With Sun Microsystems on the block - and perhaps on the chopping block - for the past several weeks, it is probably a good time to take a gander at the family jewels: Sun's software business. Sun likes to count downloads and the size of installed bases because the company's top brass believes this a kind of leading indicator for …
You've picked the software products that are high profile in the opensource market, and obviously there are other considerations there, but why ignore things like Identity Management, Directory Server, Sun Cluster, etc.? Those are products where Sun has major maket share, and not just in the Solaris market. I work for Sun engineering and am obviously in no position to share numbers, but those do make revenue and profit. Solaris may be a "loss leader" but it's far from being the only card in the deck.
A few things to Add about Solaris profitability
A key difference between Redhat and Solaris is: DTrace. Dtrace allows support engineers to resolve customer production issues a lot faster than Redhat, there is no need to send the customer debug kernels... This results in better customer satisfaction and reduced servicing costs for Sun.
Another difference is ZFS that architecturally is also more reliable by architecture with end to end checksums and the fact that ZFS is always consistent on disk and thus not requiring FS repairs after a system crash. ZFS clone and snapshot support allows for safe upgrades that can be rolled back safely... This also reduces Sun's support costs.
My point is that at least theoretically Solaris support business can be more profitable then the Linux business...
Solaris still has some catching up to do with the user interface in some areas, but still manages to out-innovate Linux with features like Time Slider ...
The future for sun and its shareholders is not that bleak, if redhat can have a market valuation of 3.5B $ with 0.5B in revenue, Sun with 13B of revenue can certainly do better than its current 5B valuation..
You are suggesting...
...that Sun should charge for their software-stack? This would work in the short-term but not in the long-term.
What Sun needs is marketing. They have to get the word out, that Sun's support is better and cheaper than Redhat's (It is!!!).
@zvonr, don't forget Sun has over 30,000 people to pay every month and hasn't made a profit in a long while, whereas Red Hat employs only 2,200. I work that out to be about $34k annual profit per employee. If Sun could get that kind of employee productivity, it would be making $1bn profit per year. Just numbers, not opinions.
@ AC and zvonr
@ AC - right on, and how about Communications Suite for that matter, zillions of deployed mailboxes worldwide:
Indeed, and if we are talking OS features to reduce downtime and accelerate resolution of production issues, let's not forget Fault Management Architecture and predictive self healing:
Show me a free OS that can do *that*, and maybe I"ll take it seriously over Solaris/OpenSolaris.
The world is dumb
"The future for sun and its shareholders is not that bleak, if redhat can have a market valuation of 3.5B $ with 0.5B in revenue, Sun with 13B of revenue can certainly do better than its current 5B valuation.."
My feelings exactly; if Sun were only turning over $500m then I'd be worried. As it is, it is still selling >$10b ($10,000,000,000) of goods a year; hardly a doomed company.
Why don't sun flog Netbeans and other dev tools for Java? It doesn't do MS much harm, they make a packet from VS and MSDN subscriptions. They could always release cut down Express versions for free etc.
RE: A few things to Add about Solaris profitability
The reality still hasn't kicked in with poor old zvonr.
"A key difference between Redhat and Solaris is: DTrace. Dtrace allows support engineers to resolve customer production issues a lot faster than Redhat, there is no need to send the customer debug kernels... This results in better customer satisfaction and reduced servicing costs for Sun...." Nice feature sell, but completely irrellevant. Apart from the fact dtrace is an engineering tool and will not be of interest to anyone above coder or sys admin level (and that means the people with the purchasing power), Slowaris could have a tool that made toast and faked harrassment calls to your mother-in-law, and it still wouldn't overcome the fact that Sun - and hence Slowaris - has a completely unsure future. Any CEO reading the news at the moment (well, for the past few years actually), will have seen article after article about the uncertainty of Sun's future. The news that IBM has backed off will have alarmed people that hoped IBM might mean a stabilising of the ship. Now, once again, nobody even knows if full Slowaris support from level one through to a commercial, code-writing lab will even be available next year. Such uncertainty is going to kill Slowaris sales.
