Kleynhans reckons that companies need to spend three months on information gathering so they know how many people, PCs and applications they have to migrate. After that, he's budgeting only six to nine months for engineering work to correct any software incompatibilities, followed by a three-month pilot test that will cover weekly, monthly and quarterly processes. "That leaves 18 months for roll-out."
The cut-off date is the end of support for Microsoft Windows XP in April 2014, but Kleynhans said companies should "aim to get off XP" by the end of 2012. He warned that "lack of ISV [independent software vendor] support will start to become common" in 2012 and "you will start to feel some pain."
Gartner slide: Windows 7 migration timeline
The move to Windows 7 could also involve companies moving to Microsoft Office 2010. In this particular webinar, 59% of the participants were still on Office 2003 or earlier, with 41% on Office 2007. All the other options -- Office 2010, Google Docs and Open Office -- had zero users. "For the most part, it's not really practical" to move to Google Docs or Open Office, said Kleynhans, "particularly when you look at the fact that [Microsoft Office] is a platform, not just an application. It's pretty difficult to extricate yourself."
Doing both migrations at the same time could be tricky. Kleynhans said many companies liked to take a mixed approach to the Windows upgrade -- swapping some old PCs for new ones while upgrading some current models -- but usually wanted to switch everyone to the new version of Office as quickly as possible.
Kleynhans added that companies needed to upgrade to Exchange 2010 and SharePoint 2010 to make full use of Office 2010, and that they should also look at the corporate versions of Microsoft's web-based Office apps.
For many companies, the migration to Windows 7 also includes upgrading from IE6 to IE8, and moving from 32-bit to 64-bit computing. These can involve significant remediation for those who skipped or didn't even test their applications on Windows Vista. "We would generally recommend you plan to go to 64-bit," Kleynhans said. "There's about a 60:40 split at this point," with most companies going to 64-bit.
Although some companies may still seem slow to respond, Kleynhans said "we've seen a very positive response" to Windows 7. "People really are moving quite quickly, compared to previous operating systems. It is really an inevitable migration."
The Big Migration: Windows 7 and Office 2010 webinar was held on 30 March 2011.
@jackschofield
Twitter 
Linkedin 
ZDNet UK App 
Facebook 
RSS Feeds 
Newsletters 
Talkback
Migrating from XP to 7 is a huge leap, and involves more than just planning by an engineering team to identify incompatible applications. It will also require the execution phase which can steal a large amount of resources from a company. There will almost certainly need to be a budget set or allocated to cover all of the costs of the migration. Machines purchased with XP (before Vista) will either need extra RAM or they will have to be scrapped and replaced with new PCs. Then, there's the issue of re-purchasing incompatible software as well. These costs can go up exponentially at the drop of a hat as items are discovered.
I've been through the migration in an enterprise environment, and it's a bumpy ride. There needs to be thorough application and hardware testing, as well as preparation to spend a large amount of money to cover time, hardware and software purchases. Some companies might not be ready for this, so I can definitely understand why some are dragging their feet. It's a huge hurdle to overcome.
@apexwm
Yes, thanks for the thoughts, which, I think, re-inforce Gartner's arguments.
> Machines purchased with XP (before Vista) will either need extra RAM
> or they will have to be scrapped and replaced with new PCs.
Vista was launched on 29 Jan 2007, so those machines are already more than 4 years old, so I hope they've been amortized by now. By the time of a last-ditch Windows 7 roll-out they would be 7 years old. Anybody not planning to replace them then is probably incompetent and could well be costing their company money.
> Some companies might not be ready for this, so I can definitely understand why
> some are dragging their feet. It's a huge hurdle to overcome.
Yes, especially if companies are struggling in a recession. However, by not upgrading sooner, they have increased their operating costs by staying on XP, and they may well face higher migration costs through having to do it in a constrained time-frame. They will also have made staff use an inferior operating system for 1-3 years after a superior alternative became available, which must be particularly frustrating for the ones already using Windows 7 at home ;-)
You can put off migrating, but it's like underpaying your mortgage for the first decade. You feel better off in the short term but you pay for it in the long run.
