VMware is perhaps the best example of this phenomenon. No longer is the company positioning itself as simply a pioneer of server virtualization, but rather it is now betting its future on the broader software-defined data center. VMware dominates the server-virtualization market (its technology lets a company run hundreds of virtual servers on one physical server). It’s no surprise, then, to see VMware accelerate its R&D schedules and M&A activity to extend its technology portfolio to also seize the infrastructure and storage markets that are up for grabs in the new software-defined data center.

In a major bid to own the leading infrastructure play in the new software-defined data center, VMware last summer acquired software-defined networking pioneer Nicira for $1.26 billion. That is a staggering sum that becomes even more impressive when one considers that, by most estimates, Nicira was generating less than $10M in sales. As part of its strategy to bite off a small piece of the emerging software-defined storage space, VMware also recently acquired Virsto for an undisclosed sum.

The rationale behind these acquisitions comes into clearer focus when you consider the larger opportunity posed by the software-defined data center. As data center workloads increasingly become virtualized, it makes sense that VMware, which already enjoys a market cap of more than $30 billion, look for ways to increase its role in managing the broader data center infrastructure.

So, with all of this in mind, what actually makes up the Software-Defined Data Center — and which companies stand to gain the most in each area?

Components of the Software-Defined Data Center

The concept of the software-defined data center revolves around making the three major infrastructure components of a data center (compute/server, networking and storage) more flexible, more automated and less dependent on the underlying physical hardware. The idea is to create a pool of available resources that can automatically adapt to changing workloads and ensure that the right resources are available whenever and wherever needed. When you look at the compute/server space, virtualization forever changed the way applications are deployed, and the dominant force behind this is VMware. While VMware has established itself as the market leader in server virtualization, offerings from Microsoft, Citrix and Red Hat are beginning to carve out sizable market share, as well. With almost 70 percent of workloads today running on virtualized servers according to IDC, this is certainly the most evolved component of the software-defined data center to date.

Networking

In the wake of the Nicira deal, along with major announcements from Cisco, Juniper and other networking giants, software-defined networking has become perhaps the next focal point of the software-defined data center discussion today. While not as mature as the server/compute side, the software-defined networking market is expected to grow; IDC predicts from $360 million in 2013 annual sales to $3.7 billion by 2016.

Cisco, which has long dominated the networking market and has a valuation of over $111 billion, has started to face new competition from startup companies like Nicira and Big Switch Networks, which designed their products for today’s virtualized IT environment. To go after this market, Cisco has invested $100M in a “spin-in” company called Insieme Networks. Cisco clearly views software-defined networking as one of the most significant technologies to emerge in decades.

Storage

The last component of the software-defined data center is storage, which is not coincidentally the trickiest part of the equation. The storage layer has traditionally been the laggard of the data center, and most venture capital firms have feared investing in startup storage companies due to the stronghold on the market enjoyed by technology giants like EMC, NetApp, HP and IBM. This has changed in recent years, however. The rise of virtualization and, more recently, cost-effective flash technology, has spurred a storage renaissance — today, storage is one of the hottest markets for venture investors.

The increased investment sexiness of storage helps explain the success of Fusion-io, which created a new memory tier based on flash technology. The company went public in June 2011, and is valued at more than $1.5 billion. Because of the huge impact of flash technology, some of the big legacy storage vendors have been looking for acquisitions to help modernize their product portfolio. Last summer, EMC acquired XtremIO for $400 million dollars to add flash to its own storage portfolio. However, flash is just one component of software-defined storage.

Flash is a very disruptive technology that has paved the way for dozens of new entrants into the storage market, but flash by itself doesn’t address the complexity and data management issues created in virtual environments. Most major storage vendors created their product architectures before virtualization even existed, meaning they were originally built for a physical world where application workloads were discrete, known and predictable. Indeed, many of the new storage startups have continued using the same architectures, albeit with faster flash storage rather than spinning disks. The problem is that the software-defined data center is possible only with virtualization. And adding new layers of software on top of these legacy architectures is an inefficient way to deal with the problem.

The move to the software-defined data center is the major technology shift of this decade, just as virtualization was in the 2000s and the Internet was in the 1990s. Like those previous shifts, there is a wealth of new opportunities for companies both new and old. It will be interesting to see how everything plays out — and, rest assured, this race has a long way to go.

Dr. Kieran Harty is a co-founder of Tintri Inc. and serves as its chairman and chief executive officer. Harty served as an executive vice president of engineering and R&D at VMware, and has more than 15 years of engineering and management experience with high tech companies. Before VMware, he was vice president of R&D at Visigenic/Borland and chief scientist at TIBCO. Harty has a PhD in electrical engineering from Stanford University and a master’s degree in computer science from Trinity College Dublin.

As of 3:05 pm ET today, Fusion shares were trading at $19.89, up $3.26 (or 19.6 percent) from a $16.63 closing price Wednesday. For one thing, the company reported quarterly results yesterday, and gave forward guidance for the current quarter that was better than anyone expected.

Another thing? There’s a new Facebook data center going up in Iowa. And as everyone who follows Fusion-io knows, Facebook and Apple are its marquee customers. A new data center means that a lot of new Fusion-io products are selling.

As CEO David Flynn pointed out in an interview this morning, that can be a blessing and a bit of a curse. Earlier this year, Facebook and Apple trimmed orders and Fusion was forced to trim its outlook. Now, with Facebook building again on a site that’s big enough to accommodate at least two more facilities just like it, there’s a brighter outlook. But if you take out the up-and-down side of Fusion’s business that caters to Apple and Facebook growth, Flynn said, there has been a nice, steady, predictable ramp.

There are also new customers to report: Box, the fast-growing enterprise cloud services company, has started adding Fusion-io products to its servers. So has music service Spotify.

Then there’s the matter of the $119 million acquisition of NexGen, a Louisville-based company that specializes in taking traditional hard-drive-based storage products aimed at mid-range companies and combining them with Fusion-io’s flash-based technology. The combination gives Fusion access to a base of customers it wasn’t previously reaching. “We started out reaching the companies at the top of the pyramid, and the fact is the size of the market opportunity in the middle market is bigger,” Flynn said.

The deal has Fusion paying $114 million in cash and $5 million in stock. It’s Fusion’s second acquisition this year. Last month, it acquired ID7, a British software firm.

As if you needed another indicator about how much the old Wintel world of PCs has flipped in the last couple of years, take a look at the earnings results of the British chip designer ARM, which just reported quarterly earnings this morning.

Sales rose by 29 percent year on year to north of 170 million pounds (or $260 million), which was better than expected. Earnings on a per-share basis were five pence versus the expected four pence, amounting to a beat of a penny per share. Its shares are rising by 9 percent both in the U.K. and on the Nasdaq in the U.S.

ARM, you’ll recall, is the company behind the designs that go into building the chips that land in most smartphones and tablets. Rather than make the chips, ARM licenses its blueprints to companies like Qualcomm, Broadcom and Nvidia, which then make their own chips. And since phones and tablets are growing a lot faster than traditional PCs (come to think of it, PCs actually aren’t growing at all), ARM is looking a lot healthier than traditional chip companies like Intel and Advanced Micro Devices. Here’s a pretty good indicator: Royalty payments for processors rose in the quarter by 33 percent versus a processor industry that’s up about 2 percent.

ARM is quickly turning out to be the company to watch in the chip space. Chips sporting ARM designs are everywhere these days, and there has been a lot of chatter of late about them heading into the data center.

Hewlett-Packard offers ARM processors as an option on its radical new server design, called Project Moonshot. Dell offers ARM-based servers, too, and there are even more plans for ARM chips in servers. I talked with CEO Warren East about this last year. (East is retiring this summer, by the way, and Simon Segars will be ARM’s new CEO, starting in July.)

The basic argument that ARM makes coming in is that its chips are good at managing power consumption, in part because they were designed from the beginning for mobile applications. And power consumption continues to be a huge problem, especially in data centers where thousands of servers are crowded together in one place.

Intel, the king of the chip world, has responded and created its own line of low-power chips called Atom. And as we learned from Mike Bell, head of Intel’s mobile chip business at D: Dive Into Mobile last week, it has gotten off to a slow start but is starting to get a little traction in mobile.

Another version of Atom, announced the week before last, will also defend Intel’s interests in the server space. But keep an eye on this, because there’s eventually going to be a rumble.

The Register says the data center will be 1.4 million square feet. There’s no word yet about what incentives, if any — tax breaks and the like — the state of Iowa is expected to provide. And there’s no word on how many jobs the facility will provide. Facebook is also said to want some tax credits for using wind power; that will require a vote by the state legislature.

State officials have been kind of hush-hush about the whole thing, and apparently there was pretty stiff competition between Iowa and Nebraska to win the deal. Google and Microsoft have recently opened data centers in the state, as well.