"....Another difference is ZFS that architecturally is also more reliable by architecture with end to end checksums and the fact that ZFS is always consistent on disk and thus not requiring FS repairs after a system crash. ZFS clone and snapshot support allows for safe upgrades that can be rolled back safely... This also reduces Sun's support costs...." But has done nothing for Sun's revenues, didn't you read the article? And here is the big hole in yet another Sun feature sell - ZFS is only great if you need it. The majority of buyers seem quite happy with OSs using tried and tested volume managers and file systems. The fact that the latest Linux kernel has gone with BTRFS will also reduce interest in ZFS in the Linux community. And then, if you think of what happens should Sun go pop, who is going to fight the NetApp WAFL lawsuit? Would you implement ZFS with the chance that the big money lawyers keeping NetApp at bay could disappear overnight, leaving NetApp a clear field to come round and ask you for a licence for their plaguerised product?
"....My point is that at least theoretically Solaris support business can be more profitable then the Linux business..." Your point, whilst debateable, is moot seeing as Sun can't make enough money to actually pay for that support service. This is called the reality of business; theory is great but profits are a lot better. The Red Hat business model has been conceived from the start as an open-source software business selling support to make money, and the company is very carefully structured to that role. Sun is still a strumbling behemoth of a failing hardware business, losing money whilst trying to morph very unconvincingly into an open-source software business.
"....Solaris still has some catching up to do with the user interface in some areas, but still manages to out-innovate Linux with features like Time Slider ..." <Yawn> What is this, Feature Sell Weekly? Get over the techworship and realise that business just doesn't work that way. If any of the features you've mentioned actually gave Slowaris any advantage in the marketplace then IBM would be choking to buy the Sun software bizz rather than the reality - very publicly backing away at speed.
"....The future for sun and its shareholders is not that bleak, if redhat can have a market valuation of 3.5B $ with 0.5B in revenue, Sun with 13B of revenue can certainly do better than its current 5B valuation.." More irrellevance - Sun and RedHat are completly different types of business, even though Sun are desperately trying to copy RedHat's. RedHat makes that money with a much smaller and much more focussed business, and the market believes that RedHat will make a profit. No-one is convinced Sun can do the same in time without someone else coming in, buying the bizz, and doing the deep and drastic whittling of deadwood that the Sun board seem incapable or unwilling to do.
For the foreseeable future, Slowaris and profit are just incompatible.
Solaris licenses a
Correction: Solaris 10 licenses (and OpenSolaris licenses) on *all systems* are free to end users, period. Doesn't matter whether it's big iron or small iron, or middling gadolinium, or whatever.
But that's not where Solaris revenue comes from -- it comes from support, which is not free, as well as OEM licensing. And yes, most people who deploy big iron-type applications do tend to get support contracts.
I have no idea what you mean by "Sun has to basically give Solaris support away for free to cover the costs of Sparc chip and server development." First of all, it's a very strange premise: how does giving something away cover costs? Second, it appears to be cut from whole cloth -- who has said that Sun is giving Solaris support away? Sun customers -- and Sun OEMS, such as Fujitsu, IBM, and now HP -- pay Sun for Solaris support.
And again, if you're deploying a mission critical application, you're not going to operate without a net; you *are* going to pay for support. It doesn't matter that the cost of the system may be lower than it was, it doesn't matter if the OS license is free, you're going to want ongoing support for the hardware and the software. It's no accident that IBM's fortunes rose as they turned their investments to their Global Services division.
Red Hat has some things too
You don't send debug kernels, the debuginfo is an extra package that you can install any time from the repository, and the information can be used by debuggers and other tools. I think it is really rare when you have to add something in the sources (debug info) and send it to the customer. And I think that more than 90% of the time the standard tools will do the job. Also system-tap is not that bad, I bet the RH technicians can use it too. Red Hat topped customer satisfaction the last years among companies.
Also form what I remember just Open Solaris has a software repository with OS packages and 3rd party packages (Linux out-inovated them here). If you want to support the in the regular Solaris what do you do ? They are in external repositories (aka blastawave) than I'm not sure how tehy are controlled by Sun. Redhat by the way has their own repos, knows exactly what they put there, test the packages and have control over them. I think that help supporting it.
