Q&A: Windows 7 Deployment

Challenges, mistakes, and best practices for moving to Windows 7.

With all the competing priorities facing IT, should one of them be migrating to Windows 7? What are the benefits (and do enterprises actually realize them)? What are the challenges and best practices for conducting a Windows 7 migration? To learn more about the strategies enterprises should be using, we turned to Aaron Suzuki, the CEO of Prowess (http://www.smartdeploy.com/) whose product simplifies such migrations.

Enterprise Strategies: A major analyst group claims that 83 percent of companies will deploy Windows 7 within a year. Do you agree with this prediction?

Aaron Suzuki: That’s a crazy metric. I guess it depends on what your definitions of “companies” and “deploy” are. It is possible that 83 percent of small and mid-size businesses will deploy Windows 7 in a year. It is possible that 83 percent of enterprises will deploy some Windows 7 desktops within a year, but it is absolutely crazy to think that 83 percent of all companies worldwide will fully deploy Windows 7 within a year.

When Windows 7 was in beta, we thought that within 18 months of release everyone would be well on their way through migration. Our current position, and this is backed up by our experience with customers, is that serious deployment will start to happen in another year, with most deployments happening in 2012-2013. I think more migrations will happen in 2014 than will happen in 2011. Many of our largest customers are barely thinking about Windows 7 migration. They’re certainly in no hurry.

What benefits and drawbacks do you see for enterprises that wait to deploy Windows 7?

We use Windows 7, but we’re a relatively small tech company. There are lots of nice features; performance is good, and security improvements including BitLocker make it hard to dislike for a CIO or tech person. However, the enterprise has largely addressed the challenges Windows 7 addresses through third-party or homegrown solutions (which is why Windows now includes them), and the enterprise probably spent a lot of time and money solving the challenge, which lessens the urgency for them to migrate. There is a valid argument that improved manageability is a significant benefit to migration, but you also have to keep in mind that deployment and migration activities themselves re extremely expensive and time-consuming for the enterprise.

The upside of waiting is that the big pains of Windows 7 migration, especially application compatibility, may be addressed by ISVs that can provide Windows 7-compatible upgrades that will significantly ease the migration process. Another side benefit is that if a hardware refresh is accompanying the OS migration, the longer organizations wait to buy hardware, the more powerful and capable the device will be at an increasingly lower cost. This is a particularly relevant point for companies considering doing more sophisticated deployments that include client-hosted desktop virtualization.

For those enterprises that have migrated, are they actually achieving the benefits they expected from Windows 7?

I think the main benefit for most organizations is escaping the closing support window for Windows XP, so from that perspective, yes, the benefits are fully realized.

All smugness aside, I think once the project is done, people are pretty happy, but the process is not always pretty. Deployment changed massively with Vista, but few organizations deployed Vista, so there has been a significant adjustment.

What’s driving the need for deployment and migration solutions in the market?

Cost. In the enterprise, deployment is monumental. It is monumentally expensive. It is monumentally time-consuming, which is more cost. It is monumentally impactful to the business in terms of productivity gains and, more to the point, lost productivity if things don’t go perfectly, which translates into still greater cost.

The thing about deployment is that there is so much potential for improvement. It is so complicated and deeply technical that solutions have traditionally been very technical which limits the solution to the enterprise, where they can afford the expertise to use such solutions. Historically, it hasn’t been very reliable (more cost).

Some solutions are starting to drive down the technical requirement and reliability is improving, but there’s so much room to make everyone’s lives easier that the value proposition is too compelling to ignore. It’s a bit surprising that more isn’t going on in terms of significant innovation. The demand is certainly there because the potential for cost savings is massive and the stakes are very high in terms of risk.

What are some of the specific deployment challenges enterprises face when deploying Windows 7?

Application compatibility is far and away number one. The same problem existed with Vista, but people are actually going to use Windows 7, so it’s time to face the music. The other related and sometimes less-appreciated but just as complicated and painful issue that I personally find more interesting is IE compatibility. Applications built for IE6 pose a huge problem for Windows 7 migration, and there are some complex workarounds, most of which employ virtualization of some type, to give you IE6. Desktop virtualization and the complicated management associated is another huge point of risk for organizations. Migrating web apps to IE7, or better yet IE8, is a huge barrier that is keeping enterprises from migrating, which I find amusing because it is not the OS but the browser that is tied to the OS that is posing the problem. The solution is to run the old, soon-to-be-unsupported OS. It really paints companies into a corner and it actually becomes incentive not to migrate.

