the stinkin’ blog

When I moved to my new position with Canonical one of the immediately noticeable items was the lack of a collaborative system (never really had a need due to their size). I am used to using a collaboration system of some sort even if they really suck. So I needed to figure out how to bridge a few gaps and be able to use my crackberry with thunderbird and allow my fellow employees to view/share my calendar.

The solution, tie Google Calendar with Thunderbird through Lightning and the Provider for Google Calendar. My first attempt failed because I was missing a library - libstd5++. After adding gcc 3.0 to my system, I had libstd5++ installed. I then added the latest Lightning and Provider for Google Calendar add-on and I was good to go after restarting Thunderbird.

The next step was simply creating a new network-based calendar, choosing Google and adding the private XML URI (from the Google Calendar settings) to Thunderbird, entering my credentials and voila, shared calendar shows in Thunderbird!

Now I can use Google Sync on my crackberry and sync it all together very easily. May seem difficult, but in reality it just required a bit of setup and I can now enjoy the freedom of Thunderbird running on Ubuntu tapping into the wonders of Google Calendar. Need to see what I am doing today, tomorrow or next month? Just ping me and I’m happy to show you!

Now here is an interesting view. A writer on GigaOM writes that he feels cloud computing will be vastly different in 2018. Ok so that is a given since it is 10 years away. I sure hope it is different. While I agree on some points, I have to digress on a number of points.

I do agree that cloud computing vendors (those that are providing the “space”) will change dramatically in the next 10 years. They will need to address the needs of the “end user” no matter who that person or organization serves. They will need to be easy to work with, provide significant service and be available to everyone. The best ones will be able to charge premiums for value added services and the worst will fall by the wayside. That is easy to see.

However, I do not agree that the user experience will be an “Apple type” of experience. While I agree that Apple does well with eye candy on their Mac and iPhone, I despise iTunes. The author seems to think that iTunes is elegant and successful. I am so sorry to inform you that the only reason iTunes is successful is because the iPod is successful. You see iTunes is REQUIRED for the iPod. It is forced upon anyone that wants to use the iPod or iPhone for that matter. So one can hardly say it is successful by itself. The other term used is “elegant” and while I agree that the iTunes application itself is fairly pretty, it is not altogether very useful and it is a pain in my rear. Why do I think that? Well, every time I go to use it, Apple forces an update on me and with that update they require quicktime (which does not save my settings!!!) and keep pushing Safari on me - I only have one windows box in my house, no macs and I HATE how Apple hawks their wares on my useless windows box. I only want iTunes and I only want it because they force me to use it to make my (free) iPod shuffle work right. Ok…so much for my complaining.

While I do not agree with the Apple user experience notion, I do agree that cloud computing will be vastly different in years to come. I know that users will force companies to change their experience and I also know that the cloud vendor with the best user experience, best presence, and best service will win. That is a given.

The question will be who will that company be? Just like we are waiting in the wind to have a better virtualization experience we will be waiting for the consolidation of the cloud computing environment for several years to come.

A writer on OStatic writes an interesting point. One that I have been pondering for a while since doing some catching up on the emerging trends around cloud computing. If open source is such a win for the cloud computing infrastructure and those that are building apps, why then are we seeing so many proprietary “clouds” which offer no interaction? Federation of clouds is the key to maintaining an edge by ISVs.

A ISV cannot afford to put all of their eggs in one basket. A great example is for those that use Amazon and S3. When S3 went down a few weeks ago everyone using S3 went down with it. Nine hours of downtime for a small company could run it into the ground quickly. The question is how do you prevent outages from affecting your application? Federation would be an answer.

Most cloud computing vendors like Amazon, Google, and others do not interact together at all. If I want to host on Amazon ECS and on Google Apps I am forced to write two different “standards”. While there are open standards that are being built for cloud computing collaboration we need to devise a way to allow cloud users to distribute applications amongst a number of providers. The best providers will stand up and can charge a premium whilst the worst ones will go away (either through acquisition or simple death).