Intel shares have traded lower since last Thursday, when the market research firms IDC and Gartner said they had tracked one of the largest year-on-year declines in sales of personal computers since records have been kept. Intel is the largest supplier of microprocessors to PC manufacturers like Hewlett-Packard, Dell and Apple, and it’s hard to see how much good news it can possibly bring to the table today.

Analysts are expecting Intel to report a profit of 41 cents per share on sales of $12.6 billion, and missing either would be seen as more or less proving that the PC market is in a state of permanent decline. So would a weak outlook for the current quarter, for which analysts currently expect earnings of 40 cents on $12.9 billion in sales.

There are other aspects to Intel’s business. It has a healthy data center business selling chips for use in servers, but out of more than $53 billion in sales last year, $34 billion, or more than 61 percent, was in its “client,” or PC, unit, while the data center group accounted for about $10.7 billion.

In the past, Intel executives have quarreled with the analyst firms, and said it was seeing more promising conditions in emerging markets. Indeed, in prior years there has been a disconnect between the dour pronouncements of Gartner and IDC and the peppier market conditions that Intel would later describe in its financial results in places like Brazil, Indonesia and Russia. In more recent quarters, the differences between their views have narrowed.

Aside from PCs, Intel has some new ideas that it hopes will kick its data center business into a higher gear. And it certainly has higher hopes about selling more chips for use in phones and tablets, but as yet they’re only hopes. It also plans to launch a TV product later this year.

Aside from the numbers, expect some questions — and maybe even some answers, but probably nothing conclusive yet — about the search for a replacement for CEO Paul Otellini. The smart money says the choice will be an internal one (here’s a rundown on the contenders), though there’s a slim chance that Intel’s board might be in the mood to surprise everyone and name an outsider. But don’t bet any money you can’t afford to lose on that.

In comments at the Intel Developer Forum in Beijing overnight, Diane Bryant, senior vice president and head of the Datacenter and Connected Systems Group, described a rethink of how data centers might be designed. Currently, individual servers, each with its own computing and storage, are being packed tightly together in a rack, and in turn packed into a room with other similar racks.

Intel sees a world where all the computing and storage portions are separated. CPUs would be grouped together so they could be cooled together. They would in turn be linked to storage infrastructure by screaming-fast optical connections running as fast as 100 gigabits per second.

Eventually, Intel sees everything being separated into its own section of the rack: CPU, memory, storage and power. The point is that you’ll be able to upgrade one — swap out older memory modules for newer ones, or upgrade the CPUs, or replace a bad hard drive — without interfering with the operation of any of the other parts.

Intel’s play, Bryant said, is to offer a reference design — essentially a basic recipe for building the concept — to the big server manufacturers. And at least some portions of this concept are already operating in China. Intel has been working with Web commerce giant Alibaba, Web search concern Baidu, Chinese Internet company Tencent and China Telecom on something they call Project Scorpio. The idea is to centralize all the cooling and fans within the rack, and to demonstrate that you can save on operating costs.

There were new Intel chips disclosed, too. A chip code-named Avoton is due in the second half of this year. It’s a version of Intel’s lightweight Atom processor, aimed at small servers — not unlike HP’s Project Moonshot — that will be built on Intel’s latest 22-nanometer manufacturing process, and with a new design. In Intel’s recent “tick-tock” parlance, where it delivers a new design — a tick — then shrinks it with an new smaller manufacturing technology — a tock — this is effectively both.

Another chip, code-named Rangeley and also due in the second half of the year, is a 22-nanometer variant of Atom that will be aimed at networking devices, routers, switches and whatnot.

There were also three versions of the Xeon chip discussed. The next E3 will get some new tricks to support video analytics workloads. The next Xeon E5 will move to the 22-nanometer manufacturing process (that’s a tock), and will be available in the fall. Finally, a new Xeon E7 will be available in the fourth quarter. Its newest trick is that it can address three times the memory of its predecessor — up to 12 terabytes.

Today is the day HP will formally unveil a product upon which a lot of its hopes for transformation and a return to health have been placed. It’s called Project Moonshot, and HP has been talking about it for about 18 months.

Basically, it’s a server, a very small server that consumes very little energy. During a conversation earlier this year, Dave Donatelli, HP’s executive VP and head of its enterprise, showed me one. Smaller than a typical hardcover book, it consumes 89 percent less energy to operate, and takes up 94 percent less space than a typical server. And, when packed into a large rack with many more servers like it, the amount of computing power that can be harnessed in one relatively small place is pretty impressive.

It’s also highly customizable, in a nearly endless series of mix-and-match combinations: It supports Intel’s Atom line of small and light microprocessors, as well as new up-and-coming server chips based on designs from the British firm ARM. It can also support graphics processing units from companies like Nvidia, as well as standard hard drives or flash-memory based solid-state storage.

HP’s argument to the marketplace is that Moonshot is unique. In a world where companies maintain data centers either for their own operations or as a means of reselling cloud computing capacity, HP’s hope is that the appeal of lower operating costs over time — energy consumption is a big one — will appeal to customers looking to swap out older machines.

And by at least one simple metric, it is. HP’s rivals, including Dell, IBM and Oracle, have nothing quite like it. That doesn’t mean they couldn’t. It is increasingly popular — companies like Google and Facebook do a lot of this — for a company to assemble their own servers from off-the-shelf components. Indeed, chipmaker Intel will be discussing new reference designs — essentially a basic blueprint — for what it calls “micro servers,” based on new versions of its Atom processors, later this week.

The message of Moonshot is also larger. It’s a signal from CEO Meg Whitman that HP can still launch important, innovative products that are different from anything else on the market. That’s a key message, given the tremendous difficulty the company has found itself in over the last few years. Indeed, just last week, Chairman Ray Lane stepped down from that role, and two other directors resigned in response to a proxy campaign by shareholders unhappy with the botched 2011 acquisition of the British software firm Autonomy.

A successful product launch will help shift the HP narrative away from recurring boardroom and executive office dramas, toward the day-to-day business of being the world’s largest technology company. And that’s just as important as the product itself. The sight of HP experiencing a win in a key market segment will go a long way toward making it look as though the longed-for turnaround — one that seemed unthinkable only months ago — is really getting under way.

Fusion-io, charter member of AllThingsD’s Flash Madness club, creator of flash-memory based insert cards that turn ordinary servers into super-fast data-crunching machines, today announced an acquisition.

The target is an obscure British software firm called ID7. Its speciality is software that allows storage systems to talk to interface technologies. But here’s why this deal is important: ID7 is also a major contributor to the open-source software project called SCST.

If you’re creating any kind of sophisticated storage scheme for a computer running Linux — which is pretty much anyone these days — you’re using SCST. That includes a lot of Fusion-io’s rivals, among them storage giant EMC, as well as its partners like IBM and Hewlett-Packard. SCST is basically a must-have whenever you talk storage, and ID7 is the biggest contributor to it. What it means is that everyone doing anything in the area of storage, or working toward any kind of software-defined storage scheme, will have to deal with Fusion-io.

Financial terms are not being disclosed, but ID7′s team of software developers will join Fusion-io immediately.

The latest market data from research firm Gartner found worldwide server sales declined by 0.2 percent, even as revenue from sales of those servers increased by more than 5 percent. For the full year though, sales increased by 1.5 percent on a unit basis, while revenue declined slightly.

Budget worries caused many companies to hold back on replacing older x86 machines, Gartner said, especially in the enterprise and in some mid-tier data centers. That no doubt factored into some worries around Intel’s outlook when it reported Q4 earnings last month. The only ones really buying were big players like Google, Facebook and China’s Baidu. Sales of RISC-based machines running Unix, as well as specialized hardware like Hewlett-Packard’s Itanium servers, were also weak.

IBM made the most on a revenue basis, clocking in sales worth north of $5 billion, while HP sold the most on a unit basis, with a little less than 664,000 units.

On a regional basis, North America saw unit shipments grow 5.5 percent, followed by Asia/Pacific at 3.4 percent, and Latin America, which was essentially flat.

Europe, on the other hand, declined substantially, and that’s where a lot of the pain was. As sovereign debt and economic concerns continued to take much of the oxygen out of the room in those countries, server sales declined by more than 10 percent on a unit basis, Gartner said. Most companies there saw their unit sales decline except for Japan’s Fujitsu and Cisco Systems. In Cisco’s case, its unit sales grew by nearly 20 percent, but that was off a low base relative to the other vendors. By comparison, Cisco sold about 14,000 servers in the region, versus HP which sold nearly 248,000.

Looking ahead to the rest of 2013, Gartner said to expect modest growth, offset a little by a boost in the use of virtualization, which allows one physical machine to run as if it’s many virtual machines. Virtualization logically tends to eat into the overall volume of machines needed. Simply put, where you once needed two or four or eight servers, it’s pretty likely you can now, given the improvement in processing power plus virtualization, replace them with one or two or three.

Here’s the worldwide breakdown for the top five vendors in the quarter, revenue first.

