Tag Archives: microsoft

.net, microsoft, mobile, silverlight, software development, windows

A glimpse into the internal battles that set the future of Windows and .NET

16 Comments

A couple of posts from Hal Berenson give insight into the internal battles at Microsoft as the company worked out its strategy to rescue Windows from irrelevance in the world of mobile and tablets. Berenson is now President of True Mountain Group LLC but was formerly at Microsoft where his roles included SQL Server development and architecture, Mobile Development Tools strategy, and General Manager of Forefront identity and security products.

image

Berenson left Microsoft in October 2010, but by that time the strategy behind Windows 8 and Windows Phone 8 would have been in place.

According to Berenson, there were two core options for evolving Windows. There may have been others, but the heart of it is this: what to do with .NET. One option was to make .NET the app model for Windows, which is what was planned for the original Longhorn, before it was reset and became the less radical update that was Windows Vista. The other was to create a new app model based on native code. Steven Sinofsky, the Windows President, chose the latter, which is why .NET is only one of three options for programming the new tablet personality in Windows 8. This meant going down the opposite path from that of Windows Phone 7, which has an entirely .NET-based programming model.

You may recall from other sources that Steven Sinofsky has never been known to be a .NET fan.  While others within Microsoft, and even senior people in the (pre-Windows 8) Windows organization, wanted to move to an entirely .NET app model for Windows Steven did not.  He (and others fyi) wanted to re-engage the native code C++ developers that Microsoft had been neglecting.  And they wanted to co-opt the huge base of web developers to create apps for the Windows platform.  Well, what had the Windows Phone guys done?  They’d implemented a .NET only app platform.  Could the Windows Phone app platform evolve to address the native and web developers?  Sure.  But with no existing library of apps and a desire not to have .NET-centric platform at the core of Windows Sinofsky apparently felt pretty comfortable ignoring the Windows Phone team’s work.

This goes a long way to explain the puzzlement many of us experienced when it transpired that having created in Windows Phone 7 the basis for a touch-friendly operating system that could easily be extended to larger form factors such as tablets, Microsoft chose instead to do a new thing entirely for its tablet strategy.

One take on this is that Berenson’s account illustrates the chaos at Microsoft. Windows Phone was created in a mad hurry in reaction to the iPhone and the ascendance of touch UIs, reusing pieces of .NET, Silverlight and Zune to bring something to market quickly. Then the company’s next move was not to build on that, but to throw it away, even in the context of a mobile and device revolution that was and is a huge threat to its core business. And where was CEO Steve Ballmer in all of this?

The other take though is how this shows the determination and strategic focus of Windows boss Steven Sinofsky. He did not believe that rebuilding the Windows user interface on .NET would save it, with the Longhorn experiment no doubt a factor in that conviction, so he refused to go down that path again, despite the cost in terms of time and, perhaps more seriously, the impact on the developer ecosystem. Microsoft platform developers were asked first to bet on .NET and Silverlight, and now to bet on this new thing the Windows Runtime, and many are disillusioned or even angry. A hard decision; but putting long term strategy ahead of the immediate demands of your customers may be the right thing, in fact the only right thing.

Berenson also confirms what many of us have always assumed: that the removal of the Start menu on the Windows 8 desktop is all about making the new personality in Windows hard to avoid:

The Start menu, and indeed the entire desktop, are legacies that will have to be removed from Windows over time.  While the desktop itself is probably with us for a couple of additional major Windows releases (though there may be truly desktop-free editions sooner than that) the start menu was something that Steven has bet he could get away with not bringing forward into Windows 8.  By doing so he forces users to start living in the new usage paradigm rather than totally avoiding it.  Yes you can still set up a system to avoid leaving the desktop most of the time.  But you can’t avoid the new world completely.  In doing so he sets people up to eventually accept systems without the desktop at all (or at least Windows RT systems for personal use even if they need the desktop at work, for example).

Personally I no longer miss the Start menu; but its absence is certainly a barrier to adoption for Windows 8, as new users struggle to navigate the operating system.