>>It is really an inevitable migration (Gartner quote)
No, it's not. Companies can stay right where they are (XP) for a good number of years yet. I've worked at companies that did this; one company that I worked at was still using Mac OS 8 in 2006, and the world didn't stop turning. Rumour has it that there are alternatives to the Windows/Office monopolies too.
@Apexwm
> I've been through the migration in an enterprise environment, and it's a bumpy ride
And what exactly did the company gain from this "bumpy ride"? What was their Return On Investment (ROI)?
> They will also have made staff use an inferior operating system for
> 1-3 years after a superior alternative became available, which must
> be particularly frustrating for the ones already using Windows 7 at home ;-)
Blah, blah, blah.. says you. Personally, I do a lot of things at home that I don't get to do at work. I do live in hope though! ;-)
The PDF on the Gartner link is a real hoot! One slide lists the reason that why you need to move to Windows 7 as these:
* AppLocker
* BitLocker To Go
* BranchCache
* Better UAC
* Updated UI
* HomeGroups
Hardly screams "must have" does it? No wonder that same slide ends with "Because you have to!"
Another slide is titled "Windows XP: ISV Support Will Wane Before Microsoft's". Errmmm... really? Are there any examples of big gun apps that have abandoned XP thus far?
Well, there is *one* example, of course; and inevitably, it's from Microsoft itself. Internet Explorer 9 requires Windows 7 (or Vista). End result? Firefox 4 wiped the floor with it in terms of number of downloads. And Firefox 4 *does* run on XP, not to mention non-Microsoft platforms.
LOL what? Gartner is a clueless organization that keeps on making stupid "unique" recommendations based on what the overall tech world says about a product. Windows 7 is just like Vista rehashed. Full of gimmicks and fancy tricks. Aero Snap? There's a far better version in XP called Tile Horizontally or Tile Vertically that isn't limited to arranging just two windows but any number you select. There are many good useful features of XP removed and broken in Windows 7. The file manager, Windows Explorer was utterly destroyed in Vista and becomes worse in Windows 7. Poor usability. See www.en.wikipedia.org/wiki/List_of_features_removed_in_Windows_7 and www.en.wikipedia.org/wiki/List_of_features_removed_in_Windows_Vista . Unnecessary GUI changes. Vista was innonative but horrible usability wise and removed things. Windows 7 is Vista with few new features and again many features removed and fancy gimmicks and shiny graphics added. Sure it's more secure and XP is also *secure enough*. There is no reason at all to migrate off it.
The difference now is that XP is running out of support and if you're a business, you don't want to be running on software you can't get support for, for compliance reasons. Thousands more group policies, dramatically improved security and the fact that you can, if you want, run Windows 7 on the same hardware as XP are appealing to many businesses, but AppLocker, BitLocker and BranchCache have significant business benefits. Or you could just stick with your penny-farthing ;-)
MB
@Mary,
Windows 7's "dramatically improved security" didn't count for much when the Conficker worm came a knocking, did it?
TrueCrypt can do everything that BitLocker does, but doesn't cost the Earth. It's free, in fact, and cross-platform to boot.
I'll stick with my penny-farthings of OS X and Ubuntu, where I have any choice in the matter.
Depends entirely on the business and what they use their computers for.
We still run Win98 on one computer (actually it dual boots with debian) It never sees the Internet so from that point of view is quite safe. However it is very important. It is the only machine that can run some ancient software needed to support some of our industry customers.
We also have 1 computer running XP - needed for Sage. That will only be changed when either the machine dies or Sage upgrades will no longer work - not that Sage is a particularly good piece of software. Actually it's crap!
The other machines have all been upgraded... to Ubuntu.
Now I don't say we are typical. In fact, clearly we are not, but I bet there are a very large number of SMEs out there that have absolutely no need to change, or if they do, are likely to think about what would be the smallest change they could make. Breaking away from Microsoft is starting to look very attractive to these people.
