in-a-box is a sealed cube and its hardware innards ain’t nobody’s business but his. So, to handicap which chip stocks should actually be moving on Ellison’s latest pronouncement, look to IP-rich components surrounding the processors, disks and interconnects.

Processors
The Exalogic box he demonstrated at the show contained 30 servers, each with two six-core Xeon CPUs. Now, before you say, “See, there’s where he’ll put the Opterons when he buys AMD,” consider instead the greater likelihood of the Xeons being replaced in the production machine with the just-announced 16-core Sparc T3, which delivers comparable performance, but runs cooler and eats less power than the x86 chip.

The interesting bit about the servers in the Elastic Cloud is that they use solid state disks (SSDs) to hold the operating system. Although SSDs are mostly made up of commodity memory chips, it takes some pretty special hardware mojo to endow them with fault tolerance and other attributes especially important in the utility (cloud) computing model.

SandForce is a four-year-old Valley start-up (sorry, traders) that just raised a $20 million D round on the strength of its patent-pending SSD controller technology. Currently SandForce packages its DuraClass controller with its own SSDs, but designed it to be easily incorporated by other SSD vendors, too. One of DuraClass’s key selling points is high and symmetrical read-write speeds, which would make it an excellent choice for OS storage in Ellison’s big box.

Disks
SandForce could also play a part in Oracle’s storage subsystem as a way-faster-than-disk element for storing frequently referenced data. But, we’re years away from all-SSD storage in large enterprise systems and it’s unclear how easily and to what advantage the SandDisk technology could be mixed with the conventional “rotating” disks that make up the bulk of the storage subsystem.

Enter Texas Memory Systems, a long-established, privately held, VC-free, engineer-run Houston-based company that is also in the SSD controller and drive business, sporting the trademark “World’s Fastest Storage,” claiming a number of large enterprise and government customers, and already an Oracle partner. What sets TMS apart from SandForce as a candidate for inclusion in the Oracle storage subsystem is its purchase a year ago of storage virtualization start-up Incipient, whose technology gives TMS the ability to transparently mix spinning disks with SSDs and enable easy migration from the former to the latter.

Interconnect
This is the part of the system that ties the processors to the disks and the processors to each other, and plays a big part in assuring the system’s overall speed, flexibility and reliability. There’s something of a battle brewing these days between two different interconnect technologies: Ethernet and InfiniBand, with many companies arrayed on either side. The Exalogic box uses InfiniBand and Ellison has made it clear he thinks it’s superior to Ethernet for his purposes.

It’s worth noting in passing that the increasingly popular VBlock cloud infrastructure technology being flogged by the Cisco/EMC/VMware VCE troika uses Ethernet and for that reason and others is likely to become one of Ellison’s favorite piñatas in future stagecraft.

One possible acquisition candidate in the InfiniBand interconnect area is a no-brainer and thus already mentioned in a number of places. Mellanox is the leading supplier of InfiniBand chips, cards and switches, one of their guys runs the InfiniBand Trade Association, they are a long-time Sun familiar, and most companies in the InfiniBand space are in some way beholden to them. They are a public company, but not gigantic, and not diversified into other businesses that Oracle can’t use.

But, Mellanox is not the only candidate in the InfiniBand interconnect area; there are others that could make good acquisition prospects, either as alternatives or in addition to Mellanox. Remember, InfiniBand is used to connect the processors to the storage and to each other. While Mellanox plays in both of those roles, there are other companies more focused on one or the other that might also fit into Oracle’s plans.

QLogic makes InfiniBand products for clustering processors, but has a longer, deeper history and considerable leadership in storage-related products. In that area, besides InfiniBand, it also offers Ethernet and Fibre Channel storage connection technology. Although Oracle wants to be the only box in the data center and wants all storage to be in that box, obviously that isn’t going to happen. Integrating with other storage products using all three protocols will likely be a requirement, and one that Oracle could meet with greater ease and control if it owned QLogic.

A third interconnect candidate is Voltaire, which offers InfiniBand switches and adapters with an emphasis on server clustering for high performance computing. Unlike Mellanox and QLogic, Voltaire doesn’t appear to be in the chip business and their unique IP seems to be in fabric management and application acceleration software. So, depending on how literally one takes Ellison’s remark about “chip” companies, Voltaire may or may not be a good possibility.

On the other hand, as of this writing, Voltaire has just become the subject of speculation as a possible acquisition by Oracle nemesis IBM as an adjunct to its purchase days ago of Voltaire partner Blade Networks. That could conceivably make Ellison more interested in buying Voltaire as a way to mess with IBM, in addition to its obvious face value.

Tell-Tales
However, one phrase that figures quite prominently in Voltaire’s positioning and may make it less attractive to Oracle: “scale-out.” In this context, scale-out represents the idea of building up computing capability by lashing together large numbers of small commodity boxes or blades, and it’s an idea with significant proponents. Google, for instance, is essentially the biggest scaled-out system ever created.

But, Ellison is more of a “scale-up” kind of guy. Scale-up means building the biggest box you can and, if you must, then building another and hooking them together. In fact, in his Exalogic demo he suggested that two of those machines could handle all of Facebook’s 500 million users. It is a safe bet that presently Facebook is running on way more than two big machines.

So, touts take note, as you consider which companies may be on Ellison’s buying radar, be skeptical about any chip and component companies flaunting “scale-out,” along with “standards,” “multi-vendor,” “open” or anything else that doesn’t make sense in the context of Ellison’s vertically integrated, all-Oracle “big honkin’ cloud-in-a-box.” And, if a company makes most of its money selling things that don’t go in a data center, forget it.