Hehe, yes, out-inovate with Time-Slider. Hey Java Desktop = GNome, remember CDE, it sucked. Linux will probably have time slider too, or something similar (it should be in Gnome, but I don't know how generic it is implemented or it is ZFS coupled) when btrfs will become main-line. Great invention time slider, heard of Time Machine from Apple ? I agree ZFS is nice (WAFL is too).
And regarding Linux <-> Solaris support, Linux is developed by many companies & community, Red Hat doesn't have to build everything, to fix all the bugs, others will help too. So I think that helps reducing the support costs.
And yes, I agree , I think Sun is under-valued on the market.
What Really Scared IBM During Talks....
Okay, so I work in Sun Software (so I have to post anonymously) and tired of listening to outsiders who don't see the forest through the trees.
Sun is a great software company that has spent the last few years silently changing the game. What IBM probably saw was they could not make Websphere make money if they adopted Sun's software approach. Oracle should be scared as well. Really scared.
By open sourcing and building communities we have built over the years a unique software business model. One that is rapidly becoming very attractive to businesses in this down economy. Red Hat may have a Linux distro (which we support as well) and a JBOSS server they have done little to nothing with (we're moving the cheese with Glassfish), Sun has products at every level, from OS to applications, identity, utilities and monitoring. And you can download them and work with them for free.
But when they want to get serious, they are willing to pay for support and licensing. We can see MySQL downloads by the thousands into dedicated IBM and Oracle shops who previously would never stray from the fold. Now these customers, as dedicated as they are to their vendor, are asking to buy our software to save on costs. We can't help but think IBM looked at their cashflow from bilking Websphere, Tivoli, and DB2 customer accounts and saw a threat of Sun's software business model. Was it surprising when it leaked during the acquisition talks that IBM could not see how it could make money with Sun software? They would lose too much money moving to the new model that it would drag down their entrenched software business.
Lets look at another disruptive marketing approach - The Grateful Dead. They encouraged everyone to download and share their music. They listened to their audience and changed to stay in touch with their loyal fan base. Everyone asked how they could make money selling their records when they gave the music away. The answer is they never made money on the bootlegged (read downloaded) music, but it built their fan base to sell them other things, like concert tickets and merchandise. They never had a #1, but they are one of the most profitable bands by selling merchandise and concert tickets, not records. And a fan following so loyal it is almost tribal. Same with us. We will make money on other products, like enterprise versions of our products and support (noticed I did not say hardware though there is a pull through there as well).
One of our biggest software transactions recently was to include the Google Toolbar with every download of the JDK. Since then, Microsoft has paid us millions to switch to down loading MSN toolbar with the JDK. Buying a seat on the free JDK downloads is highly profitable to us. Think of all the downloads that go with MySQL, Star Office, Solaris, Glassfish, OpenSSO, etc.
What IBM probably got scared at was the threat this model is to their bloatware Websphere. Same with Oracle. One CxO has said they have a $50M+ annual license from Oracle and does not see the value (repeat value) from maintaining that license. He/she is looking into MySQL at 1/10th the license cost (don't believe MySQL is a poor cousin to Oracle - it works for 98% of all DB applications across most enterprises right now). We don't have to try so hard to break into a IBM or Oracle shop like we used to; we are being invited in by the customer. Really doubt we have the reverse problem with our customers.
It has taken use 3-5 years to build communities around our software and adapt to this new software model. It makes it very difficult for our competition to follow suite. How many open source database communities can one have? How long would it take Oracle or IBM to open source their bloatware and switch to our model? That is what scares IBM. It could never make money the way they do now (by heavy-handed, perpetual licensing models) if they had to adapt to our model. The reason they leaked the "Sun software cannot make money" story during the acquisitions negotiations is IBM could never make money if they had to adopt this model. They would buy Sun just to kill the business model off. BTW, those acquisition talks had more leaks to the press than it should have. Hope IBM looks long and hard where those leaks came from.
And one last point. Customers are telling us they want out of the integration game. They are tired of buying IBM and Oracle software (and ours to some point) and paying the vendors to weld it together. So, any vendor that can provide a comprehensive suite (sorry Red Hat) and not have to its customers shell a ton for licenses and consultants, has a distinct advantage. Whereas a few years ago, our open source approach was a nice to know about for a CxO (they wanted information about it, but had little incentive to make the enterprise switch), now they are calling us to find out more because the economy has given them an incentive to look a viable, lower cost alternatives. And they are buying. Repeat, buying.
