Brian Gammage

Gartner Lead Microsoft Analyst Neil MacDonald explores the possibility of using Windows 7 Ultimate instead of buying SA on his blog: http://blogs.gartner.com/neil_macdonald/2009/10/28/windows-7-ultimate-sort-of/

The following factors are expected to drive adoption of Windows 7 in the market:
• Quick transition between Windows Vista and Windows 7 availability on new PC shipments for consumers and small business buyers. We expect that PC vendors and retailers will be managing their product inventory well to ensure that products with pre-loaded Windows Vista will be sold out in the first 4 weeks after Windows 7’s release across mature markets. In emerging markets it might take 1-3 months after Windows 7 release, due to variation of Windows 7 release dates across selected countries and slightly longer channel inventory time frames.
• Microsoft will also offer free upgrade program to Windows 7. PC buyers that will purchase their PCs between 26th June 2009 and 11th January will get a Windows Upgrade Option, limited to 25 PCs/organization. That should have helped avoid the dip in PC sales in back-to-school PC sales in 2009.
• The negative image of Windows Vista among many business companies due to application and hardware compatibility issues will result in many companies that have not started deploying Vista, will skip it and move straight to Windows 7. The delay in Windows Vista deployments in 2009/10 and increase in number of companies planning to skip Windows Vista will lead to a major boost in Windows 7 enterprise deployments from middle of 2011.
• For organizations that have remediated applications for Vista or deployed some Vista, relatively easy migration due to reduction in the application and hardware compatibility issues for users migrating from Windows Vista to Windows 7.
• Economic weakness resulting in delayed PC purchases in 2008 and 2009.
• Limited ISV support for Windows XP after 2011 will force organizations to move to Windows 7.

However, the following factors could inhibit the adoption of Windows 7 in the market:
• The current economic downturn is expected to lead to decline of WW PC shipment in 2009 (both consumer and professional), thus negatively impacting the overall Windows 7 license sales in the first 6-12 months after its release.
• Mainstream adoption by enterprise companies will not start until 2011, as it usually takes 12-18 months after new OS launch for the companies to prepare for OS deployment and ensure ISV support.
• In the current situation when IT budgets are being cut, many enterprise companies might delay tests of new OS thus delaying the start of Windows 7 deployment time until 1H 2011.
• Windows XP will be supported (even it is in the extended support phase) until early 2014, thus some companies might not feel no need to rush Windows 7 deployment.
• While the compatibility between application and hardware working on Vista and Windows 7 should be significantly improved, it does mean that the migration from Windows XP to Windows 7 will be easer than from Windows XP to Vista. The majority of enterprise PCs are running Windows XP, so companies will still need to spend as much time testing their application and hardware to migrate to Windows 7 from XP, as move from Windows XP to Vista.
• Expected strong growth in alternative client computing architectures, like Hosted Virtual Desktops (HVD) and Server Based Computing, will reduce the potential traditional professional desktop market that run Windows OS. Many organizations looking at HVD will deploy Window 7 with that technology, which will not be mainstream until 2011/12.

So Windows 7 is finally upon us and most organizations already know quite a lot about it: they know it’s the same generation technology as Windows Vista, that’s it has a smaller overall footprint, that it has many small improvements that will make it easier to adopt and use. In effect, that it’s like Windows Vista without some of the “hard edges” that deterred its adoption. They know about it because they’ve already planned for it – in a majority of cases, because they knew when they decided to skip Windows Vista that they were making migration to Windows 7 a necessity. For a majority of organizations, Windows 7 is the next major step on their client computing roadmap.

Many organizations are now building client computing roadmaps that look five or more years ahead. They are doing so because technology changes are creating more options for delivering applications to users and they need direction to navigate through the new range of choices. They are also doing so because they recognize their desktop real estate as a critical asset that defines end user capabilities and the per-user cost of IT. Client computing assets consume a big slice of most IT budgets and many organizations are stepping back to take a more strategic view of their current and future investments. This change is significant – it represents a shift from in the balance of market power between suppliers and buyers. Organizations are no longer prepared to navigate supplier roadmaps without question – they are building their own roadmaps so suppliers can navigate theirs. Windows Vista was part victim, part instigator of this change. Vista’s low marginal benefit and high cost of migration motivated many desktop planners to raise their eyes from the road and look further ahead.