I had a short conversation with Chambers last night, shortly after the earnings report was released. Here are the highlights:

AllThingsD: John, you were just on CNBC a little while ago, talking about how some of your customers are seeing a slow start to 2013, but generally feeling positive. Your shares moved after-hours in reaction to that, and people tended to focus on that. Can you unpack what you meant?

Chambers: I talked about a lot of positive things, and I think sometimes people will take one element and focus on it. When you have record earnings and record revenues, and you’ve done that for eight quarters in a row, the key takeaway about the quarter has to be that our vision and strategy is working. Many times — not entirely — but many times, we’re out-executing the vast majority of our peers. The second element is not that Cisco is off to a slow start in 2013, it’s more that 2013 is off to a slow start economically in terms of what our customers are saying to us. I believe, based on what they are telling us and based on our order trends, that it appears to be, barring government mis-execution in a big way, I think we can handle a short-term pause. But, barring a big economic surprise, it’s just going to be a slowly improving year. My customers are saying it’s looking like a better year than they’ve seen in a couple years, and the trend is up and to the right, but it is coming off a slow base. So I think we’ll see relatively slow GDP growth in the U.S.

That seems to be the sentiment right now.

Yeah, most of my customers are saying they see a point and a half of GDP growth in the first half of the year, and unfortunately that means they’ll spend to that. On the other side, our enterprise and commercial accounts are showing two quarters in a row of really solid growth. And they’re usually a good indicator two to four quarters out.

So would you say you’re expecting something of an acceleration later in the year?

I would use a different phrase for it. I’d say cautious optimism. There are some very smart people who are very negative not just on Europe but on the U.S. So our eyes are wide open to the fact that we may be of the minority view. But we have the advantage of seeing our order patterns among our customers, and being able to talk to most of the government global business leaders around the world.

What’s working and what’s not for Cisco right now?

I think what’s working are the vision and the strategy of the role of the network, and beginning down the path of being the No. 1 IT player in the world. This quarter was another proof point of that, especially in the data center, but also in software. Secondly, what’s working is key technology trends, such as cloud and data center, mobility and video — all had very good quarters. What’s not working is that we’d like to see things growing faster on a global basis. Assuming we continue to be a reasonably good barometer of the global economy, we’d like to see these things growing a little faster.

How do you feel about the competitive landscape?

We love to compete, and we try to always compete with class. And we compete to win the market transitions. So let’s look at those. In the transition in the cloud and the data center, we’re challenging the incumbents that have been there for two decades. Let’s face it, servers are a commodity. And yet our servers get a dramatic premium, and they are tied very tightly to a switch. And we’re winning with a 65 percent growth rate in a market where our peers are flat or growing in low single digits. The second thing that is working is mobility. It was an Achilles’ heel about two or two and a half years ago. Now we’re the leader in most segments of mobility that have profits. So mobile edge, mobile backhaul, mobile packet edge, wireless LANs, small-cell, integrating it all together with wired capability. And collaboration tied to it, it really feels good to win there. The third thing would be in video. We’re making a pretty good transition from a set-top box where we’re making very little profit, and we’re starting to walk away from deals where people won’t pay for an architectural play; we just don’t bid on it. And, in spite of that, we’re up about 20 percent, mostly on software from NDS, so the market is going where we want it to go. It’s too early to say how far along we are in becoming the world’s No. 1 IT player, or the Internet of everything. I’d say a little bit of progress there, but it’s way too early in the ball game to say.

The venture arm of the world’s biggest chip company joins other investors — including Goldman Sachs, Index Ventures, Khosla Ventures and Redpoint Ventures — and brings the total amount of capital raised to $45 million.

Remember, software-defined networking (SDN) aims to do to networking gear what virtualization companies like VMware have done to servers. In the same way that one server can be virtualized into many, all with different configurations, the point of SDN is to make networks as easy to spin up, configure and expand as virtual servers in the cloud, all of it done via software. VMware, for its part, is in the game via its purchase last year of Nicira, the first SDN company I ever heard of.

The idea is considered a metaphysical threat to established networking companies, specifically Cisco Systems; analysts, specifically J.P. Morgan’s Rod Hall, have worried that Cisco isn’t ready to meet the threat.

I had a quick chat with Guido Appenzeller (that’s him in the photo above, at the left of president and co-founder Kyle Forster), and he said that Intel, which does make some specialized networking chips, sees some alignment of interest with Big Switch in the future of SDN.

As I noted before, Big Switch’s approach has a lot of industry support. Its partners include Juniper Networks, Citrix, F5, Dell, Microsoft and Palo Alto Networks.

Now that the tech investment buzz cycle has pivoted in an enterprise-friendly direction, I thought it was time to check in with Levine, a partner at venture capital firm Andreessen Horowitz.

I talked to him late last year on the heels of a busy summer. In July, he led AH’s stunning $100 million investment in GitHub, then followed it up with an $11.2 million investment in Meteor, and in October, an investment in Udacity, an education startup.

Perhaps hinting that 2013 will be as busy as 2012, last week he led a $30 million investment in Data Gravity and joined its board of directors.

We had a pretty long conversation. His fundamental argument about why smart enterprise-focused startups make for good investment opportunities comes down to one simple notion: Pretty much everything that’s been running in the corporate IT environment — things on which companies will spend more than $3 trillion this year according to some educated estimates — is ripe for a significant disruption that will make it less costly to own and operatee, more efficient, less prone to failure and simpler to use. That means a lot of large companies who have a lot of skin in the game maintaining the status quo are likely to have their worlds seriously rocked in the coming years.

Sound like fun? It’s exactly what AH hired him for two years ago.

Here are some highlights from the first half of our conversation. I’ll post the second half soon.

AllThingsD: Peter, the thing I keep hearing people say these days is there’s this feeling that the venture community is sort of “waking up” to the enterprise, or that it’s suddenly cool. You, on the other hand, have been all about the enterprise from the start. What do you think about that?

Levine: Yeah, it’s a really good time to be investing in the enterprise. We see so many interesting companies in the space right now. Just for context, it’s interesting that you comment on the idea that people are just waking up to the enterprise, and that’s exactly how I have felt for awhile, that there’s this renaissance in enterprise computing. There’s an awakening up and down the stack for new products that are going to service the enterprise.

Like what?

In my mind there are three categories that are all being disrupted at once. Cloud infrastructure: All the underpinnings of it all — storage, security networking and virtualization — would all fit there. At the next layer there’s software-as-a-service, though now there’s almost anything as a service. That’s transforming all the on-premise applications. At the highest level, mobile is transforming how people are consuming information and data and applications they are using. … In the cloud infrastructure you can see someone like a Cisco as the old guard, or even server vendors to some extent, and the new folks being VMWare, which totally upended the server world. For networking, you might see Nicira and other companies in the software-defined networking space as examples. It’s early days in that. There’s a big shift in software driving new network infrastructure rather than hardware and components driving the network infrastructure. In the same way, a VMWare can turn a server into an infinite number of servers. With SDN, the hardware component, it may even be a server in this case, with software basically creating a virtual network.

I get the point on Software-Defined Networking, though lately Cisco would argue, and not entirely without merit, that it has some of its own SDN chops. But I think your point is bigger about incumbent companies in the enterprise.

Exactly. The point isn’t whether Cisco has the technical chops to go do SDN, and it’s not whether they see this trend or don’t see it. They see it as well as everyone else. The problem is that from a business model standpoint, it’s really hard to go from one side of the pillar to the other. We’re talking about the commoditzation of components that very strongly eat into revenue streams and how revenue is recognized. That may sound like a minor accounting rule. But it is huge. With software, if I have to prorate my revenue for three years over the course of a sales engagement, rather than book it all up front as you do with a hardware sale, it dramatically shifts the sales model, the revenue model, the go-to-market model. It’s not only about the technology. For someone like a Cisco, they may very well have the technology chops, but there it’s about commoditizing the very revenue stream you rely on for your existence.

So I presume there are more examples like this in other parts of the stack?

Of course. There’s software-as-a-service on one side. There’s SAP, which has a long history of running big on-premise apps. And then you go to the other side. The whole transformation began with Salesforce.com. Ten years ago, Salesforce.com was an outlier. It was heresy to even think about putting precious customer data outside your firewall. I think that Salesforce really paved the way for the entire software-as-a-service category. And now every part of that stack — from business intelligence to analytics, to performance management to CIO tools — are all moving from the on-premise equivalent to off-premise, SaaS-based equivalents. Every major on-premise vendor has an equivalent off-premise counterpart.

And yet here you have the same thing as in your first example: SAP and Oracle are buying SaaS companies and talking about how they, too, are running in the cloud.

It’s true. They can acquire their way in. They have a lot of money. However, I’ve worked at large companies, and when you’re in there, you have this revenue base, and have a history defining how you’ve done things for the past 20 years, from sales to engineering to architecture. So even at the executive level, you can have every intention to transform the business and move into the new cloud architectures and so on, as the strategy trickles down through the organization it becomes difficult to implement because of the way people are managed and it becomes very hard to swallow some of these acquisitions and have them flourish. Basically the old guard suffocates the new innovation by not letting it flourish.