Its not the software or support business that is causing Sun to post negative earnings - its the high end server and storage business, which is still the larger portion of the company. The downturn in the economy has put the brakes on everyone in that market. But that is changing. Sun software is reaching critical mass within the company. Over time, as Sun software becomes an equal revenue contributor to the bottom line, the company's fiscal status will be less whip-sawed as one sector goes through its cycles.
And one last point - to the poster using the term "Slowaris". First, dude, read a DTrace manual and tune. Solaris is used in many more high load, high reliability situations than Red Hat ever has. And it has 5 times as many applications available to run on it than RH. With OpenSolaris, customers can be assured that Solaris will have a long and happy life, with or without Sun itself. Ask some of our recent Solaris customers; IBM and HP.
Re: "What Really Scared IBM During Talks...."
@AC, I completely agree with you. If you're in the software sales game, then it's suicide to bite the bullet and go down the "Grateful Dead road" of open source distribution. Sun's way is simply incompatible with Oracle or IBM. And this means that the chances of Sun being sold to anyone anytime soon are slim to non-existent, which is good news for Sun employees and the many communities that Sun has grown over the years.
I believe Sun's software will do very well within 2 years, but that things will get a little tougher before they get better - for the very reasons you outlined: high end servers and storage slowing down in this recession. Until Sun can put the "big iron" behind it and focus on cloud services and software support, the big picture will look somewhat negative to a cursory glace over the accounts. If I could make one software suggestion it would be to wholeheartedly welcome Linux as an option for customers, and "get in bed" with Ubuntu and Gentoo. While there is a perception that Sun is about Solaris and not Linux, it will cause a great number of potential customers to shop elsewhere for support.
Good luck. You're 30% foolish and 70% brave IMHO.
Of leaks and scares
Methinks you shouldn't trust an organisation which has a vested interest in Sun's doing badly (or at least its share-price doing badly) for information on how well Sun is doing.
That said, I hope they can avoid doing a "novell" - great product, reasonable price, rotten marketing.
There is certainly a perception that Solaris is more difficult than Linux. Its probably for the same reason Novell got squeezed by Windows - Linux has a better GUI. I know Solaris is a server OS, but it doesn't help that everyone who works with it needs a Windows box for all their productivity apps. It shouts "niche product" loud and clear. They should have begged Apple to port their GUI to it instead of BSD, but that's all history now.
There is also the support problem. It isn't linux vs solaris, its linux and solaris. Most support systems rely on customers buying, but not using their support, in order to make a profit. However, when the application is very simple, Linux is often easier to install and maintain than Solaris. No support contracts for the web-server farm (linux or solaris) will be forthcoming. That rather important database however probably will get a support contract. Thus, Sun ends up with all the problems and that reduces the profitability of the support contracts. Sun could raise the price of support, but that scares off the customers, and Sun doesn't have license sales to offset the support costs. It does have hardware sales but the constant pressure from x86 pricing makes life difficult there too and unix kit tends to have much longer lifespans than your standard pc.
I'd like to suggest Sun do something to make themselves more appealing to the lower end of the server market. I know it isn't that profitable, but it builds a pipeline for sales further up. I'd like to see some management tools that do for server systems what Apple has done for the desktop. Backup, mail, calendar, directory, file & print, webdav services are all available open-sourced. Talk to Novell about directory-based management and Apple about ease of use. It doesn't have to be all-powerful, just simple enough for relatively unskilled IT staff. I"m sure Apple would welcome some better server systems what they've got and it doesn't have to be Apple-only. Overpowered hardware covers a multitude of inefficencies. It doesn't have to be free, but it does have to be easier to use - and that isn't something which comes easily to Sun.
I hope things get better for Sun. I'd hate to see them acquired just to get get rid of them or be chopped up. They are one of the very few innovators in the server space.
RE: What Really Scared IBM During Talks....