The problem with roadmaps is that they demand more information about future products than most suppliers want to offer. To make the best next step, you must know where you will want to go afterwards. For organizations planning Windows 7 deployments, that means looking beyond – to Windows 8 or whatever it will eventually be called. Microsoft has not been able to support the requirement; it is still fully occupied communicating about Windows 7 and is unlikely to begin talking about what comes next until Windows 7 adoption begins. Thankfully, we at Gartner have no such constraints – we have been telling customers what we think will come next for some time.

Our views of Windows 8 (let’s call it that for convenience) were published in February 2009. We built these on three basic assumptions:
• Microsoft will not stop innovating and adding new capabilities to Windows
• Backwards compatibility with previous Windows versions is non-negotiable
• Windows cannot keep getting “bigger”, because scale and complexity have become significant barriers to adoption and hence a problem for Microsoft (which, after all, wants to keep selling new products).

These three assumptions point to an end of the road for the monolithic architecture of the Windows OS – in which everything resides and is interlinked in the same bubble. Instead, we believe Microsoft has to separate the compatibility components from the new, delivering each in its own bubble. The technology to support these on a single device already exists – its called a privileged hypervisor and Microsoft is already on the way to integrating one of those into its Windows Server OS. Microsoft also has the technology to create a single, seamless user interface that spans two separate computing “bubbles” in its desktop virtualization product, MEDV.

So, does Windows 7 conform to this vision? So far, yes. Windows 7 is more virtualizable than Windows Vista, so would be easier to package as the “compatibility” bubble in a 2-part OS. Meanwhile, changes in the way Microsoft delivers products that help use the OS show the Windows business model is evolving too. With more enterprise-focused capabilities delivered through SA and MDOP, subscription services to use Windows are gradually becoming essential for most organizations. This is important, because while we expect the compatibility piece will continue to be delivered as a preloaded “product”, the newer functionality will likely be delivered online as a service.

So what does this tell you about Windows 7 and beyond?
• Windows 8 will come as two “execution bubbles”, one for compatibility functions and the other for new
• The compatibility piece will almost certainly be Windows 7
• The new piece will be delivered online/over the air (it may even converge with Windows Mobile)

In other words, there’s good news and bad news. Good – your investments in Windows 7 will have longer lifespan than usual because a near identical bubble will get an extended support life with Windows 8. Bad – organizations that don’t pay for SA or an enterprise agreement today will have to pay more.

The future of the Windows OS may be set to get simpler, but no one said it would cost less.

Since receiving the first Community Technology Preview (CTP) version of Windows 7 last October, I’ve installed various versions of the product about 30 times across 15 different systems.  The systems in question have run the gamut from netbooks and an Atom based net-top, through several variations of corporate and consumer notebooks, a couple of branded consumer and corporate desktops, and up to a high-end home-brew gaming system.  The installations have been a mix of 32-bit and 64-bit, on processors from Intel, AMD and even VIA, and have used both discrete  graphics and integrated graphics.  The systems have included a seven year old single core Pentium 4, right up to a dual processor, 8 core behemoth, with the most recent being a 64-bit installation on a high end Core i7 system that has just become my new office desktop. I’ve had the opportunity to see the installation process across a very wide range of scenarios including many of the ones our customers are likely to face. 
 
My experience, even from the first tentative install last October, has been pretty good.  Sure there have been a few missing drivers here and there,  but in every case the resulting machine worked and almost always performed  better than under Vista AND XP.
 
 Until one install with the final code in late August.   That was the day I decided that it was time to upgrade one particularly critical machine we use around the house. I had gotten so comfortable with the install that I clicked right through the warning messages that the installer routine plastered on the screen.  Had I read those messages about issues with the graphics card a little more closely, I would have turned back, but alas I didn’t.   I won’t go into all the gory details, but the result was painful and unpleasant.  The  warnings would have alerted me to the fact that the particular graphics card would only work with reduced functionality due to its vintage.  The issue was related to a quirk of my particular environment and not something that a typical user would ever see.  Once I realized I had a problem, my hasty attempt to fix the problem (again ignoring a warning message posted by the graphics card vendor), just dug the hole deeper.  
 