@BrownieBoy
> Conficker worm came a knocking, did it?
Conficker exploited a hole that had already been patched. No OS protects from the incompetence of the IT department.
> TrueCrypt can do everything that BitLocker does
Interesting, so it now supports TPM? Also, how does it work with Active Directory and Group Policies?
@Jack,
The TPM you say? As in the "Trusted Platform Module"? As in, Hollywood and the software industry's attempt to reclaim PCs from pesky users putting they want on them?
And you thought that Microsoft's "Genuine Advantage" was bad enough, kids? Well, the TPM is hundreds of times worse. (Remember, you don't buy software you only rent it.)
No, I should think that TrueCrypt would not be compatible with such a thing. But hang on, let's see what the TrueCrypt FAQ says:
"The only thing that TPM is almost guaranteed to provide is a false sense of security (even the name itself, "Trusted Platform Module", is misleading and creates a false sense of security). As for real security, TPM is actually redundant (and implementing redundant features is usually a way to create so-called bloatware). Features like this are sometimes referred to as 'security theater'"
Looks like you're right, Jack!
@BrownieBoy
> Looks like you're right, Jack!
Thanks, I usually am. Otherwise, I enjoy your touching faith in the TrueCrypt FAQ explaining why TrueCrypt lacks TPM support.
We've already established that it's wrong to claim that "TrueCrypt can do everything that BitLocker does", but I'd still be interested in this bit:
> Also, how does it work with Active Directory and Group Policies?
By the way, I have nothing against TrueCrypt. I have recommended it in the past, and will no doubt do so again.
BrownieBoy :
"And what exactly did the company gain from this "bumpy ride"? What was their Return On Investment (ROI)?"
That's a very good question, not much that I have seen. The migration involved throwing out about 40% of the existing hardware and a mass purchase of new PCs to replace them. It also involved re-buying up to date software and proprietary drivers that have amounted in another huge amount of money. The plusses of Windows 7 out of this is that it is newer, which brings some enhancements like better firewall protection, fast user switching in the domain, and easier proprietary software installations with UAC. But those are hardly worth the huge investment that went into the migration itself. In fact, the migration is still ongoing because of the high costs, and time needed to test and fix broken applications that don't work in 7. But what choice to companies have, the old OS is 10 years old and Microsoft is dropping support. The ultimate solution would have been to offer existing XP customers steep discounts along with a smooth upgrade path, and backwards compatibility. But Microsoft has not stepped up to the plate, and instead has leveraged its dominance in the market to force existing customers to pay almost full price, and migrate.
The problem is that Windows 7 and XP are vastly different, which is why it is called a "migration" and not an "upgrade".
I would love to hear of others that have attempted it, to compare stories and feedback.
It's worth bearing in mind though that Win7 allows companies to do stuff they couldn't do on earlier versions, and that its improved robustness should save money because of reduced calls to the helpdesk. But it's the new Group Policies that are probably near the top of many IT departments' must-have lists.
"It's worth bearing in mind though that Win7 allows companies to do stuff they couldn't do on earlier versions, and that its improved robustness should save money because of reduced calls to the helpdesk."
So far that has not been the experience. We've been witnessing the ugly sides of Windows 7 and in fact Help Desk volumes have been as high as with XP. Some repeating issues that frequently come up: Windows 7 boots into recovery mode at random and the repair fails but the PC boots normally if the user cancels the repair; offline files causing file corruption and loss of data within the local cached files; inability of IT to run explorer as an administrator while working on a user's PC (which was possible with XP -- now switch user needs to be used instead which doesn't allow for retaining the user's session while making changes in explorer which was nice before); and the list goes on. Maybe SP1 will address some of these, who knows.
I work for a company called ChangeBASE that has developed a product – AOK, to automate the testing and fixing of apps – particularly relevant to Windows 7 and 64-bit migrations
We have worked with over 1,000 companies with hundreds of thousands of apps and have some useful statistics on application compatibility problems
Across the board we see that c. 5% of apps moving to Windows 7 64 bit will have issues that stop them running properly or all. The most common issues are:-
• Non 64 bit drivers
• 16 bit files and API calls
There are additional issues but there are a lot less common. E.g.