So we talked about software-defined networks, and there was last year a lot of attention on new companies there. Where’s the next area of attention?

Three years ago we invested in Nicira under the assumption that this was coming. There are still opportunities there, but the nature of this business is such that when a successful outcome like that happens, there’s a flurry of activity, where the venture community tends to pile on the next 40 companies doing similar things. But there’s other areas. Storage is a really interesting area right now. If you look at computing being commoditized by VMware, and now networking being commoditized by SDN, storage is the most expensive component at that layer. And it has been dominated by the same architecture and the same companies — EMC, NetApp, Hitachi, IBM — for at least 20 years. Same architecture, same companies. So I do believe there is an opportunity for companies in that layer to be disaggregated. Now, people have been talking about that for a long time. I was one of the very early employees at a compay called Veritas in the 1990s. So it’s a space that I have an affinity for. The interesting part of that market is that people have talked about the commoditization of storage since then. That said, it’s actually pretty interesting.

So what is startup vs. incumbent dynamic in storage?

In the past there was only the enterprise data center. And a startup would have to go head to head against an established startup in the data center. Whatever you’re selling, storage or networking or security, you’re going head to head with the incumbent players. And for a startup it is incredibly different. There’s questions about service and support and features, and there’s a CIO who says, well you never get fired for buying X or Y. So what’s happened now — and it just became clear when I talked with a friend about this recently — is that whole cloud architectures that are being set up in parallel to enterprise data centers, maybe inside or outside a company, or whether it’s Salesforce or Facebook, or Zynga or Amazon, those are the companies that are very sensitive to price, anti-incumbent, they’re adaptable to letting startups come in. So where a lot of new startups are starting to get traction is with these new cloud architectures, and it might be within a corporate infrastructure, and where there’s green field opportunity. What you find in these environments is a lot of commoditization at every part of the cloud infrastructure layer — whether it’s compute, networking, security or storage — they’ve all been greatly commoditized as compared to the traditional data center. That parallel universe has given startups the ability to make an inroad by gaining traction and relevance. And those companies are the lead users. The cloud infrastructure has dragged the crusty old data center to innovate in a new way.

So who are the companies you’re starting to see and invest in doing that disruptive work in storage?

We recently did an investment in a company called Convergent-io, which is building a storage networking switch that will leverage commodity disks at performance and reliability rates that are equal to if not better than current storage arrays. So that is the whole magic. You have to get equal or better performance and reliability. The whole idea of using flash is really interesting, but that’s essentially using the same architecture. What I’m sort of trying to leapfrog in the storage space, is to ask how storage can become commoditized on the back of commodity components.

You’re not a believer in the idea that flash can save the world, at least inside the data center?

Flash is expensive, but it’s still a storage array that sits there with a lot of expensive stuff around it. The flash stuff is linear and sequential versus the way storage array has tended to work in the past. My belief is that fundamentally there won’t be a storage array anymore. Now, that’s a big leap. But in the same way, things will change up where software will just define the storage layer. I’m talking commodity flash and disk as being the basis for the infrastructure.

In the midst of globally difficult economic times, the Japanese saw an opening. Taiichi Ohno, working for Toyota after World War II, saw inefficiencies in the American assembly line system, especially in the inventory and quality control systems. He pioneered a new system of production for the company, called “just-in-time.” The objectives were to introduce consistency in the production process and eliminate waste of all kinds, including overproduction, excess inventory, time (waiting for parts) and ineffective or defective products being produced. American supermarkets, ironically, held the answer for Ohno; they only restocked shelves with enough product to replace inventory that was sold to customers. In the just-in-time system, this translated into keeping about 24 hours’ worth of inventory in the factory, with parts being ordered and shipped on an as-needed basis.

The success of Toyota’s transformation is now well-known; it became a global company that would eventually surpass sales of GM, the world’s largest auto maker, in 2008. The just-in-time system pioneered by Toyota would be widely copied, with U.S. automakers implementing the new system in the 1980s, to varying levels of success.

Lessons for the Data Center

You may be asking what all of this fascinating automotive history has to do with data centers today, in 2013. The fact is, the data center industry is undergoing a transformation of its own. While the U.S., thankfully, hasn’t seen oil shocks on the scale of those experienced in the 1970s, it’s clear that the era of cheap energy is over, especially with increasing demand for energy supply in high-growth markets like China and India.

Rising energy prices and volatility mean that data centers can’t afford to operate the way that U.S. auto manufacturers did in the 1950s and 1960s. Traditionally built to accommodate future growth, wastefully inefficient data centers were built in previous decades with overcapacity in mind, in an era when energy was much cheaper and more readily available. Today, data center owners and managers face the challenges of legacy builds, with these facilities suffering from the same kind of waste that plagued the auto industry before lean manufacturing changed it. As James Glanz outlined last year in the New York Times story “Power, Pollution and the Internet,” the worst offenders not only use huge amounts of energy to power parts of data centers that aren’t used, but also contribute to pollution.

Just as gains in the speed of transportation made just-in-time a reality for the automotive industry in the 1970s, the data center industry’s transformation may be driven today by recent technology advances and supplemented with an overall recognition in the industry that the old way of powering and cooling data centers will no longer be ignored by the public nor economically advantageous for companies. These recent advances include:

• The rise of readily available modular data center components, including Uninterruptible Power Supplies (UPSs), cooling and power distribution systems. These components are the infrastructure that can efficiently and reliably power and cool expensive computing equipment. Prior to the availability of modular components, data centers were almost always custom-built, and because of this, they were almost always overprovisioned to accommodate future growth or spikes in demand.
• The rise of standardized components in the data center, especially related to data center infrastructure. Part of the success of just-in-time was a recognition that relying on trusted suppliers to manufacture parts, so that the automaker could focus on its core competency of manufacturing the automobile, would actually bring about more reliable and better-performing products. Data center infrastructure is much the same — custom-built systems come with custom problems that take man-hours and money to fix. Standardized components are time-tested to be effective, reliable and predictable, with flaws having been driven out in the manufacturing and testing/validation process.
• The availability of monitoring software that provides visibility into energy usage in the data center. The just-in-time system relies upon signals that would tell managers what the rate of demand for parts was, so that production could be scheduled accordingly. Similarly, advances in data center software and analytics today provide that kind of intelligence for data center managers by monitoring equipment usage and helping to plan for future needs through benchmarking, indicating in advance when capacity may need to be increased.

Standardization is the key to consistency in the data center, while modularity is the key to flexibility. Both characteristics are needed to respond effectively to changes in the IT environment that result from changing business needs. Modular components are pre-engineered and “plug and play,” so that they can evolve with a changing data center’s design over time. Because they are scalable, changeable, portable and swappable, they confer a speed of installation that is akin to the flexibility that most car owners get today when they need to replace or repair auto parts.

Benefits of a Just-in-Time Approach

When Toyota pioneered the just-in-time system, it also built an entire system that the company would adhere to outside of production that would contribute to its success. Part of the system was to cultivate a culture of continuous improvement where its people’s talents would be stimulated and performance would be maximized. To be sure, recent events such as the tsunami in 2011 affected global supply chains and devastated production and supply in Japan, which exposed vulnerabilities in the system. However, the success of the overall system over time, as evidenced by its proliferation into all types of manufacturing as well as other markets including business intelligence in recent years, serves as a testament to the substantial benefits that can be conferred through applying the overall principles of the system.

In the data center, this requires a new approach to thinking about designing and building new systems, one which promises substantial gains in operational efficiencies. Some of these include:

• Outsourcing of non-core competencies. In the auto industry, this included logistical functions such as storage and distribution to third party providers. Modular data centers, effectively, have outsourced the headaches of production, configuration and installation to the manufacturer of the components. They are plug-and-play, making initial deployment and subsequent changes achievable without a great deal of custom work, and they are built in factories under strict quality control and engineered to deliver peak performance.
• Reduced response time to business demands, again due to availability of standardized and modular components. Growth can be planned for in the initial design of the data center, without being deployed and operationalized, which means data centers are “right sized” — operating only with the computing power and infrastructure to fit needs today rather than anticipated needs in the future.
• Improved customer satisfaction and better return on equity, due to the reliability of components as well as energy savings and reduced manual labor. Custom-built data centers can take years to achieve designed performance specifications with respect to measurements such as Power Usage Effectiveness (PUE, a measure of data center efficiency). Prefabricated modules come with design specifications that are verified in the factory before they are shipped.
• Less downtime, as failures in modular components can be identified quickly and easily and replaced in little time, compared to disparate or custom engineered components.

Data centers today, in fact, are already taking advantage of modular data center builds. The Beneficial Financial Group, one of the oldest insurance companies in Utah, needed to add more power and rack space in its data center, with better monitoring and management capabilities and minimal downtime. If it bought a new system or a larger version of the existing system, it would have had to re-engineer the data center completely; instead, Beneficial chose a modular system that integrated power, cooling management and servers with a universal rack design. The installation was completed over a weekend, with no interruption to business, and forecasts estimate the company will gain an annual ROI of 74 percent with a payback period of ten months, due to maintenance and service savings and increased user productivity. In addition, because it can now add servers as the company grows, Beneficial has expanded its data center capability by 300 percent with room for future growth.