Note: Berenson has kindly commented below. Note his point that merely working at Microsoft does not give you detailed knowledge of all decisions made there.

.net, microsoft, mobile

Microsoft’s Azure Mobile Services: node.js and more in beginnings of easy cloud to device development

3 Comments

Microsoft announced Azure Mobile Services last month and it was mentioned by Microsoft Server and Tools boss Satya Nadella at the launch of Visual Studio 2012, as an example of where Microsoft is going with its “Modern app” vision, continuous services and connected devices (but with a Windows 8 or Windows Phone 8 flavour).

Azure Mobile Services is in some ways a reworking of the WCF RIA Services developed to support Silverlight applications, and in fact I swear I saw a reference to RIA Services flash past when I was opening my first Azure Mobile Services project in Visual Studio. It consists of a service type in Microsoft’s Azure cloud combined with a client SDK which is currently for Windows Runtime apps in Windows 8, though the REST protocol used could be called from any client platform.

image

Looking at the dashboard for a Mobile Services project in the Azure portal, you can see what Microsoft is going for here. Mobile Services handles authenticated access to data stored in SQL Server Azure. It is designed to be simple and cost-effective to get started, but can be scaled out by moving from a service on a shared host, to a dedicated VM with multiple instances.

image

It is easy to think of cases where the cloud component of a cloud plus device app need do little more than authenticate users, and retrieve and update data. Azure Mobile Services also provides for server-side scripts which you can modify to handle validation and other tasks.

I was interested to see that the server-side scripts are written in JavaScript and executed by node.js. Node.js is fantastic, and one of the benefits is that if you have an HTML and JavaScript client, you can use JavaScript both on the client and on the server. On the other hand, I wonder if Microsoft’s community would rather work with C# on the server, which is more mature and more familiar. Scott Guthrie’s introductory tutorial does not mention node.js.

I had a quick go at creating my own Azure Mobile Service. I have only been partially successful so far.

Things started well enough. I created a mobile service and the Quick Start opened.

image

Both Guthrie’s blog and the Quick Start wizard in the Azure portal are based on a todo list app. I went slightly off-piste here, deciding instead to create an app to track my articles on the web. I wanted to see how Azure Mobile Services copes with related tables, as opposed to a single table.

I had a frustrating time trying to create the database tables. I had to add my IP address to a firewall rule, enable popups, and deal with connection failures caused by unknown network issues.

Finally I was able to get into the database designers. I created an Articles table joined to a Publications table, with a very few fields.

Next I downloaded an automatically generated Windows 8 app from the portal. I had hoped this would magically work with my data. Unfortunately though, it seems to be hard-coded for the todo list app. If you do not want a todo list, you have to write your own code; and so far I have not had time to figure out from the reference what to do next. I looked at the Get started with data article, and guess what, it is the todo list again.

When you create a database, you can specify simple permissions. The todo list example depends on an application key stored in your app and sent over SSL, to grant permission to read and modify data. I selected authenticated user access instead.

image

There is an article explaining how to add authentication, though note that it presumes use of a Microsoft Live ID (the service formerly known as passport). This is perfect in the context of Windows 8 and Windows Store apps, but businesses will want to use Active Directory instead, whether hosted in Azure or Office 365 or on premise. I presume Microsoft will add this at some point though it is not mentioned currently.

My initial conclusion is that Azure Mobile Services shows lots of promise, but that the introductory documentation could be usefully improved, for example not to assume that you want to make a single table todo list app.

In this context the partnership with Xamarin, which is extending the SDK to Apple iOS and Google Android, is excellent news. This makes Azure Mobile Services useful more broadly, and I have a hunch that Xamarin’s support will soon improve the documentation and tutorials. The client SDK is open source and on github.

