On Cloud Computing

While I typically blog AFTER the students have blogged, I felt pretty inspired today and I thought I would pen down my thoughts before I forget them.

I would like to thank Simone Brunozzi from Amazon for taking time to come talk to the class about Amazon Web Services (AWS) and Cloud Computing in general.

The reason why we are deploying the class assignments on AWS and why I invited Simone to give a talk to the class is because I believe that cloud computing is future (and Amazon didn't pay me to say this). I'm saying this not because the technology is cool, but because it makes business.

Simple reason: it's annoying (and expensive) run a data center. :-P

I must admit that while I have a reasonably good idea of how the technology works 'cos my area of research is networking and distributed systems, I had never used a cloud computing service before until this semester.

I believe that it would be helpful for the students to understand how it works, because while many aspects are the same as traditional servers, there are also key differences. For example, persistent storage is something that requires a little more understanding and thought, i.e. should we put the data in S3, EBS or MySQL?

While students are expected to work mainly with EC2, I hope that the students will also use the opportunity (and free credits) to explore the other AWS services that Simone mentioned.

I'm not going to talk about the technical content of the talk since that's for the students to talk about and discuss. There are three things that I would like to highlight.

One, the conversion to cloud computing will not likely be an easy one. The reason is that people don't necessarily change just because something is better. Small companies probably don't need the full capabilities of the cloud; large companies are resistant to change.

The resistance is more often than not a human rather than a technical issue. Basically, change is scary because things can go wrong. Middle management in Asia is typically conservative.

This is the question: suppose they convince their boss to switch over to cloud computing and their company saves a lot of money, will they get a bigger paycheck? Suppose something goes wrong in the process - who's going to take responsibility?

The lesson here is that it's not enough to big something bigger and better. You've gotta convince the sucker to buy it. :-)

Next, I really really pleased by how Simone answered some of the questions. Basically, to almost all the questions of "how do we do it?", his answer was "it depends".

This underscores the importance of being flexible and understanding that life is complicated and that there are often different answers to the same question "what is the best way to make data persistent on AWS?" under different situations. In a nutshell, context matters.

Finally, I was very impressed by Simone's story about him advising the customer who had 200 EC2 instances to run his own distributed file system instead of using S3 or EBS (which would cost the customer more and increase AWS's profits).

This underscores the importance of integrity in doing business.

Many of my students will probably eventually end up starting their own businesses.

My wife does not have very many good things to say about the sorts of people she has come across in the corporate world. She said that integrity is severely lacking.

I hope that however that my students will learn from Simone's example and "do the right thing". Don't screw the customer to make a quick buck. Create value and build real businesses. :-)