Yearly Archives: 2007

Burning the Midnight Oil

I wanted to take a moment to say thank you to a very important group of people working through the evenings and holidays – our employees, partners and customers. You know who you are.

For most of us, holidays are a great time to catch up with friends and family. They’re also a time of mad scrambles. As I’m sure isn’t unique to the American tradition, some among us postpone holiday shopping until the very last minute (today, even). Which leads to a late surge in infrastructure requirements from those businesses that continue to bet the world will ever get its shopping act in gear (I’ll take the other side of that bet, any day).

This leads to a late surge in purchase orders (for the record, we love that part, planned or otherwise), and then a late surge in shipment activities (not all datacenters are near airports, sadly), and then a late surge in installation activity – yielding a lot of travel for those responsible.

In a perfect world, we’d then be done, and home with friends and family.

But then it doesn’t end.

Christmas Day is a day of massively high load for Sun’s customers across the world. This year will undoubtedly set a pile of new records. Millions upon millions of network enabled gifts will be given in December, and a huge chunk will be unwrapped and turned on tomorrow. Digital still and video cameras will start pumping content to photo/video sharing services. Mobile phones will need to be provisioned, and will start downloading and sharing content (on a global basis, the network load from New Year’s Eve MMS messages goes beyond staggering). Set top boxes, networked picture frames, video game consoles, navigation devices, stuffed animals, sports equipment and automobiles – will all come on-line tomorrow. On the same day. And everyone will (and should) expect flawless service.

For some of our customers, it’s their single highest load day of the year – and single most valuable opportunity to shape their brands. For Sun and our partners, it’s a day we’re very focused on making successful. So to all of you working over the holidays – thank you. I and my team are aware and appreciative of your efforts, and you are making a big difference.

Please take the time to rest when the work is done. Remember, the difference between humans and computers is that our uptime is a function of our downtime.


Filed under General

The Internet as a Customer

It’s been a very busy second quarter to the fiscal year, and I’ll apologize up front for falling behind on my posting duties. When you write a blog, you welcome onto your shoulder an inner editor, or in my case, an outer editor in the form of a general counsel whose last missive to me was titled, “you need to write a blog.” Thanks, Mike.

I’ve been spending a ton of time with customers and independent software vendors – especially those that blur the lines (software as a service (SAAS) companies). We held an event a few weeks ago to which we invited a broad spectrum of such customers, one room filled with CTO’s from some of the world’s largest on-line companies (whose brands nearly everyone would recognize – the internet as a social phenomena is in full swing); the other filled with CIO’s – from the new world, as well as more traditional companies (banks, telcos, technology and retail companies).

Not all the attendees were customers (that is, some came just to hear what we were up to), and they came from China, Japan, across Europe and North America. Budgets ranged from $10,000 to billions. But all of the attendees were focused on using IT as a competitive advantage (why else would you spend two days with us, I suppose).

A lot of the insights were confidential, but here were a few items of interest…

  • No CIO wore a baseball cap. The same was not true of the CTO sessions.

  • The youngest company in the CTO session was started last year.

  • The fastest grower (in terms of datacenter infrastructure spending) was growing 100% per month.

  • The cost of people and change dominated the CIO room – not capital assets or power.

  • The cost of storage and bandwidth dominated the room filled with web companies.

  • Not a single company in the CTO room paid for software. Many knew Sun exclusively from our work in the open source or academic arena – validating free communities as a vehicle to meet new opportunities, before they join the Fortune 100.

  • In contrast, not a single company in the CIO room allowed free software without a commercial support contract. Not one. Validating the notion that for more mature/diverse companies, the cost of downtime dwarfs the cost of a support contract.

  • The CTO’s in the web companies wanted innovation at an accelerating pace.

  • The CIO’s (broadly) wanted innovation to slow down long enough for them to manage and exploit it.

Virtualization and open source storage were big topics in both rooms – almost everyone was aware of what we’ve been up to with ZFS (courtesy of one of our competitors). If you’re technically inclined, you can read about our approach to free, multi-platform virtualization in this whitepaper. There is a general sense the prices of both storage and virtualization will be falling in the next 12 months. We helped lend confidence to that thought.

  • The CTO’s said we were too hard to do business with, but they appreciated the ease with which our software could be freely downloaded.

  • The CIO’s praised us for being so easy to do business with, and one groused about the ease with which his developers can bring our software into his network.

This contrast made many Sun attendees want to pull their hair out (but customers obviously buy in very different ways, and we’re working hard to adapt our systems and processes for both).