Look instead for the companies making patented proprietary chips that get soldered onto other companies’ boards, and/or those making heavily engineered boards that get built into major vendors’ servers. And, give extra credence to any cases where the purchase will give Oracle immediate product differentiation in terms of speed, capacity and throughput, give it technical leadership and competitive leverage, or enable it to steal a long march on IBM. If you understand Ellison, you know those are the things he cares about the most.

OpenOffice Kidnapped

A bunch of European and Brazilian developers kidnapped OpenOffice.org (OOo), the Sun-started challenge to Microsoft Office, Tuesday to the cheers of Oracle enemies like Google, which Oracle is suing for infringing on Java, Red Hat, whose code Oracle poached for it own Linux, Novell, whose own version of OpenOffice has patches Sun wouldn’t accept, and Canonical, which has a relationship with IBM, Oracle’s biggest enemy of all.

IBM itself, however, whose Lotus Symphony product is based on OpenOffice, seemed a little taken aback by the kidnapping and – apparently more concerned with the Open Document Format (ODF) than OOo – has adopted a wait-and-see attitude to the development.

The high-handed note the kidnappers left announced “a major change in the project’s structure” and said simply, “from now on, the OpenOffice.org community will be known as ‘The Document Foundation.’”

It said that Oracle, which acquired OpenOffie.org assets as a result of its acquisition of Sun Microsystems and so owns the OpenOffice.org trademark, “has been invited to become a member of the new foundation, and donate the brand the Community has grown during the past 10 years. Pending this decision, the brand ‘LibreOffice’ has been chosen for the software going forward.”

It also wants “the other assets [Oracle] holds in trust for the Community, in due course, once legal etc issues are resolved.”

That’s as much as they want to do with Oracle, which is not exactly a hero of openness, is given to having a tighter grip than Sun, and has had little to say about its intended Google-y Oracle Cloud Office since the name fell from its lips last winter. Presumably it is closed source.

Oracle, which is the biggest contributor to OpenOffice, has yet to respond to the news that there’s a new OpenOffice developer in town that means to “liberate the development of the code and the evolution of the project from the constraints represented by the commercial interests of a single company.” But judging from the icy silence with which Oracle greeted the attention-seeking ultimatum it got from the OpenSolaris governing board, it may be best for everyone to assume that the LibreOffice fork is going to be the LibreOffice fork – and, yes, there’s some English-only beta code for Windows, Linux and Mac OS X at libreoffice.org now.

Evidently a roadmap is expected soon too.

One of the Foundation’s first changes in how Sun and then Oracle ran things was to can the requirement that contributors turn over their copyrights. And unlike Sun, which had the commercial StarOffice, there will only be the one LibreOffice, governed by both the LGPLv3 and the Mozilla Public License.

One of the renegades, Charles Schulz, a member of the Community Council and Native Language Confederation lead, said in a statement that “The Document Foundation supports the Open Document Format, and is keen to work at OASIS [on] the next evolution of the ISO standard.”

The new movement expects to accelerate ODF adoption by government and the enterprise.

Google mumbled something about LibreOffice “encouraging further development of open source office suites” and Markus Rex, who runs open platform solutions at Novell, hoped “to see LibreOffice do for the office productivity market what Mozilla Firefox has done for browsers.” It appears that the enhancements in Go-OOo, Novell’s version of OpenOffice, will be merged into LibreOffice.

Canonical means to ship LibreOffice in future releases of Ubuntu and Simon Phipps, who was Sun’s chief open source officer, didn’t make the Oracle cut and is now a director of the Open Source Initiative (OSI), is looking forward to seeing the innovation a “truly open community” can drive.

The foundation, which is being led by a steering committee of developers and national language project managers, is setting up its own web infrastructure because it seems that “since Oracle’s takeover of Sun Microsystems, the Community has been under ‘notice to quit’ from our previous Collabnet infrastructure.”

It is soliciting funds and resources from companies and telling they won’t have to worry “that they may be helping a commercial competitor.” It needs developers, localizers, documentation writers and QA testers. Its bank account is being handled by OpenOffice.org Deutschland eV.

So far endorsements have come in from the Free Software Foundation (Richard Stallman himself), the French and Norwegian OpenOffice communities, Brazil’s BrOffice.org, the Gnome Foundation, NeoOffice, an OpenOffice fork for Mac, credativ, a UK open source consulting and services outfit, Collabora, a UK-based open source consultancy, and Liberix, a non-profit Czech open source support operation.

If the Supreme Court Hears Microsoft’s Case It’s Gonna Be One Crowded Courtroom

Practically everybody and his brother is backing Microsoft’s appeal to the Supreme Court in its attempt not to pay Canadian ISV i4i $290 million for Word’s alleged infringement of an i4i custom XML patent that Microsoft claims is invalid.

Google, Apple, Intel, Yahoo, HTC, Dell, HP, Facebook, Wal-Mart, Toyota, Intuit, Netflix, Verizon, the Electronic Frontier Foundation (EFF), the Apache Foundation, the Computer and Communication Industry Association (CCIA), the Securities Industry and Financial Markets Association (SIFMA) and the Clearing House Association, 35 law professors and others have all filed amicus briefs trying to persuade the court to hear the long-shot case.

Microsoft is challenging the judicial notion of “clear and convincing evidence” used by the Federal Circuit to uphold the district court’s decision that i4i’s patent is valid. It says the standard should be changed back to the old pre-1982 rule of a “preponderance of evidence.” That way defendants would have more latitude in proving patents they are supposed to have violated aren’t valid in the first place and wouldn’t have been if the patent examiners had seen all the evidence.

The Federal Trade Commission favors the “preponderance” standard and says the “clear and convincing” proof demanded by the Federal Circuit simply whistles through too many really bad patents.

The amici, who have millions riding on the decision, argue that “preponderance” – the same standard allowed a plaintiff in proving infringement – could save endless frivolous litigation and put throw a money wrench into the troll business.

