I first encountered “the cloud” back in ’88 when working for a network comms company and back then it was a pretty “out there” concept. The internet wasn’t even a word we knew. Then sometime in the last 10 years Amazon Web Services started offering file hosting with complicated API systems to save and retrieve your files. I wasn’t in the “large file” game so I took note and moved on.
Roll onto 2010 and I saw a tweet from Dave Winer promoting DropBox and so I took a look. Around that time he was talking about using DropBox for hosting files used on his websites and I did wonder why you’d bother when hosting accounts are typically generous with disk space (a change from 2000). A quick search revealed that DropBox weren’t the only players in the market and that the offerings were all fairly similar. If necessary I could have 2 gigs here and 2 gigs there and all for free.
I needed to extend some GreaseMonkey scripts around the same time and instead of uploading the famfam icons to my site I just popped them into a public folder on DropBox. What I’ve noticed is that if the images need to reload the images on my site are there instantly but the DropBox images have a noticeable delay. It wasn’t a problem for me but I took note and was happy to use DropBox for file serving and backing up my work files but not for anything where delays would be an issue.
Roll on a few months and Dave Winer was backtracking and saying the cloud had problems for web hosting. I smiled quietly to myself.
So this morning I wake up to find DropBox has hit the 25 million user mark and without actually advertising – just by using referral codes and relying on the love of the freebie. Michael Woloszynowicz looks at the economics of it and spills the beans that DropBox use Amazon Web Services. Well, I could have told you that because as Mashable were reporting Amazon’s server outage I was seeing my GreaseMonkey scripts failing to load the DropBox hosted icons.
I have no idea what caused the Amazon outage and I’m sure there are engineers reviewing it but the company is huge and will have redunancy systems in place to stop such an event. That one happened implies a significant problem. For the rest of us its a quick lesson in who we put our trust in. We choose our web hosting company and most offer 99% uptime and we demand they meet those targets. We avoid resellers and like to know that the hosting company owns and operates their servers.
So why would we entrust our files to a “cloud” repository when we don’t know anything about the hosting, how they’re making their money (after all you get 2gigs free) and who is responsible for the uptime.
I think the future of “the cloud” will be interesting and there may come a time when my mistrust is seen as quaint. Until then – look closely at what you save into the cloud and just what you expect it to deliver.