Almost all the CIO’s were in the midst of new datacenter buildouts (due to insufficient power density for many, lack of space for others). Growth was a big issue for everyone, but in remarkably differing ways – one CTO had 9,000 systems, growing 5% per week – and storage growing even more rapidly… just to put that in context, that’s a social networking company, not a Fortune 500 bank.

All the CIO’s wanted to drive toward uniformity in their datacenters (“just like Southwest Airlines” – it’s cheaper and easier to manage an airline if all you fly are Boeing 737’s – diversity and variation is very expensive, especially at scale). Most of the CIO’s were struggling to free the resources necessary to drive a singular platform standard. Conversely, the CTO’s all had mandatory platform standards – with no variation permitted without explicit approval. Then again, the web companies mostly had the luxury of being less than 10 years old.

  • The youngest company in attendance was less than a year old.

  • The oldest company in attendance was more than 100 years old. The CIO of the latter (who’s quite a bit younger than 100) claimed he was still running processes from the 1800’s. He wasn’t smiling when he said that. We got a round of applause for working with IBM to port OpenSolaris to System Z mainframes.

Which is all to say… I feel quite good about how we’re prioritizing our investments, for CIO’s and CTO’s, alike. Around communities, efficiency, security, automation – and most importantly, innovation to address ever increasing scale. In everything we do, hardware, software, developer tools, all. Customer feedback around our technology roadmaps was, almost uniformly, “why can’t we have that sooner?” Which is a great buying sign.

But the most important conclusion was this: there is no longer a uniform definition of “customer” at Sun. From developer to startup, mammoth messaging service to 100 year old financial institution, customers, and their requirements, are as diverse as the internet, itself.

Which is both the challenge and the opportunity.


Filed under General

Solaris and Dell… and Virtualization, Of Course

It was an exciting morning… we made two big announcements at Oracle Open World.

First, we announced a key relationship with Dell, through which they’ll be OEM’ing Solaris, and directly supporting customers running Solaris on Dell systems. Second, we announced our free/open source virtualization roadmap, starting with xVM and xVM OpsCenter, our hypervisor and management product set.

With the Dell relationship, Michael joined me on stage (after I assured him there would be no uninvited hugs), and kindly offered me a Dell t-shirt (I gladly accepted). You can watch the whole keynote here.

Truth be told, the relationship with Dell has been in the making for a while – I flew down to Texas last year to have dinner at his house (with a fortuitous 180 knot tail wind – sadly, I had return the same night with a 180 knot headwind). If you’re thinking, “hm, didn’t Sun’s relationship with Intel start with dinner, too?” you’re picking up on a theme – great partnerships start with a meal, in my book. At that dinner, we began discussing ways we could work together. Since then, we’ve both heard from a ton of customers that they’re running Solaris (and Sun Software, broadly) on Dell systems – and they’d like us to work together to make the experience a seamless one. It’s important to note, of the Solaris instances distributed into the world, roughly a third run on Dell – that’s certainly motiviation for us both to work together.

Dell and Sun will work shoulder to shoulder to support joint customers. And we expect our respective sales organizations to do the same – in pursuit of the highest quality customer experience possible. We’ll be making joint investments to build new solutions for customers, working to expand the already large Solaris ISV community for Dell systems, and broadly work together to build new business. For customers, partners, Sun and Dell – win/win/win/win.

Dell’s now advantaged in the marketplace, as well, and alongside Intel and IBM, can better serve customers wanting a single hand to shake (throat to choke isn’t the experience either of us are seeking). Reciprocally, Solaris is clearly advantaged by association with Dell, the company that invented volume success in the IT marketplace. The relationship broadens the market for the both of us.

So why are we signing these partnerships, rather than simply locking Solaris to our own hardware? (Yes, I still get that question…) Because locking Solaris to Sun would be like a wireless carrier selling you a phone that didn’t roam – or an automobile manufacturer mandating you buy their gas after you’ve bought their car. There’s probably a market for both, it’s just smaller than the market we’re after – the global market. In which customers value choice.

So thanks, Michael, and the whole Dell team. We’re looking forward to building the market.

We also introduced our new virtualization offering today, the Sun xVM hypervisor, and Sun xVM OpsCenter management suite (the video, above, has a great presentation/demo by Rich Green, who runs our Software biz). I’ll be putting together some thoughts later on our approach to the virtualization market, but in short, for geeks… our xVM hypervisor is a very lightweight kernel that inherits proven virtualization technologies (like ZFS, FMA, Dtrace and Crossbow) from the Solaris kernel – while supporting Linux, Windows and Solaris as guests – imbuing guest OS’s with the properties of the host hypervisor.

