They ask a lot of questions they don't answer, and from my experience with both Windows and Linux, Linux can answer most of those questions far better than Windows. This is a joke.
edit: example: "Will patching and security lead to downtime?" There's ksplice and kgraft and more - patch the kernel with no downtime. And even without using either, I can't remember the last time I had to reboot immediately because a patch was that critical, or the last time I had to sit there waiting for my laptop to reboot 3 times trying to install the weekly updates. Windows Update blows.
edit: another example: "Proven security development lifecycle". Ha ha ha ha ha!
Updating application software on Windows can be a pain as well since replacing the running binary must be done by the binary itself (otherwise it seems the file is locked). Which means that unless an application is specifically written to support it, upgrading in-place isn't an option (AFAIK).
You owe me a new keyboard. This one is full of coffee.
Also got to love the whole "Linux doesn't do AD" BS. If you still live in 2005, perhaps not, but that's about it. You could also argue that AD is a misfeature, given the almost hilarious insecurity of NTLM password hashes, etc.
Ah, the FSF argument. I guess saying GNU/Linux was too wordy for Microsoft anyway, with their whole Metro concept being "having information density and options is bad".
The cost of Windows server licenses for a 1,000+ node system would be prohibitively expensive. Many of the Cray systems run Cray's own OS. It's a bit easier to run Linux in a light-weight configuration. There is more existing software for management of such systems in Linux (both system management and job management). Most of the top (public) supercomputers (that anyone has bothered to run linpack on and publish the results) are also running academic software (think F/OSS built on F/OSS using F/OSS tools). Windows just doesn't stand a chance in that world.
However, sometimes your solution is focused solely on server-side operating systems, and that's where Linux falls short.
I laughed hard. Also what is with all the question marks on the right side of the table? It's almost like they are not confident with their statements about Linux.
They seem to only have one real point, that microsoft products work best with microsoft products. The rest of it is just building up a sense of vague uncertainty. I would have thought Microsoft could make better arguments than that.
You should use Microsoft server solutions because
1. Microsoft servers work well managing linux solutions
2. Linux systems can't handle microsoft solutions themselves
3. We are uncertain what Linux is capable of/how secure it is
4. Therefore, you should go Microsoft (QED)
Ignoring how blatantly misleading these points might be given that linux is quite mature and we now live in a work of system virtualization (Wine anyone?), the argument here only serves to convince me that I should be using linux solutions instead of Microsoft solutions in my company. If a given solution works for linux (which, despite calling out Active Directory, they don't seem to have ruled out as a possibility) and can also be managed through Azure, then Why on earth would I create an inflexible system that can only be managed through Microsoft branded tech? Am I supposed to be convinced by their skepticism on linux? I'm perfectly capable of researching linux server solutions myself. Am I supposed to be amazed at the flexibiliy of Microsoft servers while simultaneously scared away from the flexibility of using linux solutions? Outside of server use, this doesn't seem like a good argument to use Microsoft solutions at all.
I thought the funeral one was interesting, as the iphone had lost the top spot around then IIRC - just to Android and not windows mobile, so the whole stunt was really just saying congratulations to google.
Throwing around the "cloud" buzzword, in spite of recent development on cgroups, CoreOS/etcd, Project Atomic and yes, even systemd, was funny enough. The paragraph on "security threats" was by far the most hilarious though, because you could tell they really had nothing worthwhile to say, but had to include it for the FUD value.
You also can't claim your code is developed with a "proven secure" process when you have to convince a court to let you hijack someone else's infrastructure to hack around the horrible security mess of your own products.
No I think I understand his point pretty well. The article's strongest arguments are interoperability with other Microsoft products. THAT's a weak argument. He's kind of parodying them. Even they tout Linux support as one of THEIR features.
> Will your Linux provider be able to support your entire infrastructure in the Cloud if the time comes?
We wouldn't put our entire infrastructure in the "Cloud", nor would our provider suggest it. Other companies have done so, though. While speaking informally, they're not happy with being "in the cloud" generally - regardless of provider.
> Does your provider offer a robust group of offerings, including Infrastructure-as-a-Service (IaaS) and Platform-as-a-Service (PaaS)?
I'm not very good at buzzword bingo.. but yes, they do provide their infrastructure and platform as a service.
The last outage was about 18 months ago, entirely our fault.
When was the last Microsoft cloud outage?
> Are they focused solely on the private cloud,
No, they're not focused solely on the private cloud - nor are Microsoft.
> and can they support Microsoft technologies for your existing Windows, Office, SQL Server and .NET users?
We have no need for supporting users of Microsoft technologies, so I really don't know how they'd be. Crap, I guess.
> Considering starting your next project on Linux?
Yes. Have you?
> Acquisition costs are a fraction of the Total Cost of Ownership of software projects. Will skilled talent be readily available if your project grows?
Yes, we have been successful in the past and we don't expect any problems in the near future. Personal note: Finding a good Windows developer out of the crowd of clowns is more difficult.
> What will support costs be if you choose to outsource?
This is not something we have considered.
> Could future migration costs be prohibitive?
We do have a back-out strategy. We're not locked in.
> Will patching and security lead to downtime?
No. We do not use Microsoft products in production.
> We encourage you to fully assess these factors.
Good advice, I will read more than just this "article" when deciding what platform we'll use...
> What does the vendor ecosystem look like for your proposed Linux solution?
There is more than one.
> Do your potential partners have a certification program for their professionals and how extensive is their experience?
To be honest, in my role I really don't give a damn about our "potential partners". Most non-windows shops seem to be adept, in my experience.
> Persistent threats and dedicated attackers can slow your projects and put your IT environment at risk with Linux projects.
As with all platforms, as Microsoft has experienced.
Microsoft seems to be touting their support for various Linux distributions on their server products. I can't help but be reminded of the massive miscalculation IBM made in OS/2 development: they went to great lengths to make sure it supported Windows applications, which simply encouraged developers to write interoperable Windows code rather than OS/2-specific code, despite the technical advantages of OS/2.
It's microsoft. They probably think their customers shouldn't know what a server is lest they cut into MS' license to print mone-...I mean, 'cloud' business.
Despite the page, I will mention this: corporate vendor support is sufficient for teams that don't have the technical expertise to either roll out a new technology or troubleshoot problems on their own. This is partly why Red Hat Enterprise Linux is so popular, among other reasons. When there is a technical, competent, and strong administration team in house, most of those concerns go away. When there isn't, licenses for commercial support make sense.
Ultimately, I wonder what development environment people prefer? For me, POSIX APIs and GNU/Linux extensions are far simpler to work with to do what I need and get the performance I want. Windows development just seems very heavy, and the APIs, while comprehensive, are incomprehensible.
edit: example: "Will patching and security lead to downtime?" There's ksplice and kgraft and more - patch the kernel with no downtime. And even without using either, I can't remember the last time I had to reboot immediately because a patch was that critical, or the last time I had to sit there waiting for my laptop to reboot 3 times trying to install the weekly updates. Windows Update blows.
edit: another example: "Proven security development lifecycle". Ha ha ha ha ha!