Google dark cloud

Why Gmail Failed Today:

Gmail, which recently passed AOL to become the third largest Web mail service in the U.S., is obviously having some growing pains. A few hours of downtime is not the end of the world, although it might seem like it at the time. It just better not make this a new habit.

The main issue is that Google obviously has to go down less often. But it's never going to be perfect, that's reserved for God. So the question is how often can it go down without people getting angry? It isn't as if not-cloud applications don't fail, we all know of many instances when computers won't boot up, applications won't work where restores, reinstalls and replacements are necessary. That's why many people have external hard drives which they use to back up their data. It seems the bigger issue is synchronicity and synergy*. If your computer at the office is having issues, others might be doing fine. On the other hand, if everyone at the office is using Google apps then a failure might render the whole office inoperable. One long-term dynamic will be that as Google gets more popular and widespread, the less frequent these outages have to become. Otherwise, Google apps might end up in the Dilbert strip.

* I don't mean synergy in a good way. Rather, if/when Google goes down in a Google dependent office that's something you can talk about as a shared experience around the cooler as you wait for the apps to come back online.

More like this

The majority of comments on slashdot pretty much nailed this. GMail's uptime is a hell of a lot better than any other email system I've seen (including the ones I've setup and administered).

The poo-pooing in this case probably has a lot more to do with a lack of control than any real technical concerns. (A sort of fear of flying vs driving effect.) Yeah, when a cloud resource goes down the effect is more widespread, but what is the actual impact of that? Are those supposed synchronicity effects actually worse than more frequent less synchronized failures? Seems more plausible they would actually have less of an overall impact.

Single points of failure are something to avoid. Cloud computing eliminates (or mitigates) a lot but not all of those failure points.

BTW: Before you decide that Gmail is too unreliable, try actually setting up a reliable system yourself. It isn't that simple. You need multiple servers in geographically distant locations hooked up to different segments of the backbone to start with.

BTW: Before you decide that Gmail is too unreliable, try actually setting up a reliable system yourself. It isn't that simple. You need multiple servers in geographically distant locations hooked up to different segments of the backbone to start with.

that's kind of a typical /. type of response. who the fuck cares that doing this hard? google is now selling itself as a mission critical service. technical difficulty is irrelevant, the reliability has to match marketing. end-users don't give a shit how nasty the sausage-making is. most /. readers gave no slack to M$ during the bluescreen of death era. so what if an OS designed top-down to satisfy everyone is really hard?

The big problem was people getting their older email. I think most could handle the hour delay on getting new email. Not having access to email period was nasty. Which is why I only access Google Mail via IMAP under Apple Mail. That way most of the mail is stored on my computer. (And I back it up nightly)

Clark,

I hope you realize that you could have accessed your gmail normally, just by using the gmail gadget, from your iGoogle page?

I'm speaking more generally about becoming dependent upon resources in the cloud.

While the Google outage was annoying there are lots of ways you can lose access to cloud resources. I didn't know you could access it by avoiding the web app. But the principle is important. My business partner has actually more or less completely converted over to using Google mail and was completely down during the outage.

Uninterrupted Gmail = "mission critical", really?
Me thinks this is perhaps some new meaning I am not familiar with.

If you have a "mission" dependent on not having any disruptions in email, that you are doing it very wrong. The "typical \." response is an entirely appropriate dose of reality.

In contrast to Windoze (which deserves to be brutalized IMO), Gmail is setup in a very sensible manner and works amazingly well. Constructive criticism is all good, but a bit of perspective please.

BTW: How about clarifying how synchronicity in email disruption is such a terrible thing over and above 'normal' email disruptions.

dude, it's not just about email. this isn't /., stop arguing like it is. the cloud is to replace MS office. if it does want to do it, clouders should expect people to bitch when it isn't working right. whining about how hard this all is is irrelevant.

Travc, if I'm going to run a business on Google's tools, as they've been advertising, I can't be having it go down for hours on end. Now Google's been overall pretty good. But if this happens a few times expect a lot of people to start switching.