We also announced a variety of partners today, most importantly Red Hat, who’ll offer reciprocity with their hypervisor, like Microsoft. As our mothers told us, it’s important for us to be a good guest, and a good host. We plan on doing both.

As with all our innovation, xVM is going to start first in the community, where we can engage the folks who’ll help make this a success – if you have an interest in joining the developer/administrator community we’ll build around openxVM and the OpsCenter management platform, come visit us at

(and finally… for those interested in why our ticker symbol changed again… it didn’t. After a reverse split, the exchanges append a character to the symbol for a period of a few weeks to let them adjust their systems.)


Filed under General

Offering Color on our Q1 Performance

We announced our first quarter results on Monday after the close of the stock market. You can read the press release here. And as usual, I thought I’d offer some thoughts around our performance.

We’ve heard a number of questions, as we do every quarter – this quarter, those questions came down to basically three areas. What went well within the quarter? What didn’t go so well? Why didn’t you grow – augmented with why did the share price decline?

So, without a lot of fanfare, here are some thoughts.

What went well?

I was very pleased with the leverage in our operating model – that’s just a polite way of saying we delivered our single biggest Q1 profit since 2001, and it’s good to be back in the habit of making money. Even in what’s traditionally our toughest quarter of the year, and even with a restructuring charge that took three cents off our GAAP earnings, we generated $89 million in net income.

We executed on our R&D, and delivered a seven year high in gross margin – up 5 percentage (yes, percentage) points year over year to 48.5%. Product gross margins were up 5.3% year over year, services gross margin was up 4.3%. Customers appear to value what we’re building, and our teams are executing well.

While our overall revenues were up only 1% over the prior year, deferred product revenues were up 18% (or roughly $100m). Deferred product revenue represents products (hardware and software) we shipped, that either hadn’t been installed or met customer acceptance criteria – we will recognize them in the near term, and it’s generally good to have a growing backlog of deferred revenue. In addition, we took roughly $20m out of channel inventories as we make progress toward managing our business on a “sell out of the channel” basis, vs. “sell in to the channel.” This shift makes us more transparent for investors (we recognize revenue on customer activity, not channel enthusiasm), and more effective in working with our partners.

Again, while overall revenues were 1% up Y/Y, our high end and mid range business was up 23% (year over year), our Niagara chip multi-threading systems were up a fantastic 70% -to a $750 million annual run rate – and our X64 business was up 10%, with blades generating around $40 million) and growing nicely, along with good momentum in our higher scale x64 platforms (as I’ve said before, virtualization depresses units in the short term, but increases configuration, ASP and margin). Our storage business stabilized, and grew faster than Sun overall, at 2.9%. Storage is getting a lot of focus, both at the software level (you probably saw we just introduced our CIFS implementation into the OpenSolaris community, which makes it an ideal NAS platform for Windows users), and at the hardware/system level, as well.

We delivered a big cash quarter, generating $574m, our biggest Q1 cash flow in… well, recent memory.

So net/net, we had lots to be pleased about – in addition to announcing new relationships in which IBM agreed to OEM Solaris, Microsoft agreed to support our virtualization efforts (and vice versa) and Google agreed to promote StarOffice, we also demonstrated our confidence in the business by buying back a massive $1.25 billion in JAVA shares. We put our money where are mouths are, for those wondering.

What didn’t go well?

The top line revenue growth was marginal – at around 1%. Deferred revenue isn’t counted, obviously, and we chose to take inventory out of the channel, knowing it would go against revenue within the quarter. That yielded a mixed bag on the top line, even with help from currency shifts.

Although our EMEA (Europe, Middle East and Africa) and APAC (Asia Pacific) businesses grew (some geographies, like India and China, in big double digits), our US business was down roughly 4% year over year (and at 40% of our revenues, if the US catches a cold, Sun catches a cold). Why was the US slow? We don’t have a crystal ball. Was it related to the US mortgage crisis? I can’t imagine it helped. Some financial customers grew, others shrank. Partners seem more excited than in a long while. So there’s no easy analysis.

Our volume systems business grew only 3%, so our total computer systems business grew only .5%. Our new products haven’t eclipsed our legacy businesses – even though the former are on good ramps. Our support services business was down just under 1%. And we had demand for several non-shipping or new products, notably our newest quad core x86 servers and eight core Niagara 2 systems (both now shipping), but some customers may have waited for Q2. Again, very hard to tell at a global level. On an anecdotal basis, customer sentiment is way up on Sun. But sentiment and purchase orders are authored with different pens.

