Over the last month or so I’ve been building a new blog site for my virtualisation discussions. If you are reading this post, then congratulations! You’ve found my site. I already write on storage at www.thestoragearchitect.com but I wanted to segment out the virtualisation part of my work and set some clear boundaries between some of my discussions, hence a new domain name, but similar design and appearance to the existing site.
Establishing a new blog also gave me the opportunity to look into hosting the site “in the cloud”. I’ve been experimenting with Amazon AWS, Rackspace, Nirvanix and others for some time, however I’ve not put any “real” data there. This hasn’t been from lack of trust but simply not having a good use case. Creating a new blog provided that opportunity and this post sets out some of the issues I experienced with the process.
Provider & Platform
At the outset I had two main decisions to make; which provider to go with and what platform to use as the basis for my blog. I decided to go with Amazon AWS as that is the platform with which I have the most experience. It’s pretty simple and as we’ll see in future posts, creating an instance is straightforward. In addition to provider I also had to choose a platform. For me personally, Microsoft Server platforms are what I feel most comfortable with compared to the various flavours of Unix/Linux. It’s not that I don’t or can’t work with Linux, I just prefer Microsoft and I’m very familiar with multiple site hosting on IIS. So Microsoft Windows Server 2008 it is.
Creating an AWS instance is pretty simple (I’ll cover it in future posts). Essentially Amazon are offering the same functionality that could be achieved by using VMware in the data centre. The difference of course is that AWS has bundled the services required to deliver on-demand virtual servers and more importantly to charge for them. In essence this is the benefit of using a cloud provider; on-demand provisioning of virtual servers and a well-understood charging model.
So back to the title of this post and the cost of using cloud computing over running services in the data centre. My current billing charge for running a single Windows 2008 Server instance is around $100/month. That’s mainly made up of the cost of the Windows server itself rather than the data traffic in and out. Annualised that translates to $1200/year, or over a 3 year period around $3600. Whilst this might be much cheaper than the cost of racking, provisioning, purchasing, managing a dedicated server with Windows licence, I suspect that looking at TCO in a virtual environment then the costs aren’t that competitive. Of course it depends on your organisation; the cost of that next server may mean purchasing a new data centre, in which case the cost is easily justified.
I wouldn’t move all services to “the cloud” at this stage as costs don’t seem to be that competitive (certainly with Amazon). However as competition increases, I’m sure this will change. Amazon do offer other billing methods (something for another post) so it is possible to obtain instances cheaper than I’m paying, but at the risk of operating a more complex service.
In upcoming posts I’ll look to talk more about how to establish services through AWS and probably most important, how application and data management should be handled.
- Netapp: The Inflexibility of Flexvols (9,970)
- Windows Server 2012 (Windows Server “8″) – Storage Spaces (9,412)
- Enterprise Computing: Why Thin Provisioning Is Not The Holy Grail for Utilisation (7,863)
- Comparing iSCSI Targets – Microsoft, StarWind, iSCSI Cake and Kernsafe – Part I (5,845)
- Review: Compellent Storage Center – Part II (5,480)
- Data ONTAP 8.0 – Part III (5,066)
- Why Does Microsoft Hyper-V Not Support NFS? (4,905)
- Back to Blogging (4,451)
- How To: Enable iSNS Server in Windows 2008 (4,352)
- Windows Server 2012 (Windows Server “8″) – Virtual Fibre Channel (4,183)