Note that according to Microsoft’s Kirill Gavrylyuk, in answer to a question from Roger Jennings, Microsoft plans to “roll out full support for iOS and Android including native SDKs soon”, rather than leaving the non-windows support entirely to Xamarin and C#.

development, microsoft, open source, professional, software development

Microsoft Project Austin: superb C++ code sample for Windows 8

1 Comment

No time to blog in detail about this; but developers with any interest in Windows 8 should check out Project Austin, a sample project for Windows 8 whose quality exceeds most of what is currently available in the Windows Store.

This is a simple note-taking app but beautifully rendered and with support for adding photos, sharing via Charms, and more:

It’s amazing how useful just a pen and a paper are by themselves. But when you take that concept to the computer realm and expand it to do things like add photos and annotate them right on the spot, and digitally share what you create, then the possibilities are endless.

say the project’s creators. More to the point:

Austin aims to demonstrate with real code the kind of device-optimized, fluid and responsive user experience that can be built with our newest native tools on the Windows8 platform.

Most of the code has been put on CodePlex under the Apache 2.0 license. It demonstrates C++ AMP, the new parallel library for GPU computing, as well as C++ coding for the Windows Runtime.

I downloaded and built the project with few issues, following the helpful guidance here. I did have to add the boost libraries manually to the include path.

image

It is not completely stable but is already very pretty.

image

Immersive UI of course. The tools disappear while you are working. I am still not sure how well I like this, but it is good to see examples of how Microsoft thinks this should work.

.net, development, microsoft, professional, software development, visual studio

Platform churn? If it is in Windows 8, we are committed to it says Microsoft

4 Comments

I interviewed Corporate VP of Microsoft’s developer division Soma Somasegar at the Visual Studio 2012 launch last week; see the article on the Register here. I asked about the inconsistency of the Microsoft platform, and the way the platform story has changed over the years (Win32, .NET, WPF, Silverlight and now Windows Runtime). Can developers trust in the longevity of today’s platform, especially on the client?

Here is what I thought was interesting about his reply:

Any technology you see shipping as part of Windows 8, we are very committed to that.

So what ships in Windows 8? Well, for reasons which are hard to discern for those of us outside Microsoft, Silverlight is not shipped in Windows 8. It is an optional download. In fact, the only plug-in installed by default is Adobe Flash:

image

No, that does not imply that Microsoft is committed to Flash; but it does suggest lack of commitment to Silverlight, which we knew.

What you do get though is .NET Framework 4.5. This is baked in and cannot be removed as far as I can tell, though you can add and remove some advanced features.

This means you also have Windows Presentation Foundation (WPF); and in fact Somasegar specifically refers to this alongside Win32 and the new Windows Runtime.

The inclusion of technology in a current Microsoft product has implications for its support lifecycle. The ancient Visual Basic runtime, for example, is still assured of a long life since it is part of Office 2013.

image

My guess is that Microsoft’s thinking goes something like this. Right now, as the October launch date of Windows 8 approaches, what Microsoft needs most urgently is a viable ecosystem for its new Windows Runtime environment. This, you will notice, is the focus of the forthcoming BUILD conference as so far announced.

image

What, though, of the Windows desktop, has Microsoft abandoned it as legacy? My guess is that we will get deliberately mixed messaging on the subject. On the one hand, Microsoft has relegated the desktop to a single tile in the new Start screen. On the other hand, most of us will spend most of our time in the desktop, not least developers who need it to run Visual Studio. If Microsoft succeeds in establishing the new Windows Runtime platform, it would not surprise me to see a little more love given to the desktop in, say, Windows 9.

Microsoft’s platform story is messy, without question, and especially so in mobile. We have seen Windows Mobile replaced by the incompatible Windows Phone 7, and now those loyal developers who invested in the Silverlight/XNA Windows Phone 7 technologies are finding that it is all change again in Windows Phone 8 and Windows 8 (though the exact details await the release of the Windows Phone SDK).

It seems clear though that the company’s current intent is that Windows Runtime evolves as the primary client platform for both phone and tablet, while desktop remains for legacy support and for applications that do not fit the new model, such as Visual Studio and (for the time being) Office.