But push comes to shove, the only thing I was disappointed in was our top line growth – nearly every other metric is now moving in the direction we want. Financial measures, like margins, cash, predictability, quality, competitiveness of the product line – all the right foundational elements. And longer term measures – adoption around OpenSolaris, support for Java on devices and among the developer community, analyst ratings.

So fine, why didn’t you grow? Why’d the share price decline?

As I said, we don’t have a perfect answer, other than to say – we don’t see competitive issues holding us back, the team is executing well, and the brand and reputation of the company are now amplifying opportunity, not depressing it. The single biggest issue was some customers slowing down some purchases in the US, but it wasn’t specific to financial services (although companies in the midst of CEO transitions tend to put a lot of things on pause), or even more general to all customers (several industry sectors grew, like our government business).

Why’d the share price decline? This is obviously a question lots of folks asked (especially after today’s market decline) – and the only answer I can give definitively, and without meaning to be cavalier, is that there were more sellers in the market than buyers. The opposite of what makes our share price rise.

And just in case we failed to make it clear, of course revenue growth remains our highest priority. We’re focused on taking a refreshed product and service lineup, into markets we expect to grow across the earth – and investing to deliver value to customers and shareholders. In my view, we took another step forward in Q1.


Filed under General

Congratulations Google, Red Hat and the Java Community!

I just wanted to add my voice to the chorus of others from Sun in offering my heartfelt congratulations to Google on the announcement of their new Java/Linux phone platform, Android. Congratulations!

I’d also like Sun to be the first platform software company to commit to a complete developer environment around the platform, as we throw Sun’s NetBeans developer platform for mobile devices behind the effort. We’ve obviously done a ton of work to support developers on all Java based platforms, and were pleased to add Google’s Android to the list.

The Java platform has come a long way – we’re on the vast majority of mobile devices in the marketplace today (well over a billion phones at last count), and with Google’s mobile services (like gMail and Google Maps), along with Yahoo!’s Go Mobile, alongside a massive spectrum of incredible entertainment offerings from folks like Electronic Arts, we have by far and away the most complete content ecosystem on the market today. Enabling carriers, handset manufacturers, content creators – and most of all, consumers – to get the most from their mobile devices.

And needless to say, Google and the Open Handset Alliance just strapped another set of rockets to the community’s momentum – and to the vision defining opportunity across our (and other) planets.

Today is an incredible day for the open source community, and a massive endorsement of two of the industry’s most prolific free software communities, Java and Linux.

Stay tuned for specific details – for those so inclined, download NetBeans here to check out the industry’s most popular IDE for mobile Java development.

And in the spirit of offering congratulations, I’d like to offer a similar shout out to our friends at Red Hat Linux – who today announced their support for the OpenJDK project. With friends like Google and Red Hat, it sure seems like the momentum behind Java’s on the rise…


Filed under General

ZFS Puts Net App Viability at Risk?

About a month ago, Network Appliance sued Sun to try to stop the competitive impact of ZFS on their business.

I can understand why they’re upset – when Linux first came on the scene in Sun’s core market, there were some here who responded the same way, asking “who can we sue?” But seeing the future, we didn’t file an injunction to stop competition – instead, we joined the free software community and innovated.

One of the ways we innovated was to create a magical file system called ZFS – which enables expensive, proprietary storage to be replaced with commodity disks and general purpose servers. Customers save a ton of money – and administrators save a ton of time. The economic impact is staggering – and understandably threatening to Net App and other proprietary companies. As is all free innovation, at some level.

So last week, I reached out to their CEO to see how we could avoid litigation. I have no interest whatever in suing them. None whatever.

Their objectives were clear – number one, they’d like us to unfree ZFS, to retract it from the free software community. Which reflects a common misconception among proprietary companies – that you can unfree, free. You cannot.

Second, they want us to limit ZFS’s allowable field of use to computers – and to forbid its use in storage devices. Which is quizzical to say the least – in our view, computers are storage devices, and vice versa (in the picture on the right – where’s the storage? Answer: everywhere). So that, too, is an impractical solution.

We’re left with the following: we’re unwilling to retract innovation from the free software community, and we can’t tolerate an encumbrance that limits ZFS’s value – to our customers, the community at large, or Sun’s shareholders.

So now it looks like we can’t avoid responding to their litigation, as frustrated as I am by that (as I said, we have zero interest in suing them). I wanted to outline our response (even if it tips off the folks at Net App), and for everyone to know where we’re headed.

