Sputtering Windows Instances

I had a concern about Windows OS being used for cloud computing.  The instances in Windows Azure take a significant amount of time to boot up.  In Amazon Web Services the Windows EC2 Instances also take a long time to boot up.  Compared to Linux, Windows takes 2-4x longer to spool up in the cloud.  (Compare a boot time of about ~1 minute for Linux in EC2 vs 8-15 minutes for Windows)

Before today, this just seemed like it might be a problem I was experiencing.  I tend to believe I’m doing something wrong before I go on the warpath, but today that concern that I’d done something wrong has ended.  RightScale posted a blog entry about the difficulties of Windows in EC2.  They’re seeing the same issues I was.

Another issue that they noticed, which I too noticed, was the issues around the clocks being off.  This is a similar problem to Windows being used with VMWare and setting up images.  The clock just doesn’t ync the first time, or subsequent times.  Usually a few manual attempts need to be made.

In another entry I caught another list of issues with Windows that Linux just doesn’t have.  None of these are work stoppage issues, but they are all very annoying and would push one toward using Linux instead if at all possible.

Putting Windows Azure and Amazon Web Services EC2 side by side Network World has found them to be on a collision course.

Boiling it Down, Where Does Windows Stand?

After some serious analysis by individuals of Windows running in Cloud Environments it appears that Windows just isn’t as suited to running in virtualized environments as Linux.  A number of friends have pointed out to me how much friendlier Linux is in virtualized spaces such as VMWare’s ESX Environment.

Also based on hard analysis of VMWare versus Hyper-V, the later doesn’t appear to be as sophisticated or capable of virtualized hosting.  Is this going to cause a price point issue for Windows Azure versus AWS EC2?  Just from the perspective of requiring more hardware for Hyper-V Virtualization versus VMWare & Amazon’s AMI Virtualization it makes me ponder if this could be a major competitive advantage for Linux based clouds.  Already there is the licensing price points, so how does MS own up to that?

I would be curious to see what others have experienced.  Have you seen virtualized differences that cause issues hosting Linux vs. Windows in VMWare, Hyper-V, or AWS?  Do you foresee any other problems that could become big problems?

IBM Cloud Services?

I’m kind of surprised, I didn’t even realize that IBM has made some inroads into cloud computing.  I however shouldn’t have been at all surprised, as IBM has made great changes over the years.  Their offering doesn’t seem as well laid out as AWS or Azure but it is a significant presence just being IBM.

IBM Cloud Computing looks like most of the other things the company does these days, which is primarily services based.  I don’t mean web services either, even though I’m sure they use those, but literally services based versus offering actual products, hardware, or some type of applications.  They do have their iNotes and other tools like that in the cloud.  Otherwise it isn’t immediately obvious how their cloud compares to Azure, AWS, vmware or Force.com.  IBM’s website for cloud computing seems as disorganized as vmware’s or Force.com’s, which make Azure and AWS seem super well put together and structured.  With clearly defined features and tools for developing in their respective clouds.

Overall the services look attractive if IBM is already a provider of other services.  Kind of like Azure looks more attractive to the .NET Stack of Developers and AWS looks more attractive to PHP or Java Developers.  Even though both Azure and AWS support .NET, Java, PHP, and more, just from the IaaS or PaaS perspective though.

This begs the questions:

  • What are the differences between the IBM Cloud Computing Services, Azure, AWS, Force.com, vmware, and others?
  • Besides IBM, Azure, AWS, Force.com, and vmware, what other cloud services are really doing it right and pushing forward with redundancy, uptime, compute capabilities, and other key features?
  • Anyone guessing on consolidations in the future with the current big players (AWS, Azure, Force.com)?