• Virtual DOS machine functions
• Windowing extension features
• Direct X video acceleration API’s
In addition to these issues we find on average a further 60% of applications will have what we report as ‘amber’ which means they need some form of remediation to run 100% ok.
Most common problems are
• Legacy help files
• CPL files
• Hardcoded values
• Unnecessary UAC prompts
• Custom actions running in the wrong context
• Etc
Manually trying to find and then fix these issues can be very time consuming
John Tate:
I have also seen the issues you state and in fact the percentage of apps that have issues is closer to around 25% in my cases. It's clear that you have been in the real world in the middle of the migrations rather than simply write about them as some do. Even some of Microsoft's own products need to be re-purchased (like MOM 2005 which is imcompatible, to be replaced with SCOM).
I also did not mention another large side issue that we saw with remote support of company machines. For years they had used Cisco VPN client and RealVNC which was rock solid, no issues whatsoever and it provided an inexpensive and reliable solution. With the introduction of Windows 7, both of those programs had to be upgraded. UltraVNC replaced RealVNC, and Cisco Anyconnect client replaced Cisco VPN client. The result? Unreliable support. UltraVNC has issues with dropping connections randomly, the service has to be restarted multiple times for it to answer connections again, and Windows 7 firewall has threw some roadblocks because of its 3 profiles as opposed to single profile in XP. On top of this, Cisco Anyconnect client needs an extra add-on and requires a lot more setup to get working for the support tech to be able to log out and log in as other users while maintaining the VPN connection. Due to all of these issues combined, it was determined that the best solution would be to purchase a remote support appliance, which carried a hefty price tag. But, it was either that or continue with the free tools, which simply did not work with Windows 7 reliably enough for critical applications. Other free software was looked at, but the company simply didn't have the resources to waste more time in testing and troubleshooting for weeks and weeks. Surprises like this are inevitable which only complicate the migration issues even more.
@apexwm
> The problem is that Windows 7 and XP are vastly different, which is
> why it is called a "migration" and not an "upgrade".
Windows XP is 10 years old so it's hardly a shock that they're different. Windows 7 is actually a simple upgrade from Vista. Companies that chose to skip Vista created the migration problem themselves. It was their choice.
They could also have used Vista to prepare for the eventual migration, even if they chose to skip Vista. Again, it was their choice not to do the work, and they had *years* do to it. Blaming Microsoft at this stage is just stupid.
> But, it was either that or continue with the free tools, which simply did
> not work with Windows 7 reliably enough for critical applications.
This sounds like inadequate planning and testing. As I almost said earlier, nothing protects you from the incompetence of your IT department....
"Companies that chose to skip Vista created the migration problem themselves. It was their choice."
I would have to disagree there. Companies chose to skip Vista because it was a horrible operating system which even Microsoft admitted to. In my opinion, Microsoft created the migration problem because they didn't provide an entirely smooth transition or upgrade path. Ask any consultant about how well the Microsoft tools work in a Windows XP to 7 migration, and see if they use them or if they recommend 3rd party tools. Microsoft writes and controls the operating system, so the users are at their mercy.
"This sounds like inadequate planning and testing. As I almost said earlier, nothing protects you from the incompetence of your IT department...."
Testing was done and passed. But testing is only an attempt to identify issues before they happen, not prevent them entirely. Issues can pop up at any time and in this case, did, after thorough testing was done and Windows 7 was being deployed. The issues mentioned did not happen on every single PC, they were very sporadic. Sometimes it worked, sometimes it didn't.
@apexwm
> Companies chose to skip Vista because it was a horrible operating
> system which even Microsoft admitted to.