First, the basics. Sun indemnifies all its customers against IP claims like this. That is, we’ve always protected our markets from trolls, so customers can continue to use ZFS without concern for spurious patent and copyright issues. We stand behind our innovation, and our customers.

Second, Sun protects the communities using our technologies under free software licenses. As an example, Apple is including ZFS is in their upcoming “Leopard” OS X release. This is happening without any payment to Sun (that’s how truly free software works). Under the license, we’ve waived all rights to sue them for any of the patents or copyright associated with ZFS. We’ve let Apple know we will use our patent portfolio to protect them and the Mac ZFS community from Net App. With or without a commercial relationship to Sun.

That’s true for any licensee – in fact, Net App could adopt ZFS today and receive the same protection. The port is done to FreeBSD, the OS on which Net App’s filers are built. They could use it without owing us a dime, and they’d be protected from our portfolio. (The quid pro quo? They’d have to agree to offer reciprocal protection to Sun.)

Third, we file patents defensively. Like MySQL or Red Hat, companies similarly competing in the free software marketplace, we file patents to protect the communities from which innovation and opportunity spring. Unlike smaller free software companies, we have one of the largest patent arsenals on the internet, numbering more than 14,000 issued and pending globally. Our portfolio touches nearly every aspect of network computing, from multi-core silicon and opto-electronics, to search and of course, a huge array of patents across storage systems and software – to which Network Appliance has decided to expose themselves.

And to be clear, once again, we have no interest whatever in suing NetApps – we didn’t before this case, and we don’t now. But given the impracticality of what they’re seeking as resolution, to take back an innovation that helps their customers as much as ours, we have no choice but to respond in court.

So later this week, we’re going to use our defensive portfolio to respond to Network Appliance, filing a comprehensive reciprocal suit. As a part of this suit, we are requesting a permanent injunction to remove all of their filer products from the marketplace, and are examining the original NFS license – on which Network Appliance was started. By opting to litigate vs. innovate, they are disrupting their customers and employees across the world.

In addition to seeking the removal of their products from the marketplace, we will be going after sizable monetary damages. And I am committing that Sun will donate half of those proceeds to the leading institutions promoting free software and patent reform (in specific, The Software Freedom Law Center and the Peer to Patent initiative), and to the legal defense of free software innovators. We will continue to fund the aggressive reexamination of spurious patents used against the community (which we’ve been doing behind the scenes on behalf of several open source innovators). Whatever’s left over will fuel a venture fund fostering innovation in the free software community.

And on that note, I want to thank the free software advocates from across the world who’ve offered expert testimony, and reams of prior art to defend ZFS, and the community of which Sun’s a part. Please rest assured we will use this opportunity to highlight the futility of using software patents to forestall competition – in the commercial marketplace, and among the free community.

In the interim, if you’re a Net App customer looking for alternatives, we would be pleased to talk to you about lowering the cost of proprietary storage – if you’re a technical sort, start by trying out ZFS in software form. (There are also lots of reviews available, this one just posted). We’d also be happy to send you a free trial Storage System based on ZFS (pick the x4500 here). And remember, we indemnify our customers.

The shift to commodity infrastructure is as inevitable as the rising tide – although for some, I’m sure it feels like a rogue wave.


Filed under General

Hugging Customers (Not Trees)

(Update – Interview with Dave Douglas, Sun’s VP, Economic-Responsibility at bottom of this entry)

A few folks know I like to cook. It pairs well with liking to eat. A good friend gave me a beautiful cookbook last year, Thomas Keller’s The French Laundry Cookbook.

It’s gorgeous book, but I will admit to having prepared a dish from it only once – the recipes are complicated, and take far more time than I typically have (the author admits, up front, people like me aren’t his target demographic). But one of the things I love about the book is Keller’s focus on efficiency – out of respect for the ingredients, and the economics of running a commercial kitchen. “Should a cook squander anything, ever?” Great chefs and kitchens waste nothing.

Now, waste is a generic term. It means one thing to a consumer frustrated with the 60,000,000 plastic bottles we dispose of daily in the US. It’s another thing to a CIO who just realized she’s running datacenters at 10% efficiency. At that level, as in a commercial kitchen, it’s not an annoyance, it’s a waste. Of money.

About five years ago, we made a simple but important bet – that our customers would eventually look at waste in their datacenters with a far more scrutinizing eye. We bet the cost of powering a computer would eventually exceed its purchase price – and thus focusing on energy waste would be a profitable pursuit. (Just imagine if your gasoline cost more than your car… oil hit $86/barrel today, btw.)