What if Microsoft fails to establish the Windows Runtime as a popular app platform? All I can add is that I know of no Plan B.

.net, microsoft, software development

Information Density in Metro, sorry Windows Store apps

4 Comments

Regular readers will recall that I wrote a simple blog reader for Windows 8, or rather adapted Microsoft’s sample. The details are here.

This is a Windows Store app – a description I am trying to get used to after being assured by Microsoft developer division Corp VP Soma Somasegar that this really is what we should call them – though my topic is really the design style, which used to be called Metro but is now, hmm, Windows Store app design style?

No matter, the subject that caught my attention is that typical Windows Store apps have low information density. This seems to be partly due to Microsoft’s design guidelines and samples, and partly due to the default controls which are so boldly drawn and widely spaced that you end up with little information to view.

Part of the rationale is to make touch targets easy to hit with fat fingers, but it seems to go beyond that. We should bear in mind that Windows Store apps will also be used on screens that lack touch input.

I am writing this on a Windows 8 box with a 1920 x 1080 display. Here is my blog reader, which displays a mere 7 items in the list of posts:

image

This was based on Microsoft’s sample, and the font sizes and spacing come from there. I had a poke around, and after a certain amount of effort figuring out which values to change in the list’s item template, came up with a slightly denser list which manages to show 14 items in the list. The items are still easily large enough to tap with confidence.

image

Games aside though, I am noticing that other Windows Store apps also have low information density. Tweetro, for example, a Twitter client, shows only 11 tweets to view on my large display.

The densest display I can find quickly is in Wordament, which is a game but a text-centric one:

image

I have noticed this low information density issue less with iPad apps. Two reasons. One is that iOS does not push you in the same way towards such extremely large-looking apps. The other is that you only run iOS on either iPhone or iPad, not on large desktop displays.

Is Windows 8 pushing developers too far towards apps with low information density, or has Microsoft got it right? It is true that developers historically have often tried to push too much information onto single screens, while designers mitigate this with more white space and better layouts. I wonder though whether Windows 8 store apps have swung too far in the opposite direction.

.net, microsoft, professional, software development, visual studio, windows

When will Blend, Microsoft’s Visual Studio design tool, be done? Not for a while says Soma Somasegar

1 Comment

During the Visual Studio 2012 launch last week I took the opportunity to ask Developer division Corp VP Soma Somasegar when Blend, Microsoft’s design tool for Visual Studio 2012, will be finished. A tricky question to answer, since there are multiple versions, as explained here:

  • Blend for Windows Store apps (HTML or XAML) is fully released and available as part of Visual Studio 2012

    image

  • Blend for WPF and Silverlight is in preview. For production you are meant to use the old Blend 4, unless you are targeting Silverlight 5 where you have no choice but to use the preview version.
  • Blend for Windows Phone is part of the Windows Phone SDK 7.1

The SketchFlow prototyping tool is also part of the preview Blend.

So when will Blend for Visual Studio 2012 be done? Somasegar refers to HBlend, which is the HTML version, and XBlend, which is for XAML.

“We shipped HBlend, and we shipped a preview of XBlend. It will take several months to finish. We also want to continue adding to HBlend. So I can’t tell you that Blend is ever going to be done [laughs],” he told me.

That said, the full Blend for Visual Studio 2012 will come out of preview sometime. Will it coincide with the first update for Visual Studio, announced for later this year?

“It is going to be later than the update, but I don’t have a specific timeframe,” he said.

Personally I have mixed (ha!) feelings about Blend. On the one hand, it is obvious that the simple designer in the Visual Studio IDE is insufficient, and that the rich Blend tool is needed, for those who can make sense of its intricate user interface. On the other hand, the designer aspect of Microsoft’s tooling seems to me messy, with too many versions of Blend and overlap between Blend and Visual Studio which gives developers a difficult choice: do I work with Blend, or stay within the simpler but more limited IDE tools?