But there is more than simply calling a cloud federated. There are a number of issues that need to be addressed such as authentication, data storage and syncronization, and application communication in addition to being able to run the application in a number of unrelated cloud instances.

Interesting issues. Time to find out more.

HP, Intel, and Yahoo today announced they were forming a consortium aimed at performing research on cloud computing. This initiative is part of HP’s Shane Robison’s view of “Everything as a Service” which is quite the interesting concept.

Personally, I think we are a little ways off from the singularity he is predicting where everything I use talks to everything else I use and they all think like I think, but the uses are enormous. The one drawback is that if “my cloud” is doing all of my thinking what is left for me to do? Will that make me more productive and allow me to focus on the task at hand? Or will it simply make us all Alzheimer’s patients?

Food for thought.

An interesting article that does a good job at providing some comparison of a handful of cloud computing models.

What makes cloud computing rather interesting today is the fact that the “components” of cloud computing is really being driven by open source. Obviously Microsoft is getting on the bandwagon. The laggards have to follow what is being hyped. The question remains though - Will the community rally around a compelling concept that can propel open source into what I would call the next iteration of a network? or Will Microsoft take what could be a real opportunity for everyone and make a Microsoft standard and Microsoft-isms to “cloud” cloud computing?

Time will tell, but I can assure you of one thing. Open source has an opportunity. Open source has a community. It is time for the community to show their true strength.

Citrix recently unveiled Project Kensho aimed to after the virtual appliance market with their XenSource hypervisor. While VMWare is pushing the Open Virtual Machine Format (OVF) Citrix is jumping on the bandwagon with Project Kensho to help push the standard forward and get them a head start in building virtual appliances on the Xen hypervisor.

It seems though that Red Hat and Ubuntu are pushing KVM where Novell and Citrix are trying hard to keep Xen in the mainstream. Microsoft on the other hand is driving their own standard with VHD. More “format” wars. At least now there can be even more niche players who tout working with the various hypervisors as the hypervisor begins to take over as the platform leaving the OS behind in the dust.

I will be interested to see where Citrix goes with this since the virtual appliance market is heating up especially with the marketing behind rPath driving virtual appliances, SaaS, and cloud computing into the forefront.

In my quest for knowledge surrounding cloud computing, I have come across some interesting notables. The first of which are the basis for a cloud compute type application which is logically an appliance of sorts. There are a number of vendors in the Linux/OSS space who are creating JeOS (juice) type instances to be used as an appliance. I have looked at four, rPath, Red Hat, SUSE, and Ubuntu.

Arguably within months after VMWare’s CEO mentions JeOS in this blog, Ubuntu created a JeOS for use within VMWare’s VMDK. rPath was already doing appliances. Red Hat and SUSE are currently in a beta project with their appliances. I understand how Red Hat and SUSE make money - from subscriptions and support. rPath appears to charge for the developer kit (rBuilder) as well as charge for support and development services. rPath may even make money off of each appliance distributed since they provide a kill switch inside their kit. However, how does Ubuntu or Canonical make any money?

Mark Shuttleworth has vowed to keep Ubuntu completely free. How can one monetize free? This is the $1m question (not that US$ are worth much today!). Canonical is not large enough to have enterprise support offerings that rival Red Hat or Novell. Canonical does not have a developer kit like rPath. Can Canonical make money off of the appliance? Maybe the only way is to build a developer kit and resell it? Does Canonical have the manpower? or the time to build and manage such a beast?

I have yet to see any appliances that advertise using linux underneath - of course I may not be looking hard enough. Let me state that I have not found any mainstream appliances that fit the previously mentioned bill.

I will be diving deeper - for good reason. Stay tuned.

People Over Process » Hyperic CloudStatus - Starting the Ball Rolling

I’m going to start posting some interesting articles and such for Cloud Computing. Here’s one…Hyperic’s status app for EC2.