We introduced our first energy efficient server system about 18 months ago, known as Niagara 1 (the name is a nod to the machine’s throughput capacity). We did the introduction in London, home to the world’s most expensive real estate – space being a precious asset for most of our customers, as well.

But after years of R&D, we showed up with something confusing: we rejected the notion of speed at any cost (a first in the industry), and optimized the system not for speed, but for efficiency, reducing power waste. Like a bus, not a Formula 1 racer (the former getting radically better per passenger mileage). We similarly suggested the world should move away from running one application per machine, and instead collapse many tasks onto the same machine – heresy at the time has since been renamed virtualization.

Those who love desktop computers thought we were daft. Here we had what looked like a slow chip, optimized for something no home user really cared about (lowering power bills, running multiple OS’s and minimizing space). And to make matters worse, we removed support for floating point precision math on the chip – to save more power and space. Desktop users (who play games that often feast on floating point processes) thought we were loons, but most datacenters didn’t notice (very few datacenters use floating point).

Net result? Within two quarters, revenue hit $100m per quarter. And by last quarter, systems based on Niagara 1 were nearly a billion dollar annualized business – from a cold start 18 months prior. It wasn’t for every application (don’t simulate nuclear fission on a Niagara 1 system), but for internet workloads (databases, web and app servers), it set new records for work/watt.

Were we hugging trees? That wasn’t the only point. We were hugging our customers. Customers out of space and power who, like great chefs, feel there’s no point in waste.

So a couple weeks ago, we introduced the second generation of Niagara based systems, known as Niagara 2. Niagara 2 adds an incredible breadth of new features and performance enhancements (this is a good summary).

We shrunk the process used to make the chip, and upped the clockrate to boost overall performance. We doubled the thread count (8 cores x 8 threads = 64 threads), and via xVM (formerly, Project Virginia) virtualization: a single Niagara 2 chip can collapse 64 independent operating systems onto a single system. Not separate application partitions, 64 separate operating systems – from Solaris and other real time OS’s, to Linux or BSD. (As far as we know, no one else can do that without a huge performance penalty.)

We added cryptographic ciphers (algorithms used to encrypt/decrypt data for secure storage or transmission over the internet) onto the microprocessor – so users don’t have to power, or provision space for extraneous security functions. We added (dual 10 Gigabit ethernet) networking onto the chip, eliminating yet more waste. Niagara 2 isn’t simply a server, it’s a true system – in the spirit of the Systems team introducing it. Our first Niagara 2 systems are as wondrous a head-end for storage farms (using wire speed encryption), as they will be PBX’s, firewalls, routers – and with floating point added back in force, wonderful rendering engines and high performance computing machines.

And on equivalent workloads, we dropped our power draw – that is, we lowered the power required to do the same amount of work.

Again, tree hugging? Nope, customer hugging. Does that make Sun green? Not by a long shot, but it’s a good step forward.

So our investments in eco-responsibility are starting to pay off – and the trends we were in front of a couple years ago (power efficiency and virtualization) have only become more relevant. Not to everyone, certainly – only those that care about minimizing waste. If you’d like a free trial Niagara 2 machine, just click here.

And much though I appreciate Kevin Maney’s column, I think he’s wrong in this opinion piece – to dismiss a focus on eco-responsibility as a fad. It’s one thing for waist-conscious consumers to avoid high carb diets. It’s another thing entirely when you’re talking to a government seeking to avoid new coal fired power plants, a C-level executive committed to extreme efficiency, or the employee of a company whose CEO has just said, “We will be carbon neutral.”

There is, and there will always be commercial opportunity in eliminating inefficiency – to be clear, that’s our primary motivation. Paint it any color you want.

Along the way, if we reduce our carbon footprint, minimize our waste stream, and get crisper about our views on corporate social responsibility – does that have an impact on customers wanting to do business with us? Absolutely yes. I’ve seen it firsthand.

It also has an impact on our competitiveness as an employer. Do our current and potential employees care about the efficiency and responsibility of our business? At least as much as the chefs in Keller’s kitchen, if not more.

(Interview with Dave Douglas here – Dave is leading the effort to reduce waste and environmental impact at Sun, and downstream in our customer base.)


Filed under General

Going Bollywood

I just got back from a series of customer meetings with technologists from the telecommunications, media and entertainment industry (they are, after all, converging on the same market, monetizing consumers). I was joined by Greg, and a number of folks from Sun, notably Jeff Bonwick and Bill Moore, along with Matt Ahrens, the co-inventors of the ZFS file system.

