Things I Learned from S3 and Apple in July 08
Apple has had troubles with the introduction of its new products in July 2008. While the iPhone registration glitch and the MobileMe start problems seem to have been resolved, behind the scenes there are hundreds (or thousands, who knows?) of iPhone developers who have paid money to have their apps reviewed and put into the App store, but Apple keeps their Apps “in review” for weeks on end (1 2 3). Sadly, I am one of them. Beloved Apple has gone from shining star to pretty buggy (or is it “pretty, but buggy”) in 2 weeks. What a pity. (The new iPhone still is wonderful - if you ignore that it crashes once a day on average.)
Amazon’s file hosting service S3 has been down most of today. While the majority of people still think of books when Amazon is mentioned, fact is that many Web 2.0 companies are outsourcing their raw data hosting to Amazon via a Service called “S3“. Well today it was down. And so was the data (images, files) of many websites (e.g. this one).
Lesson learned: Trust involves risk. Trusting a technology heavyweight involves risk that is beyond your control.
Dave Winer writes: “as probably a lot of others are doing today, I’m thinking of ways to avoid these outages in the future”.
I guess the future has to be: Make sure to put up your business in a way that at no point you rely on the services of a single company thus that their failure could seriously harm you. And: The fact that a company has provided excellent service up until now doesn’t necessarily enforce that it will do so in future.
Which reminds me of an interesting philosophical thought: You can’t prove that world is non-chaotic. The fact that it has been behaving according to rules up to a certain moment doesn’t entail necessarily that it will do so from that moment on.
And with this soothing thought I go to bed. The latter being a fact I might post on Twitter… if it’s working. It’s been having serious downtimes recently.
July 22nd, 2008 at 10:12 am
The place where I work is getting heavily dependent on IT. Your text made me think once again - it rang a bell. I just do not want think of a scenario where such a crash would happen here - hundreds of people, paid to come, not being able to be served. Shame international… Wow!
IT has to be used to its maximum, I think, but not naiively and completely dependent on. The most crucial things I would double-secure. In our case I would still recommend some paper and pen.