Why do I bring this up?  Well it is awfully easy to get so lulled into a sense of complacency that we forget to do the basics.  Most Windows 7 experiences will be great, but that doesn’t mean we can ignore the details and bypass normal due diligence.  For individuals this means running Microsoft’s compatibility checker BEFORE starting an upgrade (and actually paying attention to what it might tell you).  For organizations, this means conducting a formal evaluation and testing process.  Check out the note I published on preparing for Windows 7 in September.
 
By the way,  the problem was eventually resolved by replacing the offending graphics card with an inexpensive new card that actually worked far better than the original one.  I wish I would have replaced the card sooner.

Windows 7 is based on the same technology generation as Windows Vista – which despite talk of layers (more than 60) and components to decouple some components, was a very large, complex and highly integrated bubble of software. When presenting at conferences, I often compare the highly integrated nature of the Windows OS with badly cooked spaghetti – everything is stuck together. This isn’t pejorative, it’s just a reflection of where we are and audiences usually agree.

We are all victims of history, especially when it comes to key infrastructure components like Windows. In the mid 1990s, the main challenge in client computing was getting applications to work with diverse hardware. To overcome this, we built big OSs to integrate shared functions and connect the applications layer with the hardware one. Now our main challenge in client computing appears to be managing the big OSs we’ve created. Food for thought there as we plan future developments and deployments (not only of Windows, but also of all the infrastructure we use for the care and feeding of our client computing environments).

Windows 7 is necessarily more of the same, but also promises to be an important step in a better direction. It is much easier to virtualize, which means it will be an easier “bubble” to package and deliver. Virtualization decouples the OS from the layers above (applications) and below (hardware), making its configuration more tolerant of changes in their configurations. For previous versions of the Windows OS, the latter was challenging in that many functions were designed to be closely coupled to hardware. Some instructions expect to be handled directly, not by an intermediate layer of software (which might also want to share access with another OS). Windows Vista in particular presented issues when virtualized, making it a particularly bad candidate for hosted virtual desktop (HVD) deployments.

Although the changes will remain invisible to most users, Microsoft has clearly done work on the underside of Windows 7, where it meets hardware. To support XP Mode, they have extended support for hardware capabilities that enable and support a privileged hypervisor. Although Windows 7 has not moved down to occupy these new privilege levels and does not ship with a hypervisor, the extended support means that Windows 7 works with the VT instruction set extensions – in other words, the instructions that previously need to go direct to hardware can now be intercepted by an intermediate layer.

Will these changes make a big difference to you? Unlikely, unless you want to run Windows 7 in an HVD (where you will find its small footprint helps too). But they point to a much more modular future in which Windows is no longer tied to hardware and can be moved more freely from device to device. They also support our expectations for Windows 8.

So changes under the hood, but then OS architecture changes should be invisible. As we battle the complexity of the client OS, our objective should be to make the OS as unobtrusive as possible. Despite this, these changes are fundamental. Microsoft has cut some of the chains of complexity – it has started to unbundled the badly cooked spaghetti for enterprise users.

Today’s guest blogger has his own blog. Read David Smith’s thoughts on using consumer SKUs in business at:
http://blogs.gartner.com/david_m_smith/

I am Annette’s Jump, Research Director in Gartner’s WW Client Computing Team, and I’m today’s guest blogger.  I cover client operating systems on PCs (Desktop PCs, Mobile PCs and Mini-Notebooks) and I am the lead analyst behind the worldwide operating system forecast. My other areas of expertise are the mature and emerging PC markets in Europe, the Middle East and Africa.

Gartner does not expect that Windows 7’s release in October 2009 will have a major boost on WW PC sales in the fourth quarter of this year. Sales should be in line with what we have seen during Windows Vista introduction in the market in early 2007. It is true that we might see slightly weaker sales in early October, followed by slightly stronger sales in November, but for the fourth quarter overall we believe that the net impact on WW PC sales will be very minimal.