.net, development, microsoft, professional, software development, visual studio, windows

Visual Studio 2012 launch: focus on Modern Apps

Microsoft is holding a launch event in Seattle for Visual Studio 2012, attended by selected Microsoft-platform developers as well as press from around the world.

image

Corporate VP Soma Somasegar kicked off the keynote, saying that Visual Studio 2012 has already been downloaded 600,000 times since its release to the web around one month ago – a take-up, he said, which exceeds previous versions.

image

But what is Visual Studio 2012 all about? It’s for Modern Apps, said Somasegar, though you would be wrong to imagine that this means a Windows 8 formerly-known-as-Metro app. Rather, a Modern App implies continuous services and multiple client devices, connected over both public and private networks. In other words, think mobile as well as desktop, public and private cloud, and bring your own device as well as corporate desktops.

Fair enough, but it is all a bit vague and still leaves us wondering what exactly apps for the Windows Runtime are called at Microsoft. Visual Studio calls them Windows Store apps, which does not make complete sense when you consider that you can deploy the apps without going through the Store, and that the Store can include desktop apps although these are links rather than direct downloads.

Visual Studio 2012 is a vast and impressive product though. Developers were shown various new features, including asynchronous development for maintaining a responsive user interface, pixel-level debugging in DirectX, after the fact debugging using Intellitrace, and new HTML and CSS support in the editor, with error highlighting that adapts to the specified version of HTML.

There was a quick demonstration of developing for Kinect, in which we saw an avatar mimic the movements of VP Jason Zander.

image

Technical Fellow Brian Harry spent some time showing off application lifecycle features in Team Foundation Server, including improved Scrum support and an emphasis on agile concepts like whole team development (the development team is not just developers).

There were a couple of announcements, including news of an update in preparation for Visual Studio 2012 which will include new features. There will be a preview later this month and delivery by the end of 2012.

Visual Studio Express for the desktop has now shipped and is available for download. This supports development in C#, Visual Basic and C++, so there is now a free C++ compiler available for Windows desktop development in the Visual Studio 2012 family.

I spoke to a couple of the invited developers after the morning event. Visual Studio 2012 looks good, they told me, but then again what choice is there for Windows platform development? That said, they had expected a stronger push for Windows 8 development, especially on the Metro-style side. Why did Microsoft not spend more time evangelising Windows 8 app development and the Windows Store? Of course there is a lot to cover but given how critical app momentum is to the success of the new Windows, it struck me as a valid point.

adobe, internet, microsoft, open source, windows

Adobe’s Roy Fielding patches Apache to ignore IE10 Do Not Track privacy request

32 Comments

Adobe’s Roy Fielding, who is also the original author of the W3C’s Tracking Preference Expression draft, has patched Apache, the open source web server, to ignore the Do Not Track header sent by Microsoft’s Internet Explorer 10, the browser in Windows 8:

image

Under the heading “Apache does not tolerate deliberate abuse of open standards,” Fielding’s patch sets Apache to remove the Do Not Track request header if IE10 is the web browser.

Fielding’s argument, one presumes, is that IE10 breaches clause three in the Tracking Preference Expression draft:

Key to that notion of expression is that it must reflect the user’s preference, not the choice of some vendor, institution, or network-imposed mechanism outside the user’s control. The basic principle is that a tracking preference expression is only transmitted when it reflects a deliberate choice by the user. In the absence of user choice, there is no tracking preference expressed.

However the document goes on to say (highlighting is mine):

We do not specify how tracking preference choices are offered to the user or how the preference is enabled: each implementation is responsible for determining the user experience by which a tracking preference is enabled. For example, a user might select a check-box in their user agent’s configuration, install an extension or add-on that is specifically designed to add a tracking preference expression, or make a choice for privacy that then implicitly includes a tracking preference (e.g., Privacy settings: high). The user-agent might ask the user for their preference during startup, perhaps on first use or after an update adds the tracking protection feature. Likewise, a user might install or configure a proxy to add the expression to their own outgoing requests.