All very nice, but there is just one tiny, little roadblock sitting slap-bang in the middle of your sunny little software-only future - it's called Sun. You see, sitting in the software side of Sun means you think it is nice and easy to offload that hardware bizz and sail on blissfully. Problem is, Wall Street has been asking McNeedy and Ponytail to do that for years, and they didn't listen. They still won't. They are insisting on selling IBM the whole albatross. They have been trying to do the same to all and sundry for over a year without success. And it epitomises the Sun ego - they can't see that the whole deal is just not wanted by anyone.
The CMT server boys are just like you - they think their bit of the bizz is super-duper snazzy tech and should sail on without the rest of the company "dragging it back". I'm sure if you asked the storage side they'd say something along similar lines, maybe just with a different part of the company as their scapegoat. Here's the reality - you may all think you're not the problem, but unless McNeedy and Ponytail can break the company up and sell it in bits, it doesn't look like there's going to be any sailing off into the sunset for anyone. And that would be such a firesale that SAM might get worried they won't make their investment back and veto it anyway.
And even then, your appraisal of Sun's software bizz as the bee's knees just doesn't line up with the market's. As an example, Sun's own Galaxy boxes go out more often with Linux than Slowaris - by a ratio of five-to-one! Your "perfect stack" still has, for example, massive management holes. Hiding behind application numbers on old SPARC Slowaris just makes users laugh - it doesn't matter how many apps you have but which ones. If Linux has that one or two apps your business requires then you don't really care about the thousand others you don't need. And Linux has been burrowing deep into the developer community for years, to the point where I now find plenty of apps that are out and supported with RedHat but don't have a Slowaris x86 port, just an old SPARC Slowaris one.
You want to make out Sun is the Grateful Dead, but for years they attacked the Linux community, acting more like a big old record company, whilst RedHat and Novell were the Grateful Dead types. Sun's sudden switch to "open source" is viewed by the community with the same distrust and contempt that P2Pers view Sony Music Entertainment, the renamed Sony BMG.
And now for the final eye-opener, blindfold-lifter or whatever. We competitively bench before we buy. We let the vendors and application vendors tune to their heart's content as long as it is in our environment and with our test data. So it is Sun's own techies that have failed with dTrace or whatever tool you want to fantasize about. I didn't coin the moniker "Slowaris", that was other customers long ago who were just as disappointed with Solaris. But it is the continued failing of Slowaris to measure up to the competition that makes the name just as fitting today. Your continued denials that anything else could be better than Sun just reminds me of the old "Solaris on SPARC" mantra. Sun fooled themselves into thinking it was true, ignored the market, and got you into the mess you're in now.
"And one last point - to the poster using the term 'Slowaris'..."
Oh, never mind him. I think an E10K fell on some particularly sensitive part of his anatomy around the dot com bubble burst, and he's never been the same since.
Thanks for posting BTW (assuming you are who you claim to be).
RedHat only does Linux and they have a bunch of OpenSource developers doing the work for free. Even if you log a call it will end up on some Linux Q&A board. Now that is a business model. You sell other peoples intellectual property and do not have to pay for it.
Sun - has stack of salaried developers working on open sourced products which is given away for free. Now that is NOT a business model. How do you monetize that?
- Hp-Axed is dead - the axe have been swung - guess you need to get to open source Unix variant skills up to date. They have a business model for ignoramuses like you.
- you still do the Competitive Benchmarking? You still believe the TPC-C figures? The tooth fairy still visits and look in your shoe? The Easter bunny left you a big chocolate egg? HP's Unix division is making money?
RE: AC @Kevin Hutchinson
Besides your fallacies about RedHat (they have a team of engineers doing their own coding, they are not just absorbing the community's work for free), you seem to have missed a few facts that have been glaringly obvious to the rest of the market for a while.
"......- Hp-Axed is dead - the axe have been swung - guess you need to get to open source Unix variant skills up to date....." Assuming you mean hp-ux, I've reminded you Sunshiners before that would be the leading enterprise high-end UNIX going by the market figures. And that would be the nice, profitable, high-services-attach end, would it not? So, just to make it clear - hp-ux is the UNIX of choice for all those business-critical solutions where reliability, performance and assured support are key. Doesn't sound very dead to me. Want to suggest why they didn't chose Slowaris? And before you start frothing about "I know what you'd say", try actually thinking about what the customers are saying, as Sun losing share in the high-end is a very clear message.