All three industries are largely underserved by innovation – their requirements and opportunities are vastly outstripping the rate at which the industry is innovating. (That’s code for, “their technology budgets are growing.”)

There were a number of choice anecdotes from the sessions, my two favorites being the following.

The first was relayed by the CTO of a major movie studio, who explained the value of very, very long term deep storage archives.

His company had recently pulled a more than fifty year old movie out of a salt mine – where it was stored on 35mm film color separations. Before you ask, vinyl film outlasts the standards that the industry produces (a point on which I’ve written previously), and salt mines are environmentally more stable than datacenters). They pulled the separations to remaster and reissue the film on DVD.

In so doing, the resulting movie actually increased in resolution for the viewer – the devices used to display the movie today (likely a laptop or high definition TV) offer better resolution than what was available to the original viewers (likely a 50’s era TV or movie theater). The movie improved with time – the modern viewer saw the movie in greater detail than the original viewers (suggesting one should always store higher resolution data than the display you see in front of you).

After release of the digitally remastered version, the DVD rose to number 8 on the Amazon best seller list.

Cost of production? Near $0. The effort was nearly pure profit.

And they have a library of approximately 30,000 films.

That’s a ton of value in a salt mine (if all the titles are as compelling, which is unlikely, but an interesting thought exercise, nonetheless).

The second anecdote relates to the advent of very high resolution digital cameras – the highest, and most desirable are currently known as “4K,” offering 4096 x 3112 pixels (!) per frame – yielding cameras that spew 100’s of megabytes of imagery per second).

The director of one feature length film wanted to keep all the footage from a soon to be introduced new movie. He wanted to preserve outtakes and all, for the eventual “Behind the Making of…” or “Director’s Cut…” versions of the movie. The digital master for an average 4k film is roughly 9 Terabytes – that’s for the version you and I would see in a theater.

But the total archive including out takes and secondary/tertiary angles (bits are a lot cheaper than film, so why not set up three or four cameras for every shot?) was roughly (drum roll please)… a PETABYTE (or a thousand terabytes, or roughly 500,000 iPods). Equivalent to about a million feet of 35mm film.

That’s a lot of data. To be archived for… well, to the first anecdote, likely forever (like health records or airport surveillance). And now you know one of the (many) principle motivations behind a file system we built at Sun, ZFS. The focus of the ZFS team is both the scale, simplicity and quality of storage (on Mac OS X, BSD, Solaris and Linux).

This is a great overview of ZFS by Jeff and Bill – of why ZFS matters to the media and entertainment industry, and likely anyone concerned with high quality, high scale and high productivity, storage.

As I was told by Jeff (parroting a storage executive), there are only two types of disk drives in the industry.

Drives that have failed, and drives that are about to fail. And now you know what inspired ZFS, and what’s inspiring interest in it.


Filed under General

All the Wood Behind One Arrow

I’m radically increasing Sun’s focus on storage today.

Why? Because the market’s only going to grow, for as long as we’re on this earth, and I believe our talent and assets give us a big sustainable advantage – that we’re planning on exploiting. Aggressively.

How? First, I’m going to be combining our Storage and Server product teams to create a new converged group at Sun known simply as our “Systems” team. The Systems team will focus on the evolution and convergence of computing, storage and networking systems. Talk to any datacenter adminstrator, and that’s what they want to hear – they live in a world managing the (often idiosyncratic) interactions of that trinity (computing, storage and networking – and just wait until they’re virtualized). We want to be in a position to innovate on their behalf, at the system level, beyond the boxes – across blades, racks, disk and tape.

So we’ll still be strongly focused on being a multi-platform storage provider (just as our servers run multiple operating systems, and our operating system runs on every vendor’s servers), but we’re also going to start talking at a higher level to customers that see more standardization and integration in their future datacenters. That’s not everyone, but it’s definitely a trend we’re going to accelerate (and again, that’s what virtualization portends).

Now, why do I believe combining groups make sense? It’s a recipe that works for us. We combined our high volume x64 server group with our traditionally high scale SPARC server group over a year ago – leveraging the volume skills of the former with the scaling skills of the latter. What did that collaboration yield? The highest scale x64 systems in the market. And a refreshed lineup of volume systems powered, interchangeably, by SPARC, Intel and AMD. We also combined our networking expertise to build the best general purpose blade platform below (known internally as C10 or Constellation 10 – 10 blades, vs. “C48,” with 48), with integrated networking and seamless management. Can you tell where the server stops, and the storage starts? I can’t (and in a virtualized world, it’s not terribly important).