Here is what happens in Windows 8 after startup. This is among the first screens you see when installing Windows 8, before you get full access to the operating system:

image

One of the settings specified is “Turn on Do Not Track in Internet Explorer. If you click Learn more about express settings you get this:

image

If you click Customize you get this:

image

Does this respect the user’s preference? It seems to me a reasonable effort. The only objection I can see is if you consider that any user agent that defaults to setting Do Not Track on cannot be respecting the user’s preference. The draft specification does not state what the default should be.

It is also worth noting that clause 3 in the Tracking Preference Expression draft has changed; the wording about “not the choice of some vendor” was inserted in the 7th September draft, after Windows 8 was released to manufacturing. Here it is in the latest (March 2012) W3C Working draft:

Key to that notion of expression is that it must reflect the user’s preference, not the preference of some institutional or network-imposed mechanism…

Even if you agree with Fielding’s views on browser defaults, quietly patching the world’s most used web server to ignore the IE10 setting looks hard to defend, especially on a matter that is far from clear cut. Fielding is personally involved, not only as the author of the Tracking Preference Expression document, but also as an employee of Adobe, which specialises in digital marketing and may be more aligned with the vendors and their brands which may want to track user activity wherever their ads appear, rather than with end users.

Of course Apache is an open source project and Fielding’s patch has attracted the attention of the Apache community and may not survive.

It is also possible that a future draft of the Tracking Preference Expression document will state that Do Not Track must be off by default; but even if it does, patching the web server to ignore the browser’s header strikes me as a contentious solution.

Finally, it is worth noting that sending the Do Not Track header has little effect on whether or not your activity is tracked, since its meaning is unclear and respecting its value is a a choice made by third-parties, so this is a debate with little practical impact for the time being.

microsoft, virtualization, windows

Microsoft’s Hyper-V Server 2012: too painful to use?

33 Comments

A user over on the technet forums says that the free standalone Hyper-V is too painful to use:

I was excited about the free stand-alone version and decided to try it out.  I downloaded the Hyper-V 2012 RC standalone version and installed it.  This thing is a trainwreck!  There is not a chance in hell that anyone will ever use this thing in scenarios like mine.  It obviously intended to be used by IT Geniuses in a domain only.  I would really like a version that I can up and running in less than half an hour like esxi.  How the heck is anyone going to evaluate it this in a reasonable manner? 

To be clear, this is about the free Hyper-V Server, which is essentially Server Core with only the Hyper-V role available. It is not about Hyper-V in general as a feature of Windows Server and Windows 8.

Personally I think the standalone Hyper-V Server is a fantastic offering; but at the same time I see this user’s point. If you join the Hyper-V server to a Windows domain and use the administration tools in Windows 8 everything is fine; but if you are, say, a Mac user and download Hyper-V Server to have a look, it is not obvious what to do next. As it turns out you can get started just by typing powershell at a command prompt and then New-VM, but how would you know that? Further, if Hyper-V is not joined to a domain you will have permission issues trying to manage it remotely.

Install Hyper-V Server, and the screen you see after logging on does not even mention virtualization.

image

By contrast, with VMWare’s free ESXi has a web UI that works from any machine on the network and lets you get started creating and managing VMs. It is less capable than Hyper-V Server; but for getting up and running quickly in a non-domain environment it wins easily.

I have been working with Hyper-V Server 2012 myself recently, upgrading two servers on my own network which run a bunch of servers for development and test. From my perspective the free Hyper-V Server, which is essentially Server Core with only the Hyper-V role available, is a great offer from Microsoft, though I am still scratching my head over how to interpret the information (or lack of it) on the new product page, which refers to the download as a trial. I am pretty sure it is still offered on similar terms to those outlined for Hyper-V Server 2008 R2 by Program Manager Jeff Woolsey, who is clear that it is a free offering:

  • Up to 8 processors
  • Up to 64 logical processors
  • Up to 1TB RAM
  • Up to 64GB RAM per VM

These specifications may have been improved for Hyper-V Server 2012; or perhaps reduced; or perhaps Microsoft really is making it a trial. It is all rather unclear, though I would guess we will get more details soon.