Toyota, which spearheaded the just-in-time model, was able to achieve unparalleled growth and eventual industry dominance in the years following its shift. It’s an example of how turning traditional thinking on its head was able to change the entire industry and spawn whole new industries and innovations. Data centers today face a fork in the road, but there is a unique opportunity to take advantage of gains in technology to respond and adapt, making sure they are built not only for the needs of today but for the unforeseen challenges that may come in the next few decades.

Kevin Brown is Vice President, Global Data Center Offer for Schneider Electric and is responsible for leading Schneider Electric’s portfolio strategy and leading innovation to respond to emerging data center industry trends. Kevin is an experienced industry professional in both the data center and HVAC industry with over 20 years of experience in various senior management roles including software and hardware development, sales and product management, and holds a BS in mechanical engineering from Cornell University.

Using a combination of shares he already owns and personal cash, Dell will, according to most reports on the deal, nudge his ownership stake north of 50 percent. Silver Lake Management and Microsoft are also said to providing some of the financing, though the role of the software giant is, as The Wall Street Journal reported, still under discussion. Reuters reports that the deal could be announced as early as Monday.

Dell has been trying mightily to transform itself from the PC maker that no one could compete with in the 1990s to an IT hardware, software and services company. Shareholders have not been forgiving, and, in their defense, as analyst Shaw Wu recently observed in a research note to clients, Dell has spent about $13 billion on enterprise-related acquisitions since 2008, a recent highlight being the $2.4 billion deal for Quest Software; still, about 70 percent of its revenue remains tied to PCs sold both to consumers and to businesses plus ancillary products that are closely tied to PCs (printers, monitors and so on). That’s down from a much higher percentage early last decade, but turning the corner has been slow, and progress has come in fits and starts.

Meanwhile, the PC market has been in a state of decline, leaving Dell in a rush to replace slowing revenue growth in that space with other things. While the idea to transform Dell into an enterprise-focused company has merit, the PC business has been contracting so quickly that the changes aren’t having the desired effect on the shares.

Michael Dell talked about the strategy in an interview with AllThingsD in July. In that conversation, I asked him whether he thought shareholders tended not to give the company credit for the transformation strategy, and thus valued the company unfairly. His answer:

“Some shareholders get it and some of them don’t. That’s how markets work. Most who have looked at it carefully say this is the right thing to be doing. They quibble about some of the smaller points, but they see the larger logic of what we’re doing and our job is to keep doing it.”

Now it looks like he’s going to do it without the bother of having to report to shareholders.

Having risen more than 3 percent to $20.09 on the New York Stock Exchange today, Fusion’s shares fell by more than 17 percent to $16.79 in after-hours trading after the company said it expects slower revenue growth as major customers Facebook and Apple slowed orders. CFO Dennis Wolf claimed the timing of orders from those two companies slipped by two quarters. Together Facebook and Apple account for 51 percent of Fusion’s business.

I just got off the phone with CEO David Flynn, who told me that having two customers account for so much of its business was always a risk. But? When Fusion first IPO’d in 2011, Apple and Facebook accounted for 70 percent of sales, so it could have been worse. “We knew there was going to be some risk with customer concentration,” he said. “Over time the concentration will work itself out,” Flynn told me.

As yet, no other single customer accounts for more than 10 percent of revenue, and therefore requires disclosure. But five of the 10 customers that bought more than $1 million in Fusion-io products in the last quarter also bought its newer product ioScale. Also, sales in Europe grew 100 percent year on year, while sales in the Asia Pacific region grew 130 percent.

Fusion-io reported per-share earnings of 2 cents in the quarter, versus 7 cents a year ago. Revenue rose more than 40 percent to $120.6 million. The results fell short of analyst expectations, which called for earnings of 8 cents on $120.3 million.

Facebook has been a Fusion-io customer for a long time, along with a handful of other companies, including Apple, that use its flash-based technology to speed up the individual machines in their data centers.

Today, Fusion announced that its latest product, Fusion ioScale, which has been available to existing customers like Facebook for a while, is now generally available to new customers as well. The implications for data centers aren’t trivial. I talked with CEO David Flynn about this last week and he summed it up to me simply: Data centers are going all flash. Hard drives are on their way out. Get used to it.

Previously, Fusion’s play has been in enhancing system memory — DRAM chips, which are fast but only hold on to as much data as a computer’s processor needs to get immediate work done — with flash memory chips, which can store data longer, even after the system they’re in has been powered down. The point has been to speed up the amount of work a processor can get done by eliminating those long pauses — long for a computer, anyway, because they’re measured in nanoseconds — when a computer’s processor is sitting around doing nothing but waiting for the next batch of computing work.

Now Fusion is starting to bring flash into the data center’s storage tier. With ioScale, the idea is to bring as much as 3.2 terabytes of storage that’s tuned for the industrial-strength machines in data centers; if you use four in one machine, you boost that to 12.8 terabytes.

Eventually, Flynn told me, hard drives will be pushed out of data centers entirely, even for long-term archival storage purposes. “We are jointly planning with Facebook for when that becomes possible, and we think it will happen within a couple of years,” he said.

Flynn, who only half in jest refers to hard drives as “spinning rust,” says that once data centers convert to all flash, their energy requirements become a little looser. “If you get rid of all your spinning rust, you do not have to control the humidity and temperature of your data center as precisely as you have to now. You can do more open-air cooling, and cooling is a single point of failure,” he said. “These data centers consume as much power to condition the air as they do to power the systems themselves, and that makes the data center itself more than twice as expensive to operate.” Thus they’ll be able to run warmer, and less expensively.

The disclosure came at the Open Compute Summit in Santa Clara, where Facebook was expected to comment on the topic as well in a series of announcements that also involve processor giant Intel. The chip giant just revealed that it is collaborating with Facebook on the design of a new “disaggregated, rack-scale server architecture” that takes advantage of some new Intel technologies, including silicon photonics.

Fusion-io shares are up more than 2 percent today on word of the new product. The shares fell 4 percent during 2012.

The auto maker is immediately seeking software developers, project managers, database experts, business analysts and other computer professionals to work in what will be its third of four information technology centers across the U.S. Its Atlanta Innovation Center will be located in the northern suburb of Roswell and concentrate on research and development rather than everyday processing.

Read the rest of this post on the original site »

In response, the once-lethargic $20+ billion enterprise storage industry is exhibiting unprecedented innovation. Giants like EMC and Dell are vying with, partnering with and acquiring start-ups for supremacy in a morphing landscape.

It’s a serious game. More than $3.5 billion was pumped into VC-backed storage start-ups between 2007 and 2011, with more than $1 billion in 2011 alone. And $10 billion more has been poured into M&A, with the most recent example being EMC’s $430 million purchase of a company that hasn’t finished developing an initial product. Tectonic shifts come from collisions of forces. There are three major force vectors here.

First is the demand for more scalable, instantly provisionable, faster and higher-capacity storage. Even in an anemic economy, demand for data storage grows more than 50 percent per year.

Storage in a Flash

The second driver is the widespread proliferation of flash memory. Remember when you could feel the hard drive spinning inside your iPod? Today, practically every consumer carries flash memory in his or her pocket or purse.

Flash has been around for years, but was too expensive for broad adoption. Thanks to companies like Apple, which consume enormous amounts of flash, the cost is dropping like an apple from a tree. It’s still much more expensive than rotating hard drives, but its notable physics are compelling. Solid-state flash is faster than mechanical drives, and doesn’t forget everything when the power is turned off. Perfect, right?

Flash in the Pan?

Actually, not so perfect for corporate, governmental or cloud environments. In addition to high cost, flash has some unfortunate features. For example, it wears out in the same way your favorite pair of jeans will become threadbare with use. You’ll never wear out your phone from too much texting. But in an enterprise application, the number of IOPS can be so staggering that flash has to be treated almost like a printer cartridge, a consumable.

Many companies are developing techniques to deal with flash’s inherent limitations to make it suitable for data centers. It’s a gold rush, with vast sums of capital chasing big markets. That $430 million acquisition by EMC? Yup, flash.

Why pay so much for a pre-product company? There were multiple bidders. NetApp made a rich offer, which Dell topped by a lot, which EMC topped by an equally wide margin. There will be more M&A.

Interestingly, this activity is driving breakneck commoditization. This is great for customers, but not for vendors, who will not long enjoy rapacious (oops, I meant healthy) margins on proprietary technology. Ironically, the value in flash-based systems is really in the software that wrests the value from the hardware. Everyone in the industry knows that the days of differentiated flash hardware are numbered.

When Is Storage Not Real? When It’s Virtual.

