Last night I attended Cloud Camp Boston and Microsoft’s R&D space. CloudCamp (for those of you that don’t know) is a hybrid unconference / conference.  There were basic sessions at the beginning and then it moved into an unpanel discussion finally culminating in unsessions.

To my disappointment most of the users that were there were brand new users or not actual users (yet).  Many were interested in why they should use Cloud Services, or what applications were best to put in the cloud.  Sadly I didn’t get many ideas from them.  The remainder (and majority) of people there were providers of cloud services, and although normally I would run away from vendors I was able to glean some great ideas that I think I can take back to academia.

Cloud Computing is a Business Model

That’s right cloud computing is a business model.  A lot of people in the Cloud Computing community are talking about Cloud as a commodity and cloud in terms of business models.  Simon Wardley had a great presentation at OSCON 2009.  If you have 14 minutes I recommend watching Simon’s presentation.  It’ll help you understand this concept.

The problem is that the leaders in cloud understand that cloud computing is a business model and that computing itself is moving toward becoming a commodity; but the people that would actually go forward and implement those systems don’t.  The positive is that in the business world they understand that saving money is good.  And since people keep talking about how the cloud can save them money, leaders decide to move applications and services to the cloud.

However when you move to academia money isn’t necessarily the driving factor, although that is less so the case these days given the state of endowments.  The truth is that campuses aren’t interested in improving themselves via a business model.  I think it may be helpful for members of campus administration to see how this business model can help them accomplish academic goals.

I’m hoping to flesh out this idea more, perhaps look into more about cloud computing as a business model and determine how it could translate to academia (and luckily I have a husband in an MBA program that can help me with this).

Cloud Computing can Improve Efficiencies

Some might argue that this is part of my first point and they would be right, but this part of the model I have fleshed out in my head, so I’m going to write on about it.

When I first began looking at the cloud I saw how it could be very useful for development purposes.  Last night I was sitting next to a woman from TnR Global.  She was speaking about how her clients are finding it easy to move from development to production in the cloud but that cloud providers have to be wary about pushing all applications into the cloud, because some just aren’t suited for the cloud.

This gave me an a-ha moment.  Currently I’m the systems administrator for my college’s learning management system, Moodle.  Let’s pretend I decided to install and test Moodle (true story).  We run a pilot for it and everyone loves it (also a true story).  So now I want to move it into production, but it can’t sit on the small server that it currently sits on.  Luckily I’ve decided to do all of this via Amazon’s EC2.  So I take my back up (which I have saved on S3) and quickly move it over to a larger server (or in Amazon language instance).  In 5 minutes I’m in production.

Now I actually just did this myself on regular old servers; I moved Moodle from an older development server to a brand new shiny server that I purchased for it and the whole process took me forever.   I had to purchase the server, receive the server, set up the server, install Moodle, blah blah blah.

Now let’s pretend its a year later and we want to upgrade to Moodle 2.0 because we hear that the grade book is better (also a true story).  With my current version of Moodle I have to take the server “down” (I don’t want people playing while I’m upgrading) and upgrade the software.  Now if Moodle were in the cloud, I could take my back up and move it over to a new EC2 instance upgrade Moodle and quickly switch the IP address from the old version of Moodle to the new version of Moodle.  No downtime, everyone’s happy.

I just had a hallway conversation with my boss and he said he wishes I had more time to work on more projects, but unfortunately I don’t.  However, if we were using the cloud I could improve my efficiency and perhaps have more time to do the work that he would like me to get to.

Academia Needs a Private Cloud

Last night Christofer Hoff (Director, Cloud & Virtualization Solutions at Cisco Systems) defined a private cloud as something where you control the policies for the cloud and where there is some sort of self service for “users” (in this case a user is a sys admin).  This of course is a dumbed down definition of something a very smart person said, but I’m just going to go with that for now.

Academia has set up consortiums to deal with cost sharing and resource sharing for years and years now.  And there are organizations that are working together on what essentially mounts to development sharing.  In fact, Educause Quarterly had an interesting article that mentions Smith College’s hosting of a Moodle Hack Docfest.  This idea of resource sharing is not new to academia, so creating a private cloud for Academia makes sense, after all aren’t servers a resource?

I met a woman from IBM who told me that IBM is thinking that colleges and universities need to use these consortiums and cost sharing organizations to create their own private cloud.

I’ve seen the need for this within my own consortium.  Recently I’ve been working on a number of projects that other libraries in my consortium are interested in.  Unfortunately they don’t have the ability to purchase servers, nor do they have the IT staff to maintain the servers even if they could afford them.  But I have the ability to.

So let’s say that I install a library research guides application (true story) and another college in my library consortium is interested in having the same application (also true story).  If we were in a private cloud, I could take my back up of the application and give it to them.  Or better yet I could create a template of the server (in Amazon land that would be an AMI) and spawn it for them.  They could then go ahead and make the changes that they would like.

The best part of the private cloud is that you have a community to draw from that is using the same systems that you are, so if that other library got into some trouble they could come to me and ask for help, which is something we in academia are already doing.

So the short of it is that I’ll probably be emailing that woman from IBM in order to get more information on private clouds and how IBM is working with educational institutions interested in them.

Conclusions

So although I left Cloud Camp feeling a bit dumb (and that’s probably because Christofer Hoff was in both sessions I attended and he’s REALLY smart), I did get some great ideas that will help me develop training and sessions for my Cloud Computing Fellowship.