Hardware compatibility is another potential issue, but it is becoming a minor issue. If companies are running five-year-old desktops, it’s pretty well a foregone conclusion that they’re going to need hardware, and since the release of Vista, Microsoft has significantly improved driver support. However, there is still a problem with getting your deployment solution to provide enough of the right drivers to get to the point that Windows will take over and avoid the blue screens and other disruptions that increase calls to your help desk. That part of the deployment problem won’t go away and on the IT side of deployment, it is one of the most painful, time-consuming aspects of Windows deployment.

What have been some of the common mistakes enterprises make when they migrate to Windows 7?

Windows 7 is a chance to change things. There is a lot of noise in the marketplace. Everything is called cloud and everyone wants you to buy lots of everything. It makes it hard to determine what the best and right solution is.

Enterprise organizations have their existing processes, tools, and habits. The worst mistake is to stick with the long-held practices organizations have used. Our recommendation is to take the time to make a change that is going to survive the next couple of generations of operation systems. Don’t just stay with what you have because you are comfortable with it or because you will need to obtain many approvals to get something done. Enough is changing now that if you can change your processes with it, it will pay massive dividends down the road in both time and cost savings. In particular, retooling for deployment using the latest techniques and technologies can simplify processes and improve efficiency for IT teams.

The hardware refresh that accompanies migration is, again, another consideration. Hardware comes powerful at a great value. In particular, storage and processor power is cheap, but it is easy to skimp on things such as memory and robust graphics capabilities. Considering where computing is going, it is worthwhile to spec high, especially on memory, so you will be prepared for trends that are coming to bear. An extra gig of memory will go a long way in 24 or 36 months when you roll out desktop virtualization broadly.

What best practices do you recommend to avoid these mistakes?

Measure twice, cut once. Prepare like crazy. Above all, start early. Most enterprises have teams that are dedicated to this job, but things like habit, culture, and corporate politics often disrupt the most authentic intentions. The CIO has the greatest ability to set the direction and momentum behind how projects like these play out, and the impact and influence trickles down from there to the operations director and IT managers and ultimately to the IT implementers who will execute the important tasks that keep people productive. On some level, “tone at the top” must be part of the picture.

There is still a “wait for SP1” mentality that delays execution, and I think often the planning and testing gets put off, too. With Windows 7, since it was a “tock” release (with the “tick” being Vista), there really was no need to wait for SP1. There are a few important hotfixes, but nothing should have been stopping companies from conducting their application compatibility testing, determining hardware requirements, and preparing for the move.

I also think companies that are purchasing new hardware, and many seem to be, should bring that new hardware in with Windows 7 on it. It isn’t worth the strategy of keeping the entire company on XP just because everyone runs the same OS on principle. Plus, making an old OS work on brand-new hardware can be very difficult.

Finally, don’t skip OSes. OS deployment should always be something you’re working on. For example, if companies had started their application compatibility testing back in the Vista timeframe, and moved to Vista, they could be lazily doing upgrades to Windows 7 now. Even if they didn’t move to Vista, the application compatibility testing would have applied to Windows 7 as well, and they’d be ready. It appears that everyone just put it off as a low priority task or project and now it’s painful for everyone.

What role does Prowess play in the Windows 7 migration market and what value does it bring to its users?

We have a team of experienced deployment engineers with years in consulting and enterprise IT that brought a new approach to the job of OS deployment and migration. The model shifts the paradigm toward virtualization and challenges many conventional deployment best practices. We released a product just over a year ago called SmartDeploy Enterprise that realizes this alternative approach and massively reduces the effort, complexity, and cost involved in deployment, and that statement is supported by some of the biggest schools, hospital systems, government institutions, and manufacturers in the world. SmartDeploy is a potential game changer for IT especially in the smaller enterprise, which is typically more IT resource-constrained.