Additionally, Microsoft is offering an upgrade program to Windows 7 from Vista for PCs sold starting 26th June 2009. That should help to avoid any major decline in PC sales in the 1-2 months prior to new OS launch. However, users will have to do the upgrade themselves and it is not something that most consumers or small businesses without IT departments would have done before. Additionally, upgrading the client OS when you already have some personal data on a PC is not that straight forward and definitely more complicated on a Windows PC vs. Mac OS.

Windows 7, with its polished user nterface and several new consumer-friendly features, will likely reduce the gap in perception between Windows and Mac OS. However, Apple is able to innovate on a shorter time frame and they already released their update for the Mac OS X – Snow Leopard – in September 2009. The upgradeability of Mac OS is still easier and significantly more users do that on Macs vs. PCs. To encourage more Mac users to upgrade to Snow Leopard and to attract some of the traditional PC users Apple also announced big reductions in upgrade price to only $29. In its announced upgrade pricing Microsoft is planning to reduce by approximately 10% (depending on the market) the upgrade price of Home premium version, so the upgrade cost to Windows 7 will range from $119.99 to $219.99.

Additionally, Microsoft announced that for a limited period of time in June and July in a few selected countries users were able to pre-order Windows 7 at nearly 50 percent discount ($49.99 for Windows 7 Home Premium and $99.99 for Windows 7 Professional). The later is likely to be Microsoft response to Apple announcement of the special upgrade price to next Mac OS, but even with the special pre-order program Windows 7 upgrade pricing will be much higher vs. Mac OS upgrade price for ‘Snow Leopard’. Furthermore the upgrade prices in Europe are significantly higher than those in the US, which will mean that fewer European consumers will buy the upgrade to Windows 7.

Back in 2003, AMD released the Athlon 64, the first x86 processor that ran in 64-bit mode. This was followed in early 2005 Intel’s Pentium 4 660 which also supported 64-bit extensions to the x86 architecture. However, the market has pretty much ignored 64-bit operation and except for a few fringe users (like me*), it has mostly remained a curiosity. Well at least it had, until the summer of 2008, when most of the PC makers began converting a large portion of their North American consumer desktops and notebooks to 64-bit versions of Vista. The reasons for this are rooted in marketing rather than technology, but nevertheless, it marked a turning point. Last fall, I received several panicked calls from vendors who had been caught off-guard by the switch and suddenly found their support lines jammed with calls from irate consumers wondering why a product wouldn’t work on the brand-new PC they had just brought home from the store.

With a Windows 7 migration looming, it’s time for corporate customers to do some thinking about 64-bit. I’ve recently published a note detailing some of the factors organizations must consider. The bottom line is that at some point in the next 5 years, you very likely will be moving to a 64-bit operating environment and Windows 7 may be the right time to make the move. At the very least everyone should include one 64-bit environment in their testing matrix. While it may not be the right time to make the move, it is certainly the right time to start preparing for the inevitable 64-bit shift.

*I first started running 64-bit Vista in production nearly 3 years ago for my everyday desktop, and am well on the way to converting all of my production notebooks and desktop systems to 64-bit Windows 7. So far I haven’t hit a showstopper except for out corporate VPN. But even that issue has been resolved and I’ve been successfully running a beta VPN client for several months.

Editor’s note: In our 13 October Webcast, we asked our 400 attendees about their plans for 64-bit deployment. The results show lots of confusion and uncertainty.

Will you deploy Windows 7 in 32- or 64-bit?

We will deploy mostly 32-bit       34%

We will deploy mostly 64-bit       13%

Undecided, leaning 32-bit           25%

Undecided, leaning 64-bit           28%

Welcome to the Windows 7 blog. Over the next few weeks, we  will be blogging some thoughts about Windows 7 and we’re going to use Brian’s blog to capture it all. I’m Michael Silver, lead analyst on client operating systems for Gartner, and I’m the first of a number of guest bloggers. We invite you to take advantage of all the research we’ve done on this topic, schedule inquiries with the appropriate Gartner analysts to help apply our research to your specific situation, and check this blog several times a week during October for more thoughts.

