The Cloud Rains On Everyone’s Parade, Not Just Apple’s

The Cloud Rains On Everyone’s Parade, Not Just Apple’s

This morning’s Apple launch was like most previous similar occasions: as soon as the event finished, people raced to the Apple site expecting to see information about the shiny new products. Instead, for quite some time every page on the Apple site returned a bleaker message: “You don’t have permission to access [page]on this server”. Welcome to cloud reality.

To be clear, this wasn’t a major drama. There’s nothing unusual about sites either collapsing under the weight of a sudden boost in demand or becoming temporarily unavailable because of a foolishly switched setting: it happens all the time. Indeed, seeing the Apple store become temporarily unavailable is generally seen as a near-certain indication that new products are or pricing are on the way.

However, its absence is usually signalled with a neater ‘come back soon’ message, not a ultra-basic and inescapable error page. So on a day when Apple was talking up its newfound cloud capabilities, it also served up a timely reminder that while shifting services to the cloud brings masses of benefits, we can’t ever assume that cloud availability will be 100 per cent. And even organisations obsessed with control, like Apple, can’t always get it right.

The only safe assumptions are that the cloud will have outages; the relevant figure is not 100 per cent but how far below that number it will fall; and that the really relevant question for IT workers is: what happens when those services aren’t available? How quickly will they be restored, and will anyone be capable of productive work while you wait?

Apple’s site being down for a quarter of an hour or so won’t have affected its business, especially since every technology news site on the planet was busy reporting in detail what it had just announced. But a series of similar outages with iCloud itself could be more of a problem.

Since iCloud in its basic form is a free service (though you can pay for extra storage), it is unlikely to come with many solid uptime guarantees. It’s not in Apple’s interest for the service to run like a dog from the get-go, but with that said, that’s almost exactly what happened with its last attempt at a cloud-centric service, MobileMe, which was (to be blunt) a bug-ridden disaster.

In terms of offering a stable online service, Apple could still learn a lot from Google. But Google has frequent outages too. They’re a fact of life.

Evolve is a weekly column at Lifehacker looking at trends and technologies IT workers need to know about to stay employed and improve their careers.


  • That is an error message that occurs when you serve stuff through Akamai and they get an error on the Origin server, so it’s not so much the cloud going down.

    • But those kind of technologies underpin loads of what Apple (and everyone else) does. It doesn’t matter to me whether it’s Akamai’s or Apple’s “fault”; in practical terms, I can’t get to the service regardless. And if you have problems with a well-established service (like HTTP), then newer options won’t necessarily inspire confidence.

    • Really, this is just buzzword twaddle. Issues exactly like this happened with web hosted resources years before anyone used terms like ‘Cloud’, ‘SaaS’, ‘IaaS’, or ‘Virtualization’ to describe business practices. A web hosted resource crapped out under peak load, and the pretty error messages didn’t display because of the nature of the error. If Apple were hosting this website in their own comms room, or on a developers desk, it would still happen.

  • I despise the fact that the cloud “is coming”. The cloud is here; has been here for years, and will be here for years to come. “It’s coming” sounds like it’s a fabled switch that gets flicked. This is not the case.

  • Whilst cloud services will no doubt go down, they should not go down due to spikes in demand as Apple has just experienced.

    One of the benefits of cloud based web hosted services, particularly th Azure model (that Apple is rumored to use for iCloud) is vertical and horizontal scalability that responds to peaks and troughs in demand, and can price accordingly.

    With a flexible cloud service, server crashes under peak load will be a thing of the past. Accordingly, Apples main webpage went gone down precisely because it hasnt adopted the cloud to its fullest.

  • The (valid) point being made is that if critical personal or data or functions are put “in the cloud” (a.k.a. on the internet as the responsibility of a 3rd party), it comes with two potentially very large caveats. One, there will be times when you will not be able to access your data/function, and you personally won’t know it’s coming (compared to a local PC or network where although downtime is possible, it’s generally under some sort of control, and you have the chance to fix it under your own steam). Two, when such a service goes down, it will likely affect a huge number of people at once (compared with your own PC/network having local effects). Not an instant dealbreaker, but something that must be taken into account if the “cloud” is used critical things.

  • this is more of a problem with outsourcing to 3rd parties

    if you are already good at something (Google, apple, etc) then outsourcing to a third party (Amazon S3, Akamai) is bound to result in a worse solution than you originally had

  • All this sudden ‘OMG YAY THE CLOUD’ focus annoys me. The way the terms being used anyone thats had an email account in the last 20 years has been using a cloud service.

Log in to comment on this story!