As if hot, high demand and cool flash aren’t enough, the storage games are impacted by a third force called virtualization. Virtualization has transformed computing. The leading vendor, VMware (not coincidentally, owned by storage company EMC), has built a market capitalization of roughly $40 billion. All by making fake computers.

We call them virtual machines. Your iPhone may be talking to one right now over the Internet. Their magic allows the creation of what looks like a physical computer server. A virtual machine, or VM, appears to embody central processing units, memory and communication networks like physical computers. But it’s a software abstraction. Through this prestidigitation, data centers run scores of VMs on a single server box.

Enterprises can deploy vastly more applications because virtualization from Microsoft, Citrix, Red Hat and VMware saves enormously on capital and operating expenses. And provisioning is so much faster. Just a few clicks and, voila, you have a new server. By the way, you can buy server hardware from anyone. That freedom makes vendors compete harder, which you like if you run an information technology department.

What does this have to do with data storage?

Enterprise and cloud storage still live in the physical age. That is, storage system features are embedded in proprietary hardware. Want a cool software feature? You have to buy hardware to get it. Before virtualization, this was how the server industry worked. But virtualization is stressing traditional modes of delivering storage to applications. Performance problems, high costs and inflexibility cause VM users great pain on a daily basis.

Hello, Storage Hypervisor

The key enabling technology in compute virtualization is called a hypervisor. This core magic remade the server industry for the benefit of all. Until recently, there was no storage equivalent.

Now the storage industry is beginning to buzz about the concept of a storage hypervisor — the analog of the server hypervisor, but for storage. Storage hypervisors promise to increase the effective performance of hardware by an order of magnitude. By virtualizing resources to provide the administrative paradigm needed in virtualized environments — VM-centric management — they provide unprecedented flexibility and efficiency. Naturally, the ideal storage hypervisor leverages flash, just as server hypervisors unleash the power of Intel-based silicon.

Giant publicly-traded storage vendors and ambitious start-ups alike are talking up their offerings. All have differing approaches, but share the goal of giving data-center storage buyers the benefits already bestowed on server customers.

A Serious Game

Today we observe a convergence of forces transforming a multi-billion dollar market. The unending pressure for more data increases demand for high-performance flash-based storage hardware. This in turn is driving the essential requirement for virtualization of storage hardware resources. This confluence will enable vastly larger amounts of storage to be applied to every imaginable use case, all while making the economics not only affordable, but also compelling.

The winners in this game? Corporate and cloud data centers and their users. In other words, you.

Before co-founding Virsto, Mark Davis was CEO of storage resource management vendor Creekpath, where he engineered its acquisition by Opsware (now HP).

Let’s start with the wireless networks. Anecdotally, I’ve been hearing about all sorts of difficulties with voice calling today. For the first half of the day, my Verizon iPhone didn’t display the 3G icon at the top of the screen, and the few voice calls I tried to make took longer to connect than usual if they connected at all, which about half of them didn’t. I also heard that people calling me were hearing “fast busy” signals indicating an overloaded network.

That’s not surprising because the storm zone is heavy Verizon territory. In a statement to Reuters, Verizon said it was still trying to get a full picture of the damage to its network. Sprint and T-Mobile are pretty much in the same boat, and customers of all three are seeing service disruptions. So for the time being my advice about texting and not calling remains in force.

Then there were the land lines and Internet services. Again, this is mostly Verizon territory. The company said in a statement that several of its central offices in Manhattan, Queens and Long Island had been hit with flooding and had also lost power. Normally there’s backup power, but generators don’t exactly respond well to water. So they’ve been running on battery power. The problem is when the batteries die, the equipment inside the hubs gets damaged and so Verizon has had to power it all down, thus hurting its capacity for service. This applies to things like FiOS, its DSL Internet services and old-school telephone service. A fellow named Jack is doing yeoman’s work on Verizon’s customer support Twitter account trying to help customers out.

On Long Island, where most of the people rely on Cablevision for their Internet, about 90 percent of residents have lost power. That’s affecting access to the Internet and phones on Cablevision’s Optimum Online service. Time-Warner Cable is suffering outages in parts of New York City and New Jersey, where people have also lost power. The company says that as power comes back, they’ll identify people whose service has been disrupted.

Personally I can report that RCN, the regional cable company, managed to keep its TV service running, but Internet service was disrupted for about 12 hours in Manhattan. Some lost TV service too, despite the power staying on in most of Upper Manhattan.

And the woes weren’t limited to consumers and their cable modems. This may come as a surprise, but data centers don’t like water very much, and lower Manhattan flooded pretty badly. One data center, operated by a hosting outfit called Datagram, whose most high-profile customer appears to be Gawker and its network of sites, saw its lower Manhattan facility flooded. Normally backup power would have kicked in but according to a terse statement, the basement where the generator’s fuel tanks are kept was totally inundated.

Here’s what Datagram said in full:

Official Statement

As of 5pm on October 29, 2012, Datagram had thoroughly tested its emergency systems at 33 Whitehall, NYC and was fully staffed awaiting the storm to hit Manhattan’s shores. Once ConEd lost power to Lower Manhattan, Datagram’s emergency systems kicked on maintaining power to Datagram’s datacenter. Unfortunately, within a couple hours of the storm hitting Manhattan’s shores, the building’s entire basement, which houses the building’s fuel tank pumps and sub pumps, was inundated with water taking the building generator system offline – essentially shutting down the entire building. No Datagram or customer infrastructure has been damaged by the storm.

We have been working intimately with the building’s engineers to clear the water from the building’s basement so that we may restore our emergency power systems. We are providing regular updates through our website at http://www.datagram.com.

A similar problem hit a huge co-location facility at 75 Broad Street. As Data Center Knowledge noted, flooding there destroyed critical fuel pumps needed to keep backup generators online, disrupting service to customers of hosting providers like Internap and Peer 1. An emergency letter posted as an image to Twitter warned customers that they should move their workloads elsewhere. At the time it was posted, the facility had about five to seven hours worth of fuel to keep the lights on and servers humming. It subsequently lost power before noon local time.

Internap Customers LGA11 Facility at 75 Broad Street (Downtown NYC) to shut down within hrs http://t.co/LD6taNth

October 29, 2012 8:37 pm via webReplyRetweetFavorite

@jose3030

jose3030

They weren’t alone. Renesys, a research firm that tracks the health of the Internet, posted on its corporate blog a graph showing a surge in network outages in New York and the surrounding area. (I borrowed the image from Renesys. Click to make it bigger.)

There’s also some interesting intelligence from a new site called Outages.org that has a really detailed report on conditions at 111 Eighth Ave., a.k.a. the site of Google’s New York operations.

It isn’t just a Google facility, but it is one of the most heavily connected buildings in New York. The detailed report on Outages.org lists lots of companies affected by flooding, including 8×8, the Internet phone service Voxel, cloud services and hosting company Equinix and XO Communications.

I’m keeping an eye on things and will report back on more situations as they develop. If you know of any other service disruptions involving data centers or any other service providers, drop me a line.

But when you’re a company with a lot of industrial-grade Internet infrastructure, including, say, a few of your own data centers, the rules can change a little bit, and you have the option of being a little more selective in determining how your data flows. And when you’re the Enterprise cloud and file-sharing and collaboration outfit Box, you turn that advantage into something your far-flung customers can take advantage of.

When file are big — and in business, they always are — and you need to share something, uploads can be a time-consuming pain in the neck. So Box today launched a network of what it calls Box Accelerators. A network of servers distributed around the world, they serve as outposts for Box’s primary data centers. Box customers around the world will be able to upload to these Accelerators, and thus speed things up, say CEO Aaron Levie. In some cases, Box is bringing to bear its relationship with Amazon Web Services, and that company’s global footprint.

“We’re making a big push on international expansion,” Levie told me in a conversation at Box’s Los Altos, Calif., headquarters last week. Over the summer, the company opened a new office in London, and announced plans to hire 100 people there, in order to double down on opportunities it sees in Europe.

Media companies, health-care companies, and companies and institutions engaged in scientific research often have to move large, cumbersome files around in order to make them available for collaboration with colleagues. The primary factor in slowing down that process is distance. For all the vaunted rhetoric of how the Internet makes the world a smaller place, when it comes to moving gigabytes or more at a time, the one thing standing in your way is the distance between you and the server you’re uploading to.

The network of Accelerators are intended to shorten that distance. Box customers will get a choice of servers closest to them, and those servers in turn will have an easier time of communicating with Box’s main servers at its network of data centers, including a newish-one in Las Vegas, and another two in California.

The service is going live in nine different regions around the world on every continent except Africa, and is available free of charge for existing Box customers. And while today the network is specified only for uploads, it will in time enhance the speed of downloads as well, Levie told me. And it will also be addressable by Box’s API, meaning that if you’re building an application that takes advantage of Box’s network, the accelerators will be available for use.