According to Law.com, if the Supremes take the case it “could lead to a historic realignment in the patent litigation arena.”

RIM Unveils its Tablet

Having been kicked in the teeth by both iPhone and Android smartphones this year, Research In Motion unveiled its rumored iPad copycat Monday in an attempt to stop both Apple and Google from leaching more of its revenues.

The thing is called the Blackberry Playbook, a clear appeal to the company’s largely male installed base of Crackberry heads.

The tablet is positioned as a multitasking “enterprise-ready professional tablet” “perfect for either large organizations or an ‘army of one.’” However, it may have a split personality. Beside corporate power games, it’s also supposed to be good at playing traditional games.

Unlike Apple, the Playbook is “Flash-loving.” It supports the Apple-scorned Adobe Flash Player 10.1 and Adobe Mobile AIR as well as the Apple-doted-on HTML 5 so RIM can brag that Playbook will deliver “the real, full web experience to mobile users.”

Otherwise it’s smaller, thinner and lighter than the iPad measuring less than half an inch thick (5.1″x7.6″x0.4″) and weighing 14 ounces with a seven-inch 1024 x 600 touchscreen display.

Initially the Playbook won’t be able to connect to the Internet directly, which is actually a plus in enterprise terms.

It’s fitted with Bluetooth and Wi-Fi and in the absence of Wi-Fi it piggybacks on the Blackberry via Bluetooth to access the Internet (no separate service plan). That means that the smartphone’s content is viewable on the tablet, but actually stays on the phone and is only temporarily cached on the tablet (subject to IT policy controls). It disappears once the connection is broken so there’s theoretically less of a security risk and less of a manageability issue.

Eventually RIM is supposed to come up with 3G and 4G versions of the tablet.

The thing is powered by an unidentified 1GHz dual-core processor running on the Unix-like symmetric multiprocessing QNX Neutrino kernel that RIM acquired when it bought QNX earlier this year for about $200 million. For purposes of the Playbook the operating system is called the BlackBerry Tablet OS and is reputed to offer a highly responsive and fluid touchscreen experience for apps and content services. It’s eventually supposed to replace the Blackberry OS in the company’s smartphones.

The Eclipse-supporting Neutrino, which runs on the PowerPC, x86, Mips, SH-4 and ARM chips, is supposed to be one of the most reliable, robust and secure (Common Criteria EAL 4+) real-time operating system architectures in the world. It supports mission-critical applications in everything from planes, trains and automobiles to medical equipment and the largest core routers that run the Internet not to mention, as Bloomberg does, the control systems in nuclear power plants and the US Army’s unmanned Crusher tanks. (Top that for reliability.)

The OS is POSIX-compliant so C-based code is portable. It supports Open GL for graphics-intensive 2D and 3D applications like gaming, and will run applications built in Mobile AIR as well as the new BlackBerry WebWorks app platform so apps (like Java apps) written for BlackBerry 6-based smartphones will also run on PlayBook.

The widget includes dual 3 MP front-facing, 5 MP rear-facing HD cameras for video capture and video conferencing that can both record HD video at the same time, and an HDMI-out port for presenting one’s creations on external displays. It has 1GB of RAM, way more than the iPad.

The dingus won’t be in retail store in the United States until early 2011 with rollouts in overseas to follow in Q2. However, developers and some corporate customers will get the widget in October for development and early testing. An SDK is supposed to be out in the next few weeks and developers can register for early access at www.blackberry.com/developers/tabletos. The thing is supposed to use industry standard APIs.

There’s no price yet although Reuters says it should come in at the “lower range of prices for consumer tablets already in the suddenly congested market.

See http://www.youtube.com/watch?v=eAaez_4m9mQ&feature=player_embedded.

IBM Buys Blade Network Technologies

IBM says it’s buying privately held Blade Network Technologies, a data center switching outfit it’s worked with since 2002, for some unrevealed price rumored to be around $400 million. Last Monday IBM bought the publicly traded Netezza for $1.7 billion.

Blade started out in 2002 as a Nortel unit spun out four years later as Nortel went down the tubes a private company backed by Garnett & Helfrich, Terry Garnett having been a Larry Ellison lieutenant.

Blade has a blade server, some top-of-rack switches and software to virtualize and manage cloud computing and other workloads. Its devices route data and transactions to and from servers. The widgetry can improve systems performance, deliver information faster, optimize virtual environments and lower energy use.

Blade claims half the Fortune 500 as customers, many of them joint clients with Big Blue. IBM says actually upwards of 50% of its System x BladeCenters currently attach to or use Blade products.

IBM expects to close on the deal in Q4. It wants to be able to manage the new, more demanding workloads. It says that with Blade, it can drive innovation at the systems networking level to speed the delivery of key information from system to system – for workloads such as analytics (an IBM favorite) and cloud computing – while reducing data center costs.

Since Cisco got into the server business last year, IBM has been expanding its network offerings through relationships with other networking companies. Now, without abandoning the others, it’s bringing one in-house.

Brian Truskowski, the general manager of IBM’s system storage and networking unit, said in a statement that “Blade will help IBM better integrate networks with its systems, optimizing them for workloads that require high-speed and low-latency performance.”

IBM said Blade’s software helps address the massive virtualization requirements of cloud computing environments. With servers more closely integrated with the network users can deploy thousands of virtual machines to run large application workloads in the cloud and reduce complexity through simplified management.

Blade’s got nine million ports installed. It’s also got its own Unified Fabric Architecture, an interoperable converged fabric for “Ethernet Everywhere.”

