Putting a mule on a cloud: one man's battle with Amazon S3
Are we there yet?
The phrase “cloud computing” has been subjected to unnatural forms of marketing abuse. The buzzword pros have taken two nouns that we all recognize, combined them and formed a nebulous, hideous beast that means many things to many people.
Vague as the general cloud computing term is, there are places where you can encounter an actual, useful cloud service and begin to see where things are heading.
As the CEO of MuleSource, a maker of open source enterprise software, I’ve been dabbling with these early cloud/utility services looking for a distribution and cost edge. Or, if not an edge, at least something that’s easy to use and – god willing - more efficient than our current systems.
I’ll be walking you through some of our experiences with these services, particularly Amazon’s “cloud,” in a series of articles. With any luck, we’ll hit on what’s helpful and what still needs to be done in the cloud to turn it from vacuous to practical.
Journey to the Amazon
Back in January, my company reached its breaking point with managing file downloads and download sites. We decided to experiment with Amazon’s S3 (Simple Storage Service) to see if shifting to something resembling a utility service could make life easier. So far, we have been very successful.
We had been managing and maintaining multiple archives on multiple servers. After a while, a server crash, a disk blowout, lack of memory or some other fiasco kept reminding us why we hate computers.
This brings us to reason one for liking S3: you can stop buying and maintaining tons of machines - at least for file serving over the internet.
After several deep technology discussions conducted over games of Guitar Hero, we decided our main goal was to provide speedy file downloads with easy administration. There were several other things we were also looking for.
- Reduced hosting costs - It's
revoltingamazing just how many servers you have to manage in this day and age
- Easy access to storage in case we want to add more files - Even though disk space is cheap, you always need more. The idea of installing disk drives made us all want to vomit.
- Unique URLs for download tracking - We want to know who is downloading and to be able to see the associated statistics.
If you've been living in a cave and aren't familiar with Amazon.com, it's fast. [Well, you know, when it's up – Ed.] The file downloads from S3 appear to be even faster than the site itself. That’s the speed problem solved.