Actually, Vista wasn't that bad, though it was horribly late and it was let down by some third-party software houses and driver writers. Either way, you've missed the point, which is that even if companies didn't install Vista, the intelligent ones tested it with their apps and the *very* intelligent ones started remediating their in-house apps. So yes, it's their fault if they are in a mess now.
> Testing was done and passed. But testing is only an attempt to identify
> issues before they happen, not prevent them entirely. Issues can pop
> up at any time
As I said, correctly, nothing protects you from the incompetence of your IT department....
@BrownieBoy
> This is twaddle, even by your standards. As apexwm says,
Thanks but I already know you loathe Microsoft, and I assume that must bring some joy to your life. However, a bit of product knowledge might get you a bit further here. Are you still trolling the Guardian as well?
Regarding VISTA history there are some lessons to learn which are relevant to Win 7 - and other emerging platforms
Our experience is
1. VISTA was not that bad an operating system. In fact sorry if I am not joining the anti Microsoft brigade - but I had it installed on my desktop and liked it!
2. However as with many new platofrms (Citrix VDI, VMware Thin App, App-V and new broswrs such as IE8/9) many vendor and in house apps have application compatibility problems with the new operating system. So particularly for big companies they have a problem - where they have to manually test and fix hundreds or thousands or more apps.
It can take on average a day to test and app and 2-5 days to 'repackage' or fix it.
The scale of the project this creates is huge - where dozens of people or more may be needed to be dedicated to this task and carry out very detailed testing and coding work
3. Most of the large SI's offer services in this area and make a lot of money out of it. However all too often they do a poor job at this
4. Our product automates much of this work - but as is often the case with new technology there has been resistance to it's adoption. One obvious reason is that it reduces the profit margin of SI's. Another is that it can reduce the in house staff requirements for a new proejct - and so the people who often test our software stand to lose their jobs if they buy it!
5. Thankfully there are some SI's out there who realise that it is a good thing if you can offer your client a cheaper/quicker/better service!
"Actually, Vista wasn't that bad, though it was horribly late and it was let down by some third-party software houses and driver writers. Either way, you've missed the point, which is that even if companies didn't install Vista, the intelligent ones tested it with their apps and the *very* intelligent ones started remediating their in-house apps. So yes, it's their fault if they are in a mess now."
Well this might hold if businesses had an open checkbook and unlimited pool of resources. But we all know in the real world, these have limits. And, application testing, deployment and replacement with a new operating system is time consuming and costly. When businesses saw Vista and did their own research and saw what a failure it was becoming, I'm sure they pulled the plug on testing. At that time Windows 7 was only being talked about. Who would have known that Vista to 7 would be an upgrade vs. a migration? Also, you must consider applications that are purchased and not developed in house. Again, you are at the mercy of the vendor to re-write to fit the new operating system. Often when this happens, the vendor needs additional resources and passes this down to their consumer as an "upgrade" that the consumer must re-buy.
@apexwm
> And, application testing, deployment and replacement with a new
> operating system is time consuming and costly.
Indeed. That's exactly why the smart IT departments started looking at this stuff in the Vista time-frame and the stupid ones didn't. See ref above to underpaying your mortgage....
> Who would have known that Vista to 7 would be an upgrade vs. a migration?
Everybody with a working brain. Windows 7's development was discussed interminably from Thu, Aug 14 2008 with Windows' top management at http://blogs.msdn.com/b/e7/
Windows 7 then had a long beta period when companies could download it for testing.
The idea that somebody could be *surprised* by what's in Windows just beggars belief. Microsoft spends vast sums telling everybody to get their apps, drivers and hardware done well in advance so they'll be ready for the launch (which they rarely are). If your IT dept hasn't figured this stuff out more than a year after the launch then they absolutely are incompetent, and their incompetence is not Microsoft's fault.
> Again, you are at the mercy of the vendor to re-write to fit the new operating system.
But that's not unique to Windows, and in general, Microsoft has provided a remarkable amount of backwards compatibility. It has also provide a free XP box with the business versions of Windows 7.