Blade, which has around 200 people, said last September when it picked up a B round of undisclosed size that it was valued at $230 million. The money came from NEC, Juniper, its original backer Garnett & Helfrich Capital and a “silent investor” only identified as an “industry leader.” It said then that it had passed six-million Ethernet switch ports. NEC, which got a board seat, uses Blade’s Ethernet switches for the network infrastructure for its SIGMABLADE blade servers.

Microsoft Kills its Blogging Platform

Microsoft is throwing in the towel on its six-year-old blogging platform Windows Live Spaces (née MSN Spaces). It’s sending its bloggers over to WordPress.com, a competing platform, where Microsoft’s current crop of scribblers can get free space.

The two have worked out a migration scheme and Microsoft will redirect URLs. Bloggers can use Microsoft’s facilities until the end of the year but have to be out by March.

Microsoft counts 30 million “active” users. In comparison WordPress, written in PHP and hosted op Apache, for Pete’s sake, powers over 8.5% of the web, is used on over 26 million sites and is seen by over 250 million people a month.

This is the new practical “why-waste-the-money” Microsoft, which has decided not to fight the social networking crowd with look-a-like services and simply integrate with them instead.

Oracle Sues Micron Technology

Oracle sued Micron Technology last Friday for overcharging Sun artificially inflated prices, an antitrust complaint, when Sun bought upwards of $2 billion worth of DRAM from the DRAM cartel between 1998 and 2002.

Micron was never charged when the Justice Department investigated it, Hynix Semiconductor, Samsung Electronics, Infineon Technologies and Elpida Memory eight years ago for collusion and criminal price-fixing because Micron co-operated and copped a plea. The upshot was that the conspirators, including Micron, wound up paying $731 million in fines and some of their executives drew jail time.

Sun sued some of the culprits back before it was Oracle. Now Oracle wants restitution and triple damages from Micron, which Sun never got around to settling up with. The suit claims Sun lost money and customers to the high DRAM prices caused by artificially stifled production.

‘Out, Out, Damn’d Cisco’: HP

Sounding as obsessive-compulsive as Lady Macbeth, HP said Monday that it had finally ripped all the -supplied WAN routers and switches out of its six data centers and replaced them with its own networking widgetry.

HP swore to go Cisco-free in April when its acquisition of network pioneer 3COM closed. You can’t very well give aid and comfort to a one-time friend that’s turned on you and now wants to sell servers.

The jihad took a motivated HP less than six months – coming in ahead of schedule with no data center downtime. It claims it’s getting better-than-expected performance.

The replaced routers and switches let HP’s data centers in Austin, Houston and Atlanta talk to each other and to the Internet. The company says it’s got an open standard-based network architecture now. It will now move on and install its own products on the fringes of the network connecting company facilities worldwide.

The exercise should help push HP’s networking front against Cisco. According to HP CIO Randy Mott, “When we talk to customers around the world, they’re looking for a full networking portfolio provider with open architecture enabling an alternative to the proprietary protocols they’ve been locked into for decades.”

CA Technologies Buys Hyperformix

CA Technologies said Tuesday that it’s buying Hyperformix, a small 50-man Austin, Texas server virtualization capacity management ISV, without saying what it is paying.

CA figures the widgetry will bolster its virtualization management portfolio and, in turn, enhance its cloud cred.

It says, “An important aspect of our virtualization and cloud strategies is to help customers overcome the ‘VM stall’ challenges they encounter as they roll out virtualization and progress to a dynamic, cloud-based data center architecture.”

VM stall is supposed to be what happens after organizations virtualize all the low-hanging fruit – their low-risk, low-impact systems – dev/test servers, web servers, file servers, internal applications – and stop short of virtualizing their tier 1 business services, customer-facing environments, enterprise-wide systems, third-party applications, multi-platform services and composite applications.

Hyperformix is supposed to move them off the dime by discovering how physical, virtual, hardware, software, storage, and network resources are being used and predicting what resources will be needed in the future. The upshot is supposed to be a mature dynamic data center and private cloud operation.

CA expects the acquisition to close by the end of December when Hyperformix will become part of the its virtualization and automation business. CA means to expand its integration with Hyperformix solutions to include its CA Virtual, CA Service Automation, CA Service Assurance and CA Cloud portfolios.

Hyperformix started out in life developing predictive software for machinery 20 odd years ago as Scientific Engineering Software. Post-2000 as Hyperformix it raised ~$34.8 million in debt and equity funding from such as Morgan Stanley Venture Partners, M/C Venture Partners and LTI Venture Lending.

Wikipedia says it has specific applications for managing the performance of SAP software and 20 patents approved and pending related to analytic and simulation modeling. CA Technologies and Hyperformix share a bunch of customers.

RightScale Pushes into Dev & Test

RightScale, the cloud manager, has moved into development and test with a cloud management solution that software development and testing teams can use to provision and use cloud computing resources.

The new RightScale Development & Test Solution Pack supports four development languages, PHP, Java, Ruby and .NET, and includes Java and PHP solutions from RightScale’s buddies IBM and Zend Technologies.

The IBM Java stack includes the IBM DB2 database and its WebSphere application server, available either in the freebie Community Edition or the commercial Enterprise Edition.

Zend, for its part, delivers a tested stack that incorporates PHP, the Zend Framework, required extensions and drivers, and a set of application management capabilities developers might find handy on the cloud.

Zend says that with the RightScale platform, developers can leverage Zend Server capabilities, such as code tracing, performance optimization and job queuing, to build and test scalable applications faster without spending time setting up multiple server environments.

RightScale’s software pack is supposed to be an easy on-ramp to cloud computing, especially for companies that aren’t ready to move production applications to the cloud, sort of a learning experience without any long-term commitment or risk.

The widgetry includes RightScale’s core Cloud Management Platform with two different user interfaces – a self-service interface designed for developers and testers to launch servers in the cloud, and an administrative interface for customizing RightScale’s standard environments.