".... They have a business model for ignoramuses like you...." Actually, they seem to have a business model that suits most of the market, seeing as hp are also the leading Linux server vendor, leading Windows server vendor, etc, etc. Name one area Sun leads in, other than major diving stock value? As to my skills, I use RedHat on ProLiant, Integrity and xSeries; Windows on Integrity, ProLiant and xSeries; hp-ux on Integrity; and EVA, MSA and XP storage (I don't touch our legacy EMC and IBM storage if I can help it!). I also play with the few remaining Slowaris boxes (currently all old UltraSPARC) when the resident Sunshiners are unavailable, and annoy the pSeries admins every now and again. I have benchmarked more modern Sun kit such as the M-series, Niagara and Galaxy kit which means I also have experience with the latest Sun kit and latest Slowaris versions - and yes, it still is slow and/or more expensive when compared to RedHat, AIX, Windows or hp-ux. I'd say my future seems a lot more flexible than your continued and blind worshipping at the altar of Sun.
".....- you still do the Competitive Benchmarking? You still believe the TPC-C figures?..." Well, if we are doing competitive benchmarks then that suggests we don't use the TPC-C figures as anything other than a guide. Do you even realise how contradictory your own rants are before you assault your keyboard? Ar you suggesting we shouldn't do competitive benchmarks, just take your word for it that Sun are the best just because you present such a stunning technical appraisal? Yeah, right! I really hope you have a ton of Sun stock, so much that the loss means you have to sell your PC and stop polluting the Web with your squealing.
".....The tooth fairy still visits and look in your shoe? The Easter bunny left you a big chocolate egg?...." Whoa! Have you been at the lightning water, chief? I know you Sunshiners spend a lot of time in a fantasy world of denial, I suppose you're just looking for a set of mythical creations to replace Sun's vapourware now that Sun is so obviouly dead.
".....HP's Unix division is making money?" At last! Some reality is seeping through the blindfold. I'm told that hp Support has been a spearate enterprise run as a for-profit business, not a loss-leader, since around 1999. And hp-ux has been profitable for hp for all those years since the dot-com crash, when Sun hasn't been able to make a profit on anything (well, they did make their red ink supplier a fortune!).
I suggest you spend less time ranting and more time looking after your own future, now that the Sun is setting. Would you like me to post that link to the RedHat newbie training page again?
/still doing the pointing and laughing thing!
That's the funniest thing I've seen on here for a while.
Seriously though, I've decided that Matt doesn't work for HP, at least he doesn't do any actual work for HP.
Perhaps he's been on garden leave since the EDS merger?
I was just wondering BTW Matt because your posts seem to take up more web space on the Reg than even the frequent who's going to buy Sun articles...
RE: @Twatt Bryant
"That's the funniest thing I've seen on here for a while....." I'd say thanks, but then I'm not convinced you got the humour.
".....Seriously though, I've decided that Matt doesn't work for HP, at least he doesn't do any actual work for HP....." Blimey! One of the Sunshiners FINALLY got it! Well, maybe.
".....Perhaps he's been on garden leave since the EDS merger?...." Nope, looks like that blindfold is still firmly in place. I can assure you, I have never worked for EDS. I have suffered their services before, but that was around SPARC kit.
".....I was just wondering BTW Matt because your posts seem to take up more web space on the Reg than even the frequent who's going to buy Sun articles..." Well, you Sunshiners do present miriad opportunities for humour! You may have noticed there is more than one Sunshiner posting The Great Truth from The Altar of Slowaris here, between you there is plenty of FUD, myths and outright lies to expose to the readers. Hey, you guys made your own cross, now bear it! If you hadn't spent years seriosuly annoying me with your FUD I might just have ignored you, but now it's just fun watching you run around screaming while your little fantasy world collapses around your ears.
Twatt Bryant ~~~ ".....Seriously though, I've decided that Matt doesn't work for HP, at least he doesn't do any actual work for HP....."
Matt Bryant ~~~ Blimey! One of the Sunshiners FINALLY got it!
No joke! Matt does not have a job!
By the way, Matt Bryant, the unemployed - you spelled Solaris wrong again.