My 2014 Hyper-V Wish List27 Dec 2013 by 4
The end of another year is upon us. It’s given us plenty to talk about, and we will carry some of that discussion into the new year. Here are some of the things I’d like to see happen in our industry.
2. Pro-Customer Improvements in Microsoft’s Culture
If the reports I’ve read are to be believed, Microsoft’s culture could use a big change. The phrase I often see is “dog-eat-dog”. That doesn’t really work out for anyone in the end, especially not the customer. I’m sure that everyone is well-aware of the broken patches that went out this year. What if someone there knew that some of those patches were bad, but also knew that a rival would fall if one became generally available? That, to me, sounds like incentive to let a bad patch out the door. Nobody really wins in that sort of environment. Not the employees, not the customers, not the company. Since they’re getting a new CEO, there’s a perfect opportunity to get to work straightening that out.
3. Acknowledgement of Error
Along with the above, I’d like to see some contrition. Use Marissa Mayer’s apology for the recent Yahoo mail outage as an example. I don’t want to spend time on an analysis of it, but the point is, there’s an apology there. Microsoft has really dropped the ball a lot lately, especially on things like patches, and I don’t like seeing anonymous, dry, nondescript KB articles about it. You can read Ben Armstrong’s post on this thread for proof that they can do better. That behavior needs to be extended to some of their more serious goofs.
4. Improvements for Everyone
Next, I’d like to see Microsoft (and many others) acknowledge that small businesses are not just some aberration to be dealt with. I have some proof:
As you can clearly see, the vast majority of businesses in the United States have fewer than 20 employees. About half of all employees in the U.S. work for companies smaller than 500 users. While there’s no set distinction, a 500-user organization would likely qualify as “medium-sized” in most people’s minds. But, Microsoft doesn’t appear to be more than marginally aware of the needs of those smaller shops. Monolithic pricing for System Center jumps to mind. $1,500 for software licensing on a physical box is meaningless to a large corporation, but can be a burden for even a medium-sized business. Another canary of the uncaring is the death of shadowing Remote Desktop sessions. Big organizations probably mostly use Citrix for ordinary remote desktops and don’t care. I guarantee a lot of smaller organizations care very much. As I’ve said elsewhere, the small businesses train the technicians that eventually go to work in the large businesses. If you want your big companies to switch to non-Microsoft technologies, keep pretending that the smaller guys don’t matter.
It would also help if the well-known community members out there would acknowledge the needs of the small/medium business environment as well, because they’ve collectively got the influence to shape things. “Everyone must move to 10G” is a great example (you’ll notice all the people saying this have 1G in their home labs, so “must” apparently hinges on “my money” versus “your money”). Screaming at people to get off of XP before support expiration without even acknowledging the thousands of LOB applications that won’t move forward is another. Software-defined networking in a ten user organization? Yeah, right.
5. SCVMM Redone or Replaced
I’d really like to see System Center Virtual Machine Manager revamped. It is a powerful tool but very bad software. In fact, I’d probably vote it in the top ten worst pieces of software I’ve ever used. What I’d like to see that team do is write down a list of all its functions, scrap the interface entirely, hire a user interface expert or five, and rebuild the whole thing. I could probably write a small book on all the issues that SCVMM has, but I’ll just outline a few. The interface is perfectly horrid.Â Think about all the times you manipulate VMs. What are the top three things you’re likely to want to do? Let me guess: change settings, connect to a console, or perform some sort of power action. Am I right? Am I close? If I messed anything up, I’ll bet that you said, “Move the VM to another host” instead of one of those three. So, let’s start by looking at a more carefully designed interface:
I don’t know that I can say it’s perfect, but it’s at least logical. Also, if a menu item isn’t relevant at any given point in time, then Hyper-V Manager hides it. That’s sort of a mixed blessing, but it keeps the menu size down (personally, I think the world should move to radial menus, putting unusable items to the right of the click-spot if you’re right-handed and to the left if you’re left-handed, but I digress). Now, let’s go to VMM:
Seriously? The first option is to create a clone? I’m guessing that the number of VMM users in the world that need that as their first option is a statistical anomaly. The power options are OK, I suppose, but Storage Live Migration before regular Live Migration? Repair before Connect? The things you’re actually likely to want to do are approaching the point farthest from where you clicked. VMM’s context menu is a jumbled mess. In aggregate, this wastes time and is a repetitive strain injury in the making.Â Yes, user experience is a big deal.
Another issue with VMM is its utter lack of self-healing. ALL problems are complete show-stoppers. Fine, so I placed a non-highly available VM on shared storage and that violates the principle of WWMD (What Would Microsoft Do?). Does that really justify taking away my ability to turn it off? Yes, I created a virtual switch outside of VMM. Does that really mean VMM has no way to figure out how to manage it? Yes, my VM has only one Preferred Owner. Does that really mean I should be prohibited from putting its host into Maintenance Mode?
As often as the program generates errors, you’d think it would be good at it. But it’s not. Usually, you have to start something, wait a minute to see if the inconsistent Jobs window will show up this time, manually switch to the Jobs screen if it doesn’t, and then you can finally see if what you did worked. You’ll get error messages that tell you to go look in the logs, but it probably won’t tell you which log, compounded by the fact that it probably didn’t log anything anyway. If the error message has details, they’re rarely related to what actually went wrong. For instance, I attempted to configure my first test installation with distributed key management in accordance with the instructions. After 15 minutes of an unhelpfully cycling track bar, the installation stopped with this:
There was nothing relevant in the logs. In accordance with the “all errors are fatal” philosophy, there was no back or retry or skip button. Just another 15 minutes of my life sucked away by VMM with nothing to show for it.
Well, that’s enough on VMM. I didn’t really mean to make this an anti-VMM rant. Quick summary for this point: I’d like to see this product improved. A lot.
6. Bring back TechNet subscriptions.
A lot of pixels have already been blitted on this topic and I doubt I could say anything that hasn’t already been said.
I wish I had some hope that any of these things would be addressed in 2014. But, we are legion. Raise your voice. Who knows what we can accomplish?
I want to close 2013 by thanking everyone who reads this blog. I enjoy producing content for you and hope that you find it useful and meaningful. My number 1 wish for 2014 is that everyone enjoys a safe and happy holidays and joins us again in January for more Hyper-V adventures.
Have any questions or feedback?
Leave a comment below!