It also comes with two pre-configured server environments tailored for different stages in the dev & test process: an “All-in-One” configuration for doing basic integration testing on a single server in the cloud and a “Multi-Tier” environment for more complex testing with separate load balancing, application and database servers.

Of course once in production, operations staff can use RightScale’s Cloud Management Platform to monitor and manage system performance, users and costs.

The widgetry, which is supposed to maintain more consistent, reproducible environments from design through production, is available as Software-as-a-Service (SaaS) for an introductory subscription price of $250 a month for the first year.

FalconStor CEO Leaves after Shenanigans

FalconStor Software president, CEO and chairman ReiJane Huai resigned Wednesday following “his disclosure that certain improper payments were allegedly made in connection with the company’s contract with one of its customer.”

The Long Island company issued a statement saying it “has fully cooperated with law enforcement authorities with respect to the ongoing investigation, and it will continue to do so. In addition, a special committee of the board has been formed to conduct a full internal investigation of these matters and the special committee has retained counsel to assist it in its investigation.”

Eli Oxenhorn, a current board member who’s association with Huai goes back to their Cheyenne Software days, has been named non-executive chairman; chief strategy officer James McNiel has taken over as interim CE0 and interim president; and CFO James Weber is also now interim COO.

The $90 million-a-year publicly traded availability house lost money last year. Its widgetry is sold OEM or branded through Acer, EMC, Fujitsu, Hitachi, Huawei, SGI, and others. Cheyenne was bought by CA, which is how CA came by its ArcServe backup line.

PalmPad Headed This Way

Stop me if this sounds familiar.

HP’s troops were out in force this week talking up the company’s mobile strategy. The company will major on the webOS 2.0 it got in July when it closed its $1.2 billion acquisition of Palm.

It’s dumped whatever Android tablets it had on the drawing board except for the ones it’s using to run its newfangled cloud-based printers and will field a Windows 7-based tablet ahead of its webOS-based PalmPad due out early next year.

It expects to own 17% of the global tablet market in 2013 and that market will probably be worth about $40 billion, it thinks.

WebOS is supposed to be better at running multiple applications simultaneously than Apple is.

HP has also dumped any Android- and Microsoft-based phone plans for Palm phones, which it will “aggressively” push through its giant PC base. It expects to stream music from the cloud to its smartphones, which will store the ones most listened to.

CodePlex Foundation Now OuterCurve

To avoid any more confusion between Microsoft’s open source forge site codeplex.com and the year-old open source CodePlex Foundation that Microsoft originally sponsored, the foundation has renamed itself the OuterCurve Foundation.

It says the forge shared its name on the hunch that there would be more affinity between the forge and the foundation but that hasn’t proved true. (Surprise. Surprise.)

It says it wanted a name that would make it an attractive investment for additional sponsors.

A branding company in Boston called Protobrand came up with the new name.

The foundation’s job is to provide a software IP management process and project development governance for corporate projects looking for open source developers. It currently has seven projects under management.

They Don’t Sell the Chips Larry Wants at Fry’s
HP Hires Ex-SAP CEO To Replace Hurd
OpenOffice Kidnapped
If the Supreme Court Hears Microsoft’s Case It’s Gonna Be One Crowded Courtroom
RIM Unveils its Tablet
IBM Buys Blade Network Technologies
Microsoft Kills its Blogging Platform
Oracle Sues Micron Technology
‘Out, Out, Damn’d Cisco’: HP
CA Technologies Buys Hyperformix
RightScale Pushes into Dev & Test
FalconStor CEO Leaves after Shenanigans
PalmPad Headed This Way
CodePlex Foundation Now OuterCurve
iPad Speculation
Oracle by Accident
The Promise HP’s New CEO Has To Deliver
Don’t Hold Your Breath for Facebook IPO
HP Wastes No Time Sopping Up 3Par
IE9 Watch
EC Stops Looking Up Apple’s Skirt
Chinese Crazy for iPhone
Office Web Apps Watch
DOJ Squelches No-Poaching Deals
Salesforce Reportedly Acquire Activa
How Much Does IBM Like Voltaire?
Tibco Buys OpenSpirit
Compal To Build New Laptop Factory

It’s unlikely that anybody in the industry needs three anonymous Reuters sources to tell them that NetWare has been the stubborn donkey on the bridge holding up the sale of Novell for lo the six months the company’s been in play.

Everybody has known for years and years that that thing has been an albatross around Novell’s neck. Heck, if Novell had been able to solve its NetWare problem it wouldn’t have bought SUSE Linux.

Anyway, Reuters says that Novell can’t get whatever ridiculous price it’s demanding for the little bloodsucker and that if it can’t unload the thing at its valuation it may abort the sale of SUSE and stay independent, or IPO a division, or something. It doesn’t want to sell off SUSE to VMware or CA or whoever else has come a calling and be left there with NetWare in its hand.

Reuters identified Los Angeles-based Platinum Equity as having tired of the “onerous” and “difficult” process of trying to buy Novell’s legacy assets and quitting the field. Its bid some time back didn’t meet management expectations.

Other candidates to buy the assets were identified as Platinum’s buyout partner Gores Group and Attachmate, fingered last week by the Wall Street Journal as a possibility, as well as Vector Capital, Vista Equity Partners and Symphony Technology Group.

SUSE could reportedly fetch a rich multiple but evidently not enough to cover the low value placed on the rest of the company, which one of Reuters’ sources aptly called a “dying cow.”

Another source said a combined offer might reach $6.25 a share, 50 cents better that the rejected Elliott Associates bid of $5.75 that set the Novell pinball in motion.