For Microsoft, Windows 7 is the most important release of Windows ever. Sure, there were other times when Microsoft had important releases of Windows. Windows 3.0 was a breakthrough release of the product and really was responsible for Microsoft capturing the position it eventually did in the market. Windows 95 further cemented that position as it finally broke the 640K DOS memory limit and consumers waited on lines that stretched around the block of mid-90’s computer stores. Windows XP was an important release as Microsoft officially ended the Windows 9x architecture (and had to make up for the dismal Windows ME) and successfully transitioned a huge consumer installed base to the Windows NT installed base. And Windows XP SP2 moved to put the widespread security problems behind Microsoft and revitalize its reputation for improved quality.

But Windows Vista was by far, the worst release of its flagship product. After years of delay, Microsoft scrambled to get a product out the door. But its huge ecosystem largely abandoned it and machines purchased after release had flaky drivers and many were underconfigured to keep PC prices down and volumes up. The product improved over the last three years, but its reputation never did. The weak economy meant that organizations were buying fewer new PCs, so organizations that had been predisposed to bring in new PCs with Windows Vista didn’t bring in too many new PCs and it did not make sense to bring in the ones they did with the new OS. The Windows ecosystem dragged its feet in supplying Vista-compatible software. The failure of Vista was compounded by a successful and unanswered attack by Apple, which made belittled Vista (and Microsoft).

Over this past summer, many people compared Apple’s Snow Leopard launch with Microsoft’s for Windows 7. In truth, they are nothing alike. Apple’s new release included many small improvements, but in the best case scenario, a successful Apple product would mean a continued, steady increase in their sales and if the product turned out to be a total disaster, it probably would not have impaired their sales by that much. Conversely, a second consecutive bad outing for Microsoft would probably be enough to shake the confidence of the market and could potentially lead to the decline of the company.

The initial feedback on Windows 7 has been excellent. The OS feels solid and performance is significantly improved. It is even winning some fans from the Apple Mac OS fan base, although there are few fan bases that are as loyal and vocal as Apple’s. Feedback from clients indicates that their preparations to deploy Windows 7 are already under way and we expect the new OS to be an extremely successful product.

However, Microsoft is not out of the woods yet. While the majority of applications installed in enterprises today require Windows, Microsoft face an enterprise software market which is increasingly moving to OS-neutral applications, and a consumer market that is embracing products from Apple, embracing devices other than traditional PCs, and having more influence over the technologies used in the workplace.

While Microsoft is trying to distance Windows 7 from Vista, Windows 7 is very much Windows Vista under the covers. It builds on the architectural improvements of Windows Vista and indeed, there could not be a Windows 7 had their never been a Windows Vista. That’s another problem, though – how can Microsoft release a version of Windows with major architectural changes and make it easy to absorb. Or perhaps the better question is whether that’s an attainable goal. But that’s a topic for another blog…

The IT headline of the week was Google’s announcement that it plans to deliver a new PC Operating System (OS) based on its Chrome browser, which engendered raging debate over the imminent death of Windows and the ultimate victory of cloud computing. However, such arguments only capture part of the picture.  

Yes, Windows will have to get smaller. And yes, many applications will move to the cloud. But the real news is that ties between applications and the OS are also slowly diminishing. New applications are increasingly OS-agnostic, particularly if web-based. At the same time, packaging technologies like virtualization are reducing the direct value and influence of the OS in supporting applications across different hardware. 

Many, if not most PC applications run under Windows and that will keep most customers from switching OS overnight. This is particularly true for businesses, who are unlikely to pay the price of upgrading a broad application portfolio only so as to use a different OS.

However, the PC OS is already set to get smaller. Windows 7 will present a smaller “footprint” than Windows Vista and we expect the releases that follow Windows 7 to change architectural direction. As the OS diminishes in scale, it will likely also diminish in perceived value and visibility. 

In this context, Google’s announcement is much more than just competition – it is affirmation that we are at the end of the dynasty of the big PC OS. The future PC OS must be thinner and more interchangeable (Microsoft needs this too), which naturally also opens the door to competition. The PC OS will become more commoditized, but will we really care? In the end, an OS is only there to run applications – it is a means to an end, not the end itself. Perhaps the best thing we can aspire to say of an OS is that we don’t know its there.