Blog on website

I am going to keep this page up (for now), but I am using my website at http://www.tintedgreen.net/ for blogging.

Feel free to follow me there. :)

At the Atlantic Eco Expo

September 18th and 19th I participated in the Atlantic Eco Expo to show off my new desktop virtualization in a box prototype. It was great to actually show this to people who had never heard of anything like this before (which was pretty much everyone). People really seemed to like the idea.


I introduced a lot of people to desktop virtualization. I described how this one tiny machine built with a few hundred dollars worth of computer parts could run dozens of thinclient computers using virtualization and the Linux terminal server project.


All of the computers I used in the demo were old laptops that my friends had abandoned. They ranged from 3 to 5 years old and all ran great the whole time.

This message attracted a lot of people to the booth. This machine was connected to a Windows 7 virtual machine. :)

Besides resurrecting old computers, what else can this system do? Here is a video posted by Adam Hartling (@XenosNS on twitter and http://acoss.wordpress.com/ for his blog). Enjoy.

This system is very easy to deploy and manage and creates a great user experience. I will be posting more video of this system over the next few days.

Mike Lovell and the Quest for the 95W Hexacore CPU

Whoa, it's been a long time since my last post.

Anyway, I'm still trying to lock down the hardware on my "desktop virtualization in-a-box" prototype lovingly codenamed greenBox. The specs so far are:

AMD Phenom II X6 1055T Thuban 6 Core Processor 2.8GHZ Socket AM3
Shuttle SA76G2 Black Barebones
Western Digital Caviar Black 1TB x 2
OCZ Platinum XTC PC2-6400 8GB 2X4GB
Possibly a second NIC

Now, there is a snag I've run into. The little shuttle motherboard here has a max TDP rating on the CPU power of 105W. Unfortunately, the TDP on most Phenom II 1055Ts is 125W. Layman's speak, this setup doesn't have the juice to run this processor (that's as close to layman's speak as I can get).

Notice, I said most. There is a version of the 1055T with a TDP of 95W which should work fine in this setup. The reduction of heat would also be great in this small build. Unfortunately, I have no idea where I'm going to get one, yet.

Is the whole project in jeopardy? Not really. If I have to, I'll use something else. I just really think this CPU would be the best for the price range I am aiming for. So, I find myself searching for a rare version of a CPU. Not exactly something new to me. I'm very picky about my computer hardware, so I'm use to bugging retailers about a particular version. I just hope I can get it in time.

greenBox update

I have been working on the software side of my greenBox Project as of late. I just wanted to make a quick update on my progress.

I have the setup working pretty much the same as it did in Ubuntu 9.10 on Ubuntu 10.04, and, I have to say, Linux Terminal Server Project (LTSP) seems to work even better in this new version of Ubuntu. Loading is significantly quicker and the overall feel is just smoother.

On the features side of things, I now have sound and disk redirection working when connecting to the Windows 7 virutal machine. These things worked already on the LTSP, but I had not added them to the RDP connection. I have been testing sound out with YouTube videos and Windows Media player and it works really well. I have actually been enjoying listening to some music playing from my thinclient test computer while typing this post. :)

With disk redirection, it is now possible for the user to connect a USB drive or put in a cd and read the files. Simple, but a feature that is completely necessary.

For those that haven't seen my previous setup demo, please check out my YouTube video. http://www.youtube.com/watch?v=_q_l8p7fn3Q&feature=player_embedded
Now, picture the exact same thing with sound. I hope to make a new video with the new features sometime soon.

And now for something I have not said this in this blog that I think is worth saying...

I don't think anything I'm doing here is overly unique or even hard to produce. I think what makes this project special is my target audience. I'm aiming to make computer maintenance and deployment more cost effective while giving a better user experience in one of the most overlooked sectors of IT; public access. These are the places with some of the worst funding (especially with the federal government cutting so much funding to CAP programs nationally), worst computers, and a lack of technically knowledgeable staff (don't get me wrong, there are some amazing IT people working in public access, but they are stretched very thin and are often underpaid). If I can build an all-in-one solution that is cheap, easy to deploy, and environmentally friendly, I think many people will be interested. Even people that cannot afford to buy a computer deserve a great user experience with the latest software.

That being said, there will be an opportunity for people to see this system working in the future. My current plans are to have the full prototype working for the Atlantic Eco Expo in Halifax in September. Stay tuned for more updates and check out the Expo at http://www.atlanticecoexpo.com/ and on twitter at http://twitter.com/AtlanticEcoExpo

I hope to see you then. :)

Start: Project greenBox

I decided to give a little "code name" for my desktop virtualization in a box project. I don't think I will keep it if I ever choose to sell it as a real product, but I need to call it something, for now.

I began the first step today; software. I am now building my previous Linux Terminal Server Project (LTSP) setup from scratch using the latest version of Ubuntu (10.04).

After the software is setup, I plan on building upon what I have and create a layer over the whole package to make it easier to manage. Setting up new virtual machines currently involves editing text files and using VMware products. I would like to create a web-based interface to streamline the setup and maintenance of this system.

I'm still working on the hardware specs of the box, but I think I have a pretty good idea of what it will look like. I hope to start buying hardware for the prototype early in June.

Affordable Desktop Virtualization in Box...

That is the key here; to create a system that would be cheap and reliable. Could this be achieved using consumer hardware?

AMD has just released their 6-core Phenom II processors. Utilizing these cheap and powerful CPUs, I believe I could build a server powerful enough to meet the requirements for a small desktop virutalization deployment.

Combine one of these CPUs with 8GB of ram, 2x1TB hard drives in Raid, and a shuttle barebones case and you have a small desktop virtualization server for less than $1000 in hardware. At that price point, it wouldn't be a big deal to add additional servers for greater capacity and redundancy.

My next step is to build a prototype for testing.

The black box

Check out my Virtualized Desktop Demo here (if you haven't already):
http://www.tintedgreen.net/?page=article&id=18

The question I have been trying to answer lately:

Could I sell this type of desktop virtualization system as a product to public access sites?

The idea is to sell a "black box". The customer would get a very small server that is ready to plug into their network and start running virtual machines.

This has a couple of advantages. It does not require us to go on site to set the system up. However, we can still login to the server remotely and setup or fix anything the customer needs. We can still have complete control (if the customer chooses). The customer gets full documentation on how to add computers, modify settings, and maintain the system.

I think the key to this is allow the customer to have a much better idea of the costs of this type of system. They pay the base price for the server and initial setup. We could then provide a service plan after to provide tech support and maintenance.

This system would also allow for a lot of expandability. Extra servers could easily be added to provide more capacity or redundancy.

My goal right now is to build a prototype. Goals of the prototype:
-Capacity for 20 windows machines each sever (plus more with Linux Terminal Server Project).
-Build it very small. I want to go small form factor to reduce shipping costs, environmental impact, and avoid adding huge server setups to public access sites.
-Build it cheap, but reliable. It has to be inexpensive, but still have enough redundancy and stability.

I want to be able to price the server and setup and one year of tech support for a reasonable price. What I need now is feedback from public access sites that might be interested in such a system to find if this would work for them.