If a sale falls through, one can easily imagine Novell sinking back to its pre-Elliott sub-five dollar price followed by a hostile tender for, oh, $5.50. One can also imagine investors filing lawsuit against management.

As the Reuters story got read Wednesday night, Novell’s stock price fell 7% after-hours from its $6.51 close to $6.09. It closed at $6.08 Thursday after spending time around $5.93.

A 75-man start-up called Aryaka Networks, backed by a $14 million A round from Mohr Davidow Ventures, Trinity Ventures, Nexus Venture Partners and Trinity’s little friend Stanford University, is lifting application acceleration and WAN optimization into the cloud.

It’s supposed to give enterprises a high-performance, highly available and scalable application-aware network whenever and wherever they do business. Its SaaS widgetry is said to be a first of its kind, liberating users from the current box-based appliances and managed service solutions that can take months to test and implement.

It claims to be the only true innovation in the WAN optimization market in the last five years that addresses the business requirements for high capacity and real-time optimization of all applications.

Being SaaS, Aryaka’s widgetry, a fundamental architectural and business model shift that supposedly couldn’t be done, can be deployed in minutes with no capital expense. The company claims users can see a 158x performance improvement and a 500%+ return on investment.

Trac Research considers it disruptive, “taking what has traditionally been a hardware play and turning it into a service, and, on top of that, they’re providing the connectivity – all in a single solution.”

However, Trac also says that Aryaka’s news “should be placed into the context of…a recent announcement from Virtela, who are providing a managed service for WAN optimization at $5 per branch office per day, with a full guarantee for response times of applications that are being delivered over the WAN. This solution comes in two flavors: software and cloud-based, both of which mitigates the need for deploying any dedicated WAN optimization hardware at the branch.”

The 10-year old managed network services company promises to speed applications 10x-25x and says it will refund 250% of the service cost if users don’t see that kind of performance boost. Its customers reportedly include Honeywell, Deutsche Bank, Sony, Kodak, FedEx, Pitney Bowes and Paramount Pictures and for channel partners it’s got IBM Global Services, Verizon, KDDI and SingTel.

Unlike Aryaka and the PoPs it leases from network carriers, Virtela says its service is based on a proprietary architecture it calls its Enterprise Services Cloud (ESC) that consists of 50 “Local Cloud Centers” in all the major financial centers around the world. It’s planning to add additional cloud-based security and mobility services based on its ESC architecture.

Anyway, back to Aryaka, Sanskrit for noble in a truthful kind of way. It’s the brainchild of CEO Ajit Gupta, who started Speedera and sold it to Akamai, which also revs up the delivery of Internet-borne content, for over a half a billion dollars. Speedera was a B2C public-facing web traffic play; Aryaka is an internal-facing enterprise play, meant to appeal to anybody with a branch office – the further away the better – particularly folks frustrated with the latency and packet loss of their communications.

At least it is to start with. If it doesn’t get bought up first it’ll probably eventually add a Speedera-style charter. But that’s a ways off. First it means to concentrate on the eight million branch offices in the world, particularly the mid-market crew belonging to companies with revenues between $50 million and $100 million, a market that competitors like Riverbed and Blue Coat have only penetrated to the tune of 5% or 300,000 boxes.

Aryaka’s platform, a play-on-words exercise in “thinking beyond the box,” can piggyback on any long-haul pipe and tickle it into handling 5x-10x more traffic quicker or, if that doesn’t work, find an alternate route. It works with Exchange, Office, VoIP, video and your basic file transfer, stuff like NFS, FTP, CIFS, MAPI, HTTP and checking code in and out, roughly 95% of what’s communicated.

Gupta says the company is going to focus on the world’s busiest intersections and guarantee 5Nines uptime. It has 10 hubs in places like New York, Silicon Valley, London and Bangalore and expects to have 25 by the end of the year. It promises enterprise-grade security complements of encryption, visibility through a web-based MyAryaka portal with executive dashboards, and of course control.

The solution offers comprehensive real-time data and alerts on the mix of applications, as well as performance and availability criteria, integrated into the management console accessible anywhere, anytime.

Being in the cloud, Gupta says, makes it the equivalent of a private jet at commercial prices or a private cloud with the economics of a public cloud. Pricing depends on the link size connecting to the network and the associated geography but Aryaka figures an organization spending $6,000 on dedicated connectivity with six locations across the globe could benefit to the tune of a million dollars over three years, bearing in mind the network savings and productivity gains – translating to 567% ROI.

Aryaka, which got organized in November of 2008, has multiple patents filed on its proprietary, built-from-the-ground-up technology, a mixture of deduplication, compression and caching.

Gupta may go out for more financing before the end of year but expects to make Aryaka cash flow positive for under $20 million. The A round will be used for R&D in new applications for its platform, extending global deployment of its network and expanding sales and marketing.

The total addressable market for WAN optimization is projected to reach $4.27 billion globally by 2014 according to Gartner Dataquest.

Start-Up Puts the Network in the Cloud
Novell Sale May Fall Through: Reuters
AMD Feels Revenue Pinch
Larry’s Got Himself a Cloud Machine
IBM Buying Netezza Likely To Annoy Oracle
Adobe Abandons Linux Flash Builder
The Chip That Will Not Die
HP Drops its Suit against Hurd
Oracle Countenances EC2
Oracle Wants Its Apps To Run on Its Linux
Oracle Trundles Out MySQL 5.5 RC
Oracle Overhauls its Pet Exadata Box
Moffat To Go to Jail Sooner Rather Than Later
RightScale Takes in $25m Third Round
Oracle Sketches Out a Java Roadmap
SCO’s Unix Assets Up For Sale
Amazon Backs Cloud Storage Start-up
Red Hat Revenues Jump 20%
Big Data House Gets $30 Mil
Enterprise Software Good For $232b: Gartner
EC2 Gets New Widgetry
HP Plays to Oracle’s Card
Google Apps Claims 30m Users
iPad Shrinkage
RIM To Unveil BlackPad: WSJ
Dell To Try 7-Inch Tablet Next
Middle East Next Chip Mecca
Teradata Recruits Ex-BusinessObjects CEO for Board
Oracle Forms New Unit
Dell Gets New CMO
Microsoft Ups Dividend
Hitachi Moves on Hush-Hush IPO
Greenplum & Cloudera New BFFs
SUSE Optimized for SAP Apps

HP confirmed Monday morning that it means to buy ArcSight, the 10-year-old anti-hacking house, for $43.50 a share or around $1.5 billion all told, a 24% premium for a company that it reputedly wanted to buy for $800 million a couple of years ago before it went public.