Secondly, as our servers clearly show, we’re heading to a general purpose world – in which open and general purpose platforms will be the dominant drivers of growth, for us and the market broadly. The first general purpose storage system from Sun was Thumper (our x4500) – powered by an open source operating system (Solaris), and file system (ZFS – soon to be parallelized by Lustre, a recent acquisition from Cluster File Systems). Thumper rocketed to a $100,000,000 annual runrate within its first two full quarters of shipment (on a $13 billion dollar revenue base, that’s hard to see, but we certainly took notice – at least one competitor did, too).

Combine these assets with some of our recent network innovations (like Magnum, the world’s largest Infiniband switch – which is not the smallest variant we’ll build, btw), the Crossbow community in Solaris – and it begins to look like we’ve got all the right ingredients to reinvent the datacenter.

So in this instance, I’m expecting our Systems team to be just as focused on standalone storage and networking – leveraging disk, tape (and all future removable media) – as they are on building great integrated systems (like the Constellation System, above, or our Thumper platform). I’m expecting to see more innovation, faster time to market, and a breadth of opportunities emerging from serving our current customers better than ever, while inviting new customers with a constant stream of high value innovation.

And before I end, I want to focus on one particular group, whose value only grows to Sun every day – our Tape and Archive business. From a market perspective, some data lasts forever – surveillance video, health and insurance records, trading histories, etc. In our view, the market for permanent data will only grow. Today, only tape can maintain the integrity of that data without electricity. And for the datacenters we serve, many are seeing the cost of electricity threatening to eclipse their hardware budgets (yes, I’m serious). For disk storage, over a decade, that’s easy to see – just look at the power bill to run a SAN, mulitply it by a decade.

So while we’ve been selling very large libraries to the largest companies in the world, in mainframe and open system shops, we’ve also begun meeting a variety of startups and web 2.0 companies (some household names, even). With the need for very, very large pools of archived storage (when you collect user generated high definition videos or satellite imagery for planetary social networks, it’s easy to find yourself with peta-scale archive problems).

Tape, with effective indexing and retrieval, represents the most economically responsible (that is, eco-responsible) archive platform for long term storage. Broadly speaking, tape (and in the future, other forms of removeable media) are a core part of Sun’s archive plans. We think there’s a ton of innovation we can bring to that market – now that we’ve done the basic integration (did you know a Solaris powered server is now embedded in our libraries – bringing new meaning to Free Inside!). As we converge high performance networking, virtualization and file system innovation – along with an overall Systems approach – our archive business will benefit just as much as our blades and rackmount systems. They are all, after all, members of the Systems family.

So like I said in the opening, I’m dramatically increasing Sun’s focus on storage today. By bringing to bear the talent and assets we have from across Sun to ensure our success. From where I sit, we have the right leaders and assets, and the right target in front of us.

Now’s a great time to put all the wood behind one arrowhead.


Filed under General

Value of Design

Last week, we joined our friends at Intel at their Intel Developer Forum, to unveil the second generation of our Systems platforms. A few of our folks from the event took photos with their phones and sent them my way.

A few evenings later, I was at a dinner presentation with about 30 CIO’s – one of whom had seen some of the IDF coverage. The customer (who does no business with Sun currently) asked me, “why should we bother talking to you? How can you possibly differentiate – it’s all a commodity to me.”

This is one of our biggest opportunities – and biggest challenges. But not as you might think. Most of the world doesn’t do business with Sun – that’s the good news. That’s also the challenge – how do we reach customers when we don’t have a pre-existing purchasing relationship based on PC procurement (like our peers)? On a global basis, it’s a tough nut to crack (and no, it’s not as simple as “Just advertise on TV!”).

So this was the picture I showed during my presentation – see any difference between our system (codename, Tucani, real name, x4450) and the others? (Answer: we’re half the size – the power and performance difference is admittedly harder to see in a picture.) If your business or datacenter’s in a place where space, power density or performance is at a premium, we’d be pleased to engage – running, of course, Windows, VMware, Linux and Solaris.

Step one of reaching new customers is being better by design – and fueling word of mouth. After all, the internet reaches more people than a television.

(And for the technical among you, what’s shown is our first four socket Caneland system, with 32 Dimms, 6 PCI-Express slots, built in RAID, up to 8 hot swap disks, redundant power & cooling, full remote systems management… all in 2U). Read more here and here. Also available as a two socket/1U platform…)


Filed under General