I took advantage of a few minutes with Levie to ask him about his reaction to last week’s disclosure by Salesforce.com CEO Marc Benioff at the TechCrunch Disrupt conference that he’s close to announcing a service called Chatterbox that will essentially compete with Box. Benioff will likely talk in detail about Chatterbox in his keynote address at Salesforce’s Dreamforce conference on Wednesday. Expect collaboration and file-sharing to become part of Salesforce’s Chatter social platform soon.

Levie said he’s known about Salesforce’s intentions in this area for about four months to six months. “We’ve known about it, frankly they kind of have to do it,” he said. “If you think about Chatter, and Jive, Yammer, that kind of social world, it necessarily has to connect up to your content. And Salesforce is trying to carve out the space that Chatter can play in. It’s not enough to be a standalone social platform. Salesforce has basically realized that Chatter has to be able to solve more problems for the enterprise before they can have a serious conversation with a CIO.”

Content that companies share both internally and with partners, vendors, suppliers and customers has to be enhanced with social collaboration features. This is the very essence of companies like Jive and Yammer, and Salesforce’s product in this area known as Chatter. “Salesforce will take what it has as a social platform and add content to it,” Levie said. “What we’re doing is more like the inverse. We have a content platform, and we bring social aspects to it, where relevant. We work with all of the social platforms out there, and we’re going to be building more collaborative capabilities into Box.”

You might think this would be a tad awkward, given the fact that Salesforce is an investor in Box, and is even said to have offered to acquire Box last year, for north of $500 million.

Not at all, Levie says. “There are lots of precedents for companies being both investors, partners and competitors.” To me, it sounds like a diplomatic way of saying “game on.”

Shares of Salesforce are falling and falling big in after-hours trading as the cloud-software concern reported quarterly results that beat all the forecasts of Wall Street analysts, but which included an outlook for the quarter and year ahead that fell short of expectations.

Soon after markets closed, Salesforce shares took a 5 percent haircut, falling by $8.31 to $138.46. That drop comes on top of a drop of more than 1 percent during the regular session, when the shares finished at $146.77.

Earnings for the quarter were 42 cents per share, beating the consensus of 39 cents. Revenue was $732 million versus the consensus of $728.3 million. So far so good.

Then came the outlook: Salesforce said it expects revenue to come in at between $773 million and $777 million, soundly beating the expectation of $771 million and change. Still good.

The problem was with the EPS expectation, which one must remember is a non-GAAP profit to begin with. Salesforce says it will be 31 cents to 32 cents a share, versus a consensus of 34 cents. (On a GAAP basis, which no one pays attention to, Salesforce still expects to report losses in the range of 26 cents to 27 cents, and of 72 cents to 75 cents for the full year.)

Okay, everyone freak out. More as I go through the numbers.

Here’s the Salesforce announcement.

Salesforce.com Announces Fiscal 2013 Second Quarter Results
- Quarterly Revenue of $732 Million, up 34% Year-Over-Year
- Quarterly Operating Cash Flow of $136 Million, up 64% Year-Over-Year
- Deferred Revenue of $1.34 Billion, up 43% Year-Over-Year
- Unbilled Deferred Revenue Increases to Approximately $2.8 Billion
- Raises FY13 Revenue Guidance to $3.025 – $3.035 Billion

SAN FRANCISCO, Aug. 23, 2012 /PRNewswire/ — Salesforce.com (CRM), the enterprise cloud computing (http://www.salesforce.com/cloudcomputing/) company, today announced results for its fiscal second quarter ended July 31, 2012.

“Our second quarter revenue growth was outstanding at 34% in dollars and 37% in constant currency,” said Marc Benioff, Chairman and CEO, salesforce.com. “Salesforce.com’s social enterprise strategy is enabling companies to connect with customers, partners, and employees in completely new ways – and it’s creating new opportunities for their growth and ours.”

Salesforce.com delivered the following results for its fiscal second quarter:

Revenue: Total Q2 revenue was $732 million, an increase of 34% on a year-over-year basis. Subscription and support revenues were $687 million, an increase of 35% on a year-over-year basis. Professional services and other revenues were $44 million, an increase of 20% on a year-over-year basis.

Earnings per Share: Q2 GAAP net loss per share was ($0.07), and non-GAAP diluted earnings per share was $0.42. The company’s non-GAAP results exclude the effects of $85 million in stock-based compensation expense, $20 million in amortization of purchased intangibles, and $6 million in net non-cash interest expense related to the company’s convertible senior notes. Non-GAAP EPS calculations are based on approximately 146 million diluted shares outstanding during the quarter, including approximately 3 million shares associated with the company’s convertible senior notes. GAAP EPS calculations are based on a basic share count of approximately 139 million shares.

Cash: Cash generated from operations for the fiscal second quarter was $136 million, an increase of 64% on a year-over-year basis. Total cash, cash equivalents and marketable securities finished the quarter at $1.8 billion.

Deferred Revenue: Deferred revenue on the balance sheet as of July 31, 2012 was $1.34 billion, an increase of 43% on a year-over-year basis. Current deferred revenue increased by 38% year-over-year to $1.27 billion, benefited in part by longer invoice durations. Non-current deferred revenue increased by 293% year-over-year to $69 million. Unbilled deferred revenue, representing business that is contracted but unbilled and off balance sheet, ended the second quarter at approximately $2.8 billion, up from approximately $2.7 billion at the end of the fiscal first quarter.

As of August 23, 2012, salesforce.com is initiating revenue, GAAP EPS and non-GAAP EPS guidance for its fiscal third quarter of fiscal year 2013. In addition, for the full fiscal year 2013, the company is raising its revenue and non-GAAP EPS guidance previously provided on June 4, 2012, and initiating GAAP EPS guidance.

Q3 FY13 Guidance: Revenue for the company’s third fiscal quarter is projected to be in the range of $773 million to $777 million, an increase of 32% to 33% year-over-year.

GAAP net loss per share is expected to be in the range of ($0.27) to ($0.26), while diluted non-GAAP EPS is expected to be in the range of $0.31 to $0.32. The non-GAAP estimate excludes the effects of stock-based compensation expense, expected to be approximately $99 million, amortization of purchased intangibles related to acquisitions, expected to be approximately $27 million, and net non-cash interest expense related to the convertible senior notes, expected to be approximately $6 million. EPS estimates assume a GAAP tax rate of approximately 37%, and a non-GAAP tax rate of approximately 35%. The GAAP EPS calculation assumes an average basic share count of approximately 142 million shares, and the non-GAAP EPS calculation assumes an average fully diluted share count of approximately 151 million shares.

Full Year FY13 Guidance: Revenue for the company’s full fiscal year 2013 is projected to be in the range of $3.025 billion to $3.035 billion, an increase of 33% to 34% year-over-year.

For the company’s full fiscal year 2013, GAAP net loss per share is expected to be in the range of ($0.75) to ($0.72) while diluted non-GAAP EPS is expected to be in the range of $1.48 to $1.51. The non-GAAP estimate excludes the effects of stock-based compensation expense, expected to be approximately $382 million, amortization of purchased intangibles related to acquisitions, expected to be approximately $95 million, and net non-cash interest expense related to the convertible senior notes, expected to be approximately $24 million. EPS estimates assume a GAAP tax rate of approximately 30%, and a non-GAAP tax rate of approximately 37%. The GAAP EPS calculation assumes an average basic share count of approximately 141 million shares, and the non-GAAP EPS calculation assumes an average fully diluted share count of approximately 150 million shares.

It has been only two years since the enterprise storage wars. In fact, it was two years ago this month that Dell and Hewlett-Packard bitterly battled over 3Par. HP won that fight, paying $2.4 billion for the storage concern. Dell bought Compellent instead, while IBM took out Netezza. All told, it was a $5 billion M&A bonanza that ended just as suddenly as it started.

By joining the loose collection of companies I’ve dubbed the Flash Madness Club — companies that have tossed their hats in the ring, either selling or relying on flash memory in a big way — IBM, a lot of people are speculating, will force rivals like HP, Dell and maybe Oracle to start rolling up the other flash-technology players. Nor is IBM done buying in this area.

The main target of all this is Fusion-io, the Utah-based flash-technology company that first went public last year, and which is remembered mostly for its data-center supply relationships with Facebook and Apple.

Fusion’s primary product is the ioDrive, an insert card that essentially speeds up conventional servers by feeding data to the main processor faster than the relatively poky hard drive. Fusion shares rose 7 percent on the speculation, closing yesterday at $28.23.

At that price, it trades at a market cap of about $2.6 billion. Assuming a 50 percent premium, it would, in a hypothetical deal, probably go for about $4 billion, assuming there wasn’t a crazy bidding war.

But there would be. Fusion’s primary strength is its OEM relationships. (OEM is industry lingo for Original Equipment Manufacturer.) In these relationships, Fusion sells its products like the ioDrive to such companies as HP, Dell and IBM, which then offer them as part of their own distinct servers.

The product is considered strategic enough that if one big IT player were to try and buy it, the others would put up a fight and offer competing bids to prevent Fusion from falling into the hands of a competitor. Fusion would become a much more expensive target rather quickly.