Since the unconfirmed report in the Wall Street Journal Sunday afternoon suggested an auction, the punters bid ArcSight’s stock price up over the $43.50 HP said it’s going to pay, hoping for a repeat of the 3PAR run-up.

The Journal said it heard ArcSight was shopping itself around to a bunch of big technology companies and looking for $42 a share. It was at $28 on August 26 when paper first reported it was for sale and closed Friday at $35.10.

During an early morning conference call Bill Veghte, the ex-Microsoftee now running HP’s software unit, refused to speculate on the chance of another bidder coming forward.

He also wouldn’t comment on whether HP’s overpaying for growth and whether its tight-fisted former CEO Mark Hurd, now gone to Oracle, would have countenanced the ArcSight premium on top of the $2.35 billion it’s paying for 3PAR.

Hurd reportedly didn’t want to pay more than $750 million for 3PAR and passed on trying to outbid Dell when he was still at HP and the price was still at $1.15 billion. HP is also carrying $5 billion in debt and promised to buy back another $10 billion worth of its stock – on top of the $4.9 billion still to be used for that purpose – to make up for the thrashing it’s taken since it ousted Hurd. It has $14.7 billion in the bank although all of that is now theoretically spent.

At least ArcSight, unlike 3PAR, is profitable.

The company, which monitors corporate networks for suspicious activity, earned $28.4 million, up from $9.9 million, on revenue of $181.4 million, up 33%, in its fiscal year ended April 30. HP said it is growing at the rate of 40% a year with double-digit operating margins and has a reportedly thousand customers including HP-style Fortune 100s and 1000s.

HP said it expects the acquisition to close by the end of year and not dilute its earnings.

Veghte said HP would combine ArcSight’s real-time view of security events with its own view of corporate resources.

It will give HP broader visibility into events across IT operations, security and compliance; the ability to detect threats and risks by correlating both activity and state changes in real-time; and a constant feedback loop between build, manage and monitor to ensure that enterprises remain secure.

ArcSight CEO Tom Reilly said HP would be able to create “a new type of security solution.”

ArcSight is supposed to control the internal and external risks associated with cybertheft, cyberfraud, cyberwarfare and cyberespionage.

Last month Intel said it was buying McAfee for $7.7 billion. HP also bought Fortify Software, another security house, last month on undisclosed terms.

Novell, which was put in play in March when Elliott Associates, a hedge fund, offered to buy it for $2 billion, has struck an agreement-in-principle to break in two and spin off SUSE Linux to a strategic buyer with most of the rest of the company going to a private equity house, the New York Post said Wednesday.

The Post said it got it from “people close to the process” but the paper couldn’t shake the names of either potential buyer out of them only that the deal was at a “sensitive stage,” still three or four weeks away from being signed, and could fall apart at any minute. Both sides of the deal are supposed to close simultaneously.

Well, everyone immediately assumed VMware was the likely buyer of the SUSE piece having tipped its hand when it cut an OEM deal with SUSE recently and the Wall Street Journal waded in late Thursday saying it was VMware and that Novell and VMware still differed on valuation.

If SUSE goes to VMware, which Red Hat figures is its main competitor, is could be a problem for Red Hat and for Microsoft. It certainly levels the technical playing field some.

The Journal offered private equity-backed Attachmate, which sells software to companies with older applications, as maybe getting the rest of Novell including NetWare, the company’s faded and fading network operating system.

The Journal said that not only is price still an issue but so are how to split the sales force, patents and other IP. It repeated what the Post said about how any deals were still weeks away if it doesn’t all come to naught.

Novell, which has around a billion dollars in the bank, rejected Elliott’s $5.75-a-share offer as too low and went off to find an alternative. A tough sell. Buyers haven’t exactly been clamoring at the gate demanding to be let in. According to the Post the Solomon-esque exit will return more than Elliott’s $5.75 a share to stockholders.

Elliott’s unsolicited bid pushed Novell’s burned-out stock price up over an optimistic six bucks a share but with the long wait it has fallen back. After the Journal piece, it closed Thursday at $6.06 for a market cap of $2.13 billion, a lot perkier than it had been before Elliott.

Reuters quoted Brad Zelnick, an analyst at Macquarie Research, as thinking Novell could fetch better than $7 a share, say, about $2.6 billion altogether, a 35% premium to Tuesday’s close, Tuesday being the day before the Post story broke.

Uncertainty over the company’s future has had a negative impact on sales. For the first nine months of its fiscal year Novell earned $56 million on revenues of $605 million.

Some 20 companies were supposed to have expressed at least passing interest in Novell but they were winnowed down and the Journal says there was a bidding round in May but that the “process has dragged on from there.” Novell recently asked for bids for the whole company as well as pieces.

