91d437f0fb56cfb5b64de81b60bd6738

One of my colleagues was impressed by the 3D desktop, driven by CompizConfig on my Ubuntu 10.04.

He then asked me how to install one on his ThinkPad T43 machine. Giving the steps:

1. Find a 2G USB key. Click Download @ http://www.ubuntu.com/desktop
2. You’d be given an ISO file. Then follow the step 2 & select “Create a USB stick” @ http://www.ubuntu.com/desktop/get-ubuntu/download
   
3. Restart your Windows then boot from USB key
4. From the same page you can either run Ubuntu from USB to start the installation
   

91d437f0fb56cfb5b64de81b60bd6738 ->

$ echo “a man on linux” | openssl dgst

or

$ echo “a man on linux” | md5sum

The debate occurs quite often on how to stack software in Cloud Computing solution (specifically Development & Test Cloud), amongst Service, Software & STG. I don’t complain the variety of software stack, as I like the more choices. I’m going to write out how my architectural solution looks.

I have one philosophy: When I could use OpenSource + Pure Blue software in my solution, I will do it without any hesitation.

IBM Development & Test Cloud solution is based upon Tivoli Service Automation Manager v7.2 for now. There’re 3 types machines in its Cloud environment. 2 of 3, in Managing environment , are Management Server & Administrative Server. The last one, in Managed environment, is the managed system(s).

My only rule to follow is to check the product infocenter, which tells the truth of technology & product. The one is http://tinyurl.com/tivsam (I talk about x only for my current engaged case @ one Telco customer in Beijing)

For Management Server, we have 2 choices on 64 bits x – RHEL5.4 or SLES10.2

For Administrative Server, SLES10.2 is ONLY our choice for Linux. Obviously Microsoft Windows isn’t in my list either. The reason is clear.

For Managed environment on x, we get rid of VMware and we have RHEL5.4 and SLES10.2 (CentOS isn’t in my list, due to lack testing & support)

To combine the above consideration, I use SuSE Linux Enterprise 10.2 in management environment and Redhat Enterprise Linux 5.4 in managed environment.

Source & Credit: http://www.ibm.com/cloud…

What are the specifications for the sizes of images?

For the Beta (Poughkeepsie) environment:

For the Pilot (Southbury) environment:

Installed Eucalyptus on Ubuntu Karmic 910. Capture screenshots

I placed a lot of effort on Cloud Computing engagement and had tours around the country. Piling my mileage

First off of everthing, I demo my compiz 3D desktop enabled Ubuntu 9.04 to show off how we live in a pure open source environment without having to touch any of Microsoft.

Get back to our topic. Last week, I visited one of largest mobile device / equipment manufacturer to discuss how a Cloud Computing could transform customer’s IT. We spent an hour to talk about Return of Investment (ROI). To list what are supposed in consideration of an ROI

• Hardware
• Image Instance – build, migration, administration
• Software – non-free operating system and hypervisor. If Xen, or KVM being used on LAMP, the cost is close to zero. And all software to build a Cloud environment with licensing and maintenance.
• Labor – build and operate a Cloud environment
• Provisioning Process – time taken to complete provision

Input the above consideration comes out in 3 charts as the following (the data is under a certain assumption and might vary in different environment) :

• Comparison of without- cloud and with- cloud

• Categorized

• Cost in Transformation

The information reflects a scope of ROI to be analyzed and doesn’t have to mean this is a mandatory.

I discussed the subject with one of colleagues who is a Cloud Computing architect. I’m logging what my comment here

=-=-

To continue our previous talk is a huge discussion. I’d recommend you could start to get involved in http://twitter.com and follow

http://twitter.com/mashable
http://twitter.com/j3ffyang -> me
http://twitter.com/OpenIBM -> I own this too as of being admin for IBM Open Source Global community

When you hit any of above, you can see the bytes come from http://aws.amazon.com actually, which is an IaaS with no doubt. Twitter is a kind of social collaboration network -> an application running over Amazon Web Services (AWS). Twitter doesn’t own any hardware resources. Whenever Twitter needs computing capability, it goes to AWS and AWS fulfills its request… on demand. Even though Twitter sometimes out of service due to its overload.

@ IaaS, AWS not only gives power of hardware (CPU, memory, disk and network…), but also provides plugin / API to connect Twitter with Hadoop… and Simple Queue Service (SQS), and Simple Database (SDB).

You shouldn’t be surprised @ Twitter’s power to gather thousand of thousand developers around it, if keeping our eyes open. (Some reasons of the motivation of social collaboration in term of Web2.0). See these:

http://tweetwheel.com
http://ftags.com
http://tweetvalue.com/
http://www.tweetizen.com/
http://twittersnooze.com/
http://mrtweet.net/home/j3ffyang
http://tweettrail.com/search/hadoop
…
…

This list is almost endless. They’re all independent of twitter.com. The above are all SaaS, plugable into Twitter.com where provides API -> http://apiwiki.twitter.com/ and http://twitter.com/downloads – @ PaaS

This is an ecosystem.