It is worth noting that if you do have a Windows domain and a Windows 8 client, Hyper-V Server is delightfully easy to use, especially with the newly released Remote Server Administration Tools that now work fine with Windows 8 RTM, even though at the time of writing the download page still says Release Preview. You can use Server Manager as well as Hyper-V Manager, giving immediate access to events, services and performance data, plus a bunch of useful features on a right-click menu:

image

In addition, File and Storage services are installed by default, which I presume means you can use Storage Spaces with Hyper-V Server, which could be handy for hosting VMs with dynamically expanding virtual hard drives. Technically you could also use it as a file server, but I presume that would breach the license.

For working with VMs themselves of course you have the Hyper-V Manager which is a great tool and not difficult to use.

image

The question then: with all the work that has gone into these nice GUI tools, why does Microsoft throw out Hyper-V Server with so little help that a potential customer calls it “too painful to use”?

Normally the idea of free editions is to entice customers into upgrading to a paid-for version. That is certainly VMWare’s strategy, but Hyper-V seems to be different. It is actually good enough on its own that for many users it will be a long time before there is any need to upgrade. Microsoft’s hope, presumably, is that you will run Windows Server instances in those Hyper-V VMs, and these of course do need licenses. If you buy Windows 8 to run the GUI tools, that is another sale for Microsoft. In fact, the paid-for Windows Server 2012 can easily work out cheaper than the free editions, if you need a lot of server licenses, since they come with an allowance of licenses for virtual instances of Windows Server. Hyper-V Server is only really free if you run free software, such as Linux, in the VMs.

Personally I like Hyper-V Server for another reason. Its restricted features mean that there is no temptation to run other stuff on the host, and that in itself is an advantage.

microsoft, virtualization, windows

Upgrading to Hyper-V Server 2012

1 Comment

After discovering that in-place upgrade of Windows Hyper-V Server 2008 R2 to the 2012 version is not possible, I set about the tedious task of exporting all the VMs from a Hyper-V Server box, installing Hyper-V Server 2012, and re-importing.

There are many reasons to upgrade, not least the irritation of being unable to manage the VMs from Windows 8. Hyper-V Manager in Windows 8 only works with Windows 8/Server 2012 VMs. It does seem to work the other way round: Hyper-V Manager in Windows 7 recognises the Server 2012 VMs successfully, though of course new features are not exposed.

The export and import has worked smoothly. A couple of observations:

1. Before exporting, it pays to set the MAC address of virtual network cards to static:

image

The advantage is that the operating system will recognise it as the same NIC after the import.

2. Remove any snapshots before the export. In one case I had a machine with a snapshot and the import required me to delete the saved state.

3. After installing Hyper-V 2012, don’t forget to check the date, time and time zone and adjust if necessary. You can do this from the sconfig menu.

4. The import dialog has a new option, called Restore:

image

What is the difference between Register and Restore? Do not bother pressing F1, it will not tell you. Instead, check Ben Armstrong’s post here. If you choose Register, the VM will be activated where it is; not what you want if you mistakenly ran Import against a VM exported to a portable drive, for example. Restore on the other hand presents options in a further step for you to move the files to another location.

5. For some reason I got a remote procedure call failed message in Hyper-V Manager after importing a Linux VM, but then when I refreshed the console found that the import had succeeded.

6. Don’t forget to upgrade the integration services. Connect to the server using the Hyper-V Manager, then choose Insert Integration Services Setup Disk from the Action menu.

image

Cosmetically the new Hyper-V Server looks almost identical to the old: you log in and see two command prompts, one empty and one running the SConfig administration menu.

Check the Hyper-V settings though and you see all the new settings, such as Enable Replication, Virtual SAN Manager, single-root IO virtualization (SR-IOV), extension support in a virtual switch, Live Migrations and Storage Migrations, and more.