HP Buying ArcSight for $1.5b
VMware After SUSE: WSJ
Neon Asks Judge for Early Decision in IBM Antitrust Case
Oracle Beats All Forecasts
HP & IBM Vie over Next Billion-Dollar Acquisition: Globes
Parallels Puts Windows on iPad, iPhone, iPod Touch
IE9 Betas
HP Paints Target on its Back; IBM CEO Shoots Arrows
Verizon Seeks To Overshadow Amazon & Rackspace Clouds
CSC Uses Skytap To Get to the Cloud
AMD’s First Fusion Chip Evidently Smokes  
Terracotta Claims To Solve Java’s Memory Conundrum
Ex-IBM Bigwig Sentenced to Six Months in Jail
Intel Leads Investment in Joyent
Microsoftee To Run Nokia
RightScale Gets Elastic Data Scaling from NorthScale
Intel AppUp Store Up
IBM Buys OpenPages
GoDaddy Reportedly For Sale
Intel Backs Cloud Manager
Dell Spending Big in China
Cisco Throws Lure of Dividend in the Water
Hurd To Speak
Red Hat, NetApp Potential Takeover Targets: UBS
Biggies Reportedly Close to DOJ Deal on No-Poaching Agreements
iPad Hits China
Red Hat Hires Public Policy Guy
The Americanization of Linus Torvalds
HP’s Got More than One Problem
Appeals Court Squelches Second-Hand Software Sales
Happy Birthday
Rackspace Taps New EMEA Chief
RIM Buys Office-Compatible Widgetry
CA Hires IT Management SaaS Boss

Apple’s unyielding and highly vocal opposition to Adobe Flash moderated Thursday morning when all of a sudden it opened up its vaunted App Store to iPad and iPhone programs written using other people’s widgetry.

Although it wrapped the surprise announcement up in “we’re listening to developers” ribbons, its real reason for changing parts of its iOS Developer Program license is more likely the reportedly parallel investigations of its business practices in the mobile software market by the Federal Trade Commission and the European Commission.

The dead giveaway would seem to be that the restrictions that have changed are exactly what get up a regulator’s nose.

Apple’s decision may also have been colored by the widespread adoption of the Flash-toting Android and the fact that – as it was – developers had to write two versions of the same app, one for Android, one for Apple, a potentially self-limiting proposition.

Not that Apple is going to let iPads and iPhones visit Flash-enlivened web sites or access Flash-based applications. No. No. No. No. Good heavens, it hasn’t gone that far. That would mean installing Flash on the things. No, but developers can now write programs for the widgets in whatever they please and then translate them into Apple-speak.

As Apple explained in a statement “We are relaxing all restrictions on the development tools used to create iOS apps, as long as the resulting apps do not download any code. This should give developers the flexibility they want, while preserving the security we need.”

Adobe indicated it was “encouraged.” And it surely didn’t mind that its stock spiked 12% on the partially gratifying news that its all-important Creative Suite 5 would have another reason for being.

In April right before CS5′s arrival – replete with its one-button transformation of an app written in Flash into an iPhone app – Apple slammed the door on its App Store shut, changing its rules and restricting developers to using only Apple technology. Private APIs were outlawed and languages used restricted to C, C++, Objective-C and JavaScript, making CS5 a no-go area.

A war of words then ensued with Steve Jobs making some very blunt comments about Flash’s technical shortcomings. Adobe complained to the antitrust regulators.

Apple’s sudden permissiveness includes permitting the use of Google AdMob analytics widgetry on iPad and iPhone ads.

Apple also ripped the veil off its mysterious and seemingly arbitrary criteria for admission to the App Store and published its App Store Review Guidelines detailing what is acceptable and what is not. It expects this unprecedented transparency to result in “even more successful apps for the App Store.”

There are now 250,000 applications for iPad and iPhone in the Apps Store, and altogether they have been downloaded 6.5 billion times earning developers a collective billion dollars, Apple said.

The parts of the license that Apple relaxed are sections 3.3.1, 3.3.2 and 3.3.9.

See http://developer.apple.com/appstore/guidelines.html.

NetApp and Oracle have called off their ZFS litigation, settling without telling anybody what the terms of the settlement are or if anybody’s licensing anything from anybody or paying any royalties. However, NetApp said Thursday that it and Oracle want the lawsuits dropped without prejudice, meaning neither of them is retreating from their position and the thing could flair all over again or presumably with somebody else.

In case you’ve forgotten the muddle, NetApp sued Sun Microsystems back in 2007 claiming that the vaunted Zettabyte File System (ZFS) that Sun open sourced in 2005 infringed on seven of its patents.

At the very least it wanted Sun to stop insisting that NetApp was infringing on three Sun patents and definitely wanted Sun to stop demanding royalties.

A very putout Sun, after claiming NetApp had tried to buy the Sun patents it was now looking down its nose at and wanting ZFS un-open sourced, then sued NetApp back for infringement (twice actually), claiming NetApps’ patents were invalid.

In disclosing the settlement Thursday NetApp CEO Tim Georgens made a point of saying, “Moving forward, we will continue to collaborate with Oracle to deliver solutions that help our mutual customers gain greater flexibility and efficiency in their IT infrastructures,” perhaps revealing a pressure point who knows.

Sun also got some favorable rulings from the court and the Patent Office last year like a decision that it didn’t infringe NetApp patent 6,892,211 while the PTO rejected some of NetApp’s claims involving patent 5,818,292 on the basis of prior art but not enough for NetApp to turn tail.

The settlement should presumably – underscore presumably – make Coraid, Nexenta, Compellent and GreenBytes, which all use ZFS, happy. Coraid just got a cease-and-desist a few weeks ago from NetApp concerning its ZFS-using EtherDrive Z-series NAS widgetry, but that’s as far as NetApp’s lawyers went down the list, suggesting they may be safe.

© 2012 LinuxGram Suffusion theme by Sayontan Sinha