But, at the same time, the winning bidder would own an asset that would instantly lose value. If, for example, HP closed a Fusion-io acquisition, can you realistically see Dell and IBM continuing to do business with it? Future growth of Fusion’s products would have to offset that lost revenue. And given the concentration, the loss of one partner would be a big blow: 82 percent of Fusion’s revenue in its most recent quarter was made up of individual companies, each accounting for 10 percent of sales or more. A target that would get progressively more expensive in a bidding war and then lose value right away doesn’t look like a good deal.

Longer-term, there’s a lot of value in Fusion’s relationship with the end customers. Even when an OEM sells a server with Fusion’s technology inside it, it is usually Fusion and not the OEM that supports the product with software upgrades and maintenance. As more companies begin running hardware with Fusion’s technology inside them — banks and financial institutions are big fans of it, for one thing — owning Fusion would make sense for one of the big IT companies.

But here’s another thought: If Fusion is to be acquired, the thinking goes that it would have to acquired by either a traditional IT company or a dedicated storage player. HP and EMC are the ones being mentioned as possible buyers in the analyst notes today. But why not Intel?

The world’s biggest chipmaker has the cash — $13.7 billion, as of its most recent quarter — and is constantly looking for ways to expand its relationships with the hardware vendors. And since Intel already does business with all the OEMs that Fusion does, none would lose access to its technology. There are probably a lot more things to consider, but I just don’t see why Intel is not part of this conversation.

Sales were $106.6 million, and per-share earnings were 9 cents, easily besting the consensus of $96 million and 3 cents. Great, but that’s not what got investors so excited that they kicked Fusion shares up by 26 percent in the after-hours session. The company said it expects sales in its fiscal year 2013 to grow 45 percent to 50 percent, well ahead of the 37 percent analysts had been expecting.

At least part of that outlook stems from things like the super-secret deal Fusion did with enterprise storage concern NetApp last week, and also the trend of servers going all flash, thanks in part to Fusion’s software.

“We got new products and new opportunities within existing markets, so we feel pretty good,” CEO David Flynn told me in a brief conference call after Fusion reported earnings today. All the new stuff going on is offsetting what has been the traditional criticism against Fusion since its IPO last summer: That it relies too heavily on large purchase orders from a small number of customers, specifically Apple and Facebook, who buy a lot of Fusion’s technology for use in their data centers.

“They’re representing a smaller overall portion of our business as these new opportunities develop,” Flynn told me.

One other thing he said: Some of those big customers have some pretty aggressive plans to scale out their data centers. “These relationships have afforded us some visibility on their plans, and we’ve factored that into our guidance with high confidence.” It looks to me like Fusion-io’s guidance is quickly becoming a pretty good barometer for the overall state of the data center business.

There’s been a significant uptick in M&A and VC funding activity around this idea lately. For months I had been covering the plans of the cool start-up Nicira to try and mess up the best-laid business plans of existing networking players like Cisco Systems and Juniper Networks. Then, just as suddenly as it burst onto the scene, VMware, the company that basically defined virtualization inside servers and other computers, stepped up and acquired it for $1.3 billion.

Then a week later, software giant Oracle took out Xsigo Systems, another virtual networking concern, though we don’t know how much it paid.

Now we have a new player to talk about. It’s called Plumgrid, and today it will announce that it is taking a $10.7 million Series A round of venture capital funding from US Venture Partners and Hummer Winblad Venture Partners. Chris Rust of USVP just joined its board of directors, and Lars Leckie, a Hummer Winblad partner, is joining the board, too.

But Plumgrid is not coming out of stealth mode, at least not just yet.

I talked for awhile with Awais Nemat, (pictured) who has assembled a team of engineers from companies like Cisco, chipmaker Marvell, Sun Microsystems, VMware and even Nicira, all of them people who have contributed a lot to the concept of network virtualization and software-defined networks.

He wouldn’t tell me much about what the company is working on productwise — that’s still the stealth part. But he did tell me a bit about the kinds of problems it is aiming to solve. “There is a monumental shift coming in networking infrastructure … Networking functions that used to require a dedicated box we do in software running on standard software, running on standard silicon,” he told me.

Different kinds of companies have different kinds of networking needs, but today more often than not, all companies pick from the same set of hardware products that Cisco and Juniper and Hewlett-Packard and other networking companies offer; the primary difference has to do with whether one or the other is big enough or too big.

“We see a new kind of customer emerging, one whose demands and needs are not being satisfied by the traditional network infrastructure vendors,” Nemat said. To that end, Plumgrid is working with some early customers to define exactly what it is going to offer, essentially by defining the problems those customers have with what they can buy off the shelf today. “Empathy is turning out to be a pretty good way to design a product,” he told me.

Nemat himself is an old Cisco hand who struck out on his own to run D5 Networks, a company that designed chips for networking security. It was acquired by Marvell in 2006. He stayed on there until last year when he founded Plumgrid and raised $2 million and change in angel funding. It’s nice to have another secretive networking start-up to watch.

But the questions stemming from the dispute don’t stop with the decision by Judge James Kleinberg ruling that Oracle must abide by the terms of promises it made to HP in 2010 when it settled an unrelated lawsuit concerning Oracle’s hiring of former HP CEO Mark Hurd.

First off for HP is the question of whether the damage to its Business Critical Server business unit, which sells the Integrity line of servers that use the chips, can be reversed. The damage is plain to see in HP’s financial results. For the first six months of the year, sales are off by $275 million, in a business unit that last year saw sales north of $1.1 billion. The uncertainty brought about by the lawsuit has hurt sales, and HP has made it clear it expects them to remain under pressure.

But it’s not lost sales that are hurting HP the most: A half-billion-dollar drop in sales at a company on track do $123 billion this year isn’t much to get excited about. It’s the profits. Legal filings made public over the course of the suit showed that HP derives a healthy portion of its profits from ongoing service and support contracts with companies that buy its Integrity servers. While HP doesn’t routinely break these numbers out in regulatory filings, documents showed that in 2010, HP derived about 15 percent of its profits on an EBIT basis (Earnings Before Interest and Taxes) on business related to Itanium.

Reversing the trend will be hard. For one thing, the legal fight isn’t over. Oracle has promised both to appeal the decision and to continue to press its counter-claims against HP in court, so the uncertainty among HP customers will continue, though as HP’s enterprise chief Dave Donatelli put it to me in an interview in June, the first step toward saving the BCS business is winning the lawsuit.

HP does have a plan to move Itanium customers onto more mainstream servers. It’s called Odyssey, and it involves building a new generation of Business Critical Servers on a more mainstream platform, probably Intel’s Xeon. Over time — and it would take several years — Integrity customers could be persuaded to move in this direction. Exactly how HP preserves the highly profitable service and sales contracts upon which it has relied all these years isn’t entirely clear. One key piece of the strategy would likely involve the creation by HP of a version of its Unix operating system, called HP-UX, that runs on Intel’s mainstream x86 chips.

For Oracle’s part, while its appeal is pending it will have to rejigger its plans and issue an update to its database software for Itanium systems. Existing customers had nothing to worry about in the first place. But it now faces the prospect of paying out a significant damages award to HP. Even if it is as high as $4 billion as many reports have suggested — and it likely won’t be — Oracle’s balance sheet, flush with almost $31 billion in combined cash and short term investments, can take the hit.

But why let it go that far? HP CEO Meg Whitman has referred to the historical relationship that existed between HP and Oracle as “one of the great partnerships in IT history.” And Oracle CEO Larry Ellison has said he likes Whitman.

There have been at least two rounds of settlement talks held before and during the trial. Now that the primary issue of the trial — whether Oracle was bound to stick to an agreement it made in 2010 — has been decided, perhaps there’s ground upon which to build the foundations of a third way out of the dispute.

Oracle’s point that the Itanium chip is nearing the end of its life has more merit than either HP or Intel would care to admit. There may indeed be a few more generations left of Itanium, but nothing can change the fact that the world of enterprise computing is turning its back on Unix and non-x86 chips. The research firm IDC noted in May that the size of the market for non-x86 servers as a percentage of the overall server market has declined to the lowest it has ever seen.

Both HP and Oracle have to respond to this. HP’s response is Odyssey. Oracle, which owns the legacy Sun Microsystems business of SPARC-based servers running Solaris, has its Exadata, Exalogic and Exalytics hardware systems, all of which are based on Intel x86 chips. As the migration away from Unix plays out, customers of both HP’s Integrity line and Oracle’s SPARC systems are going to be forced to choose a way forward.

Depending on the case, both Oracle and HP will be jockeying for this emerging segment of post-Unix customers. One would think they’d want to do so with the maximum amount of customer goodwill. These are specialized customers — shared customers were the basis of the partnership in the first place — who don’t make their computing choices lightly and who tend to stick with one vendor for a long time. For them, the sight of these two tech industry heavyweights fighting so bitterly must be getting tiresome.