Why don't you go find a Mac site and whinge about Apple instead? XP came out in the same time-frame as OS X 10.0 (Cheetah). Since then, Mac users would probably paid good money for Puma, Jaguar, Panther, Tiger, Leopard and Snow Leopard. If they were trying to jump from Cheetah to Snow Leopard today, they'd have found their whole Classic environment had disappeared and they'd need to switch from PowerPC to Intel hardware.
"> Who would have known that Vista to 7 would be an upgrade vs. a migration?
Everybody with a working brain. Windows 7's development was discussed interminably from Thu, Aug 14 2008 with Windows' top management at http://blogs.msdn.com/b/e7/
Windows 7 then had a long beta period when companies could download it for testing."
Running Windows 7 beta would have been an indication of what was to come, but definitely should not have been used as a decision maker until a final release was obtained. It's BETA and there were a lot of changes thrown in before 7 was officially released.
@apexwm
> Running Windows 7 beta would have been an indication of what was to come,
> but definitely should not have been used as a decision maker until a final release
What a bizarre comment: the beta is precisely where you start to make important decisions, and some "lead customers" actually install it in production. If you think you're supposed to write or remediate programs so that they only work in the beta, or only work in the final version, then no wonder you're confused. However, my main point was that the beta should be used to find out if you are going to have problems, where they are, and how you are going to deal with them.
Those of us who live in the real world also appreciate that all operating systems continue to change as patches and service packs appear, get new browsers and other updated programs, and sometimes new features. Business also change their processes and develop new requirements. If you can't cope with change then you shouldn't be in the IT business.
I have worked with IT for over 30 years. Today the desktop/personal device market is more complex that I can ever remember. First we have Win 7. Then VDI technology from the likes of Citrix. Then new browsers eg IE8/9 and Crome. Then the I-pad. If that isn't compelx enough we add in mobile device. In the business environment all of these are potentially required to run applications and the complexty/cost of managing this is huge.
So we could blame Microsoft for creating the first widely used 'thick client' and argue that they are the root of all evil. Once upon a time mini computers with dumb terminals worked just fine. However Apple is showing today that people like buying new and 'pretty' products/devices. So you could argue if Microsoft had not done their Windows bit someone else would have taken the leadership in this space. If we just had to manage say Windows 7 that might be ok - but all the other opertng environments makes the job just about impossible.
At the end of the day consumers/businesses are buying new product and seem to prefer the latest thing over reliability/security. Until buying behaviour changes it will tkae us a long long time to get to a really stable/easy to use IT infrastructure
"Why don't you go find a Mac site and whinge about Apple instead? XP came out in the same time-frame as OS X 10.0 (Cheetah). Since then, Mac users would probably paid good money for Puma, Jaguar, Panther, Tiger, Leopard and Snow Leopard. If they were trying to jump from Cheetah to Snow Leopard today, they'd have found their whole Classic environment had disappeared and they'd need to switch from PowerPC to Intel hardware."
I agree and I haven't gone down that path. Let's look at Red Hat Enterprise Linux, or even CentOS (the free alternative) instead. About the time that XP came out, Red Hat 7.x was out. The difference here is like Mac OS X, upgrades would have been done along the way, not directly from Red Hat 7.x to Red Hat EL6. And upgrades in Linux generally go smooth especially when incrementally done. Data is retained, packages are upgraded in place, etc.
Microsoft did a good job at releasing service packs along the way with XP to keep it updated. But what about a service pack 4 to allow an in place upgrade to Windows 7 (which would be an incremental upgrade)? What about providing a clear upgrade channel for current XP customers to receive a steep pricing discount for 7?
"However, my main point was that the beta should be used to find out if you are going to have problems, where they are, and how you are going to deal with them. "
I agree, it's a starting point. And I definitely agree that the beta can be used to assess. My comment was in response to the migration path itself, and testing the migration, not the re-writing of software to fit the migration.
I would love to hear of hands-on experiences with XP to 7 migrations as a comparison. What were the roadblocks vs. milestones? Was the migration tool used, or a 3rd party solution?