Does Gmail’s Crash Forecast “Dark Cloud Computing”?

By Phil Butler
We teamed up with SiteGround
To bring you the latest from the web and tried-and-true hosting, recommended for designers and developers. SitePoint Readers Get Up To 65% OFF Now

gmailToday started like any other, with me searching for pertinent news for the SitePoint readers. Then it happened, Gmail crashed. Suspecting it was a client side problem, I spent the better part of an hour trying to figure it out, when a familiar “server error” message appeared. Not an hour after the outage occurred, the news was on every major outlet announcing a possible worldwide system failure. This is not the first time Gmail has flopped, but today’s episode appears to have been much wider spread. Interestingly, the event may be more significant for computing in the cloud than it is for email.

Cloud computing has an equal share of detractors and proponents these days. Google and others have been trying to sell “cloud based” apps like Gmail to the mainstream for years. But, today’s incident added fuel to the fire of criticisms that for many operations, cloud computing is simply not feasible. Gmail is currently the third most popular online mail service behind Hotmail and Yahoo!, with between 95 and 110 million users according to most reports.

Ironically, Google utilizes Gmail for much of its communications, so users could not even effectively complain about today’s problem. The service was only down for a couple of hours, but within an hour Twitter went crazy with complaints, and as I said, the news spread across media outlets like wild fire. We don’t need to trash the idea of cloud computing over these incidents, but even small instabilities like this over such a broad scope does suggest scrutiny. Email, like many other cloud applications, is relatively inert as far as application sensitivity to crashes are concerned. As for other more “sensative” ones, seeing Google crash anything should at least raise eyebrows for lesser capable companies.

The good news is that Gmail is back up, and Google apologized for the incident. No doubt the fervor will settle down in a day or two, and everything “G” will be back to normal. But the larger picture for cloud computing (particularly the e-commerce and enterprise types) took another hit. How big of a hit remains to be seen of course, but for “the cloud” to reach its potential as a multi-billion dollar industry (and particularly a long term one), redundancy and other safeguards will have to be addressed more fully. The rationale might be expressed like this; “If Google can go down for even an hour, then what might happen to much smaller enterprise cloud solutions?” Does today’s mishap darken the outlook for cloud computing? It represents nothing more than a soft Winter rain, but it does remind us that it can rain.

We teamed up with SiteGround
To bring you the latest from the web and tried-and-true hosting, recommended for designers and developers. SitePoint Readers Get Up To 65% OFF Now
  • Mail is so spotty on a shared server I have come to easily tolerate this sort of thing – the fact that Gmail is free also makes it easier to handle – at the same time people are so reliant on email, and on Gmail and without the sky always seems to fall.

    Drink some tea and read a book for a bit perhaps…

    • LOL Brandaggio! I agree. It did not actually bother me that much except I had several deadlines and an important IM going. Not a big deal, just inconvenient. Thanks for the snicker and the insight too. Does make on wonder what else can have glitch.


  • flashparry

    Because I’d recently enabled the labs offline gmail feature, I still had access to old emails and was able to type a few replies that queued up waiting for the things to be fixed.

    I think a cloud-based app that continually syncs to a local copy of the data, and remains usable (if limited) when offline is still an attractive proposition.

    • You are so right Flash. there still exists problems that I won’t go into, but all I am ever suggesting is that more safeguards need to be in place.


  • curtismchale

    downtime happens there is nothing that can be done about it. Downtime happens on corporate email all the time (speaking from experience). at least gmail can be used offline as well. our coroporate is shut down if the mail server goes offline. drafts don’t send later… There is no fool proof way to make sure 100% up time. Any claim to the contrary is a lie.

  • flashparry pretty much summed up my thoughts on cloud computing. I don’t think we will ever have a true, ideally-implemented cloud computing system because of things like this. However, a mostly cloud computing system which has some functionality, even if it is only limited will probably thrive.

    Take for example some sort of document creating and sharing system. Now, when everything is fully functional you can create new documents, edit them, and share them through the cloud without having any real limitations.

    Now, the cloud goes down. If you had some offline portion (such as have the editor and saved copies of local documents, perhaps running in an AIR app). Even without the net (on planes, etc.), or with server outages, you can still look at documents you’ve already been given and create and edit new documents which you can share once you can access the cloud again.

    I think this provides a nice, real-world-practical version of cloud computing that could be very applicable.

    On the same note though, there is always the chance your computer is inaccessible for one reason or another. If you compute in the cloud this isn’t an big problem. You would just grab a notebook or go to a library and keep on going.

    There are trade-offs either way, which is why I like the idea of a balanced middle ground that never completely fails one way or another, even if it becomes limited.

    • Good points all. This is where we are, in the middle. Email is one thing. corporate mail goes down all the time as Curtis illustrates. For some companies however, a deep inventory or CMS aspect going down on a scale like Gmail, might just be catastrophic. I can think of 100 scenarios. 100 million people depending on email in itself (or morning coffee for that matter9, is significant just because of the numbers and the law of averages. What do you want to bet some guy was communicating today from deep in the amazon and got cut off from his only source of outreach? Or, perhaps some other goofy scenario.


  • israelisassi

    It works amazingly well for a free service…

    • That is does Israelisassi, that it does.


  • Jhorra

    I like how Gmail suddenly became cloud computing at some point. Isn’t it just regular old webmail like Hotmail and Yahoo? How is it cloud computing?

    • Hi Jhorra, Thanks for the comment. For the definition of “Cloud Computing” I thought I would look at Wikipedia.

      Cloud computing refers to the use of Internet (“cloud”) based computer technology for a variety of services.[1] It is a style of computing in which dynamically scalable and often virtualised resources are provided as a service over the Internet[2][3][4][5]. Users need not have knowledge of, expertise in, or control over the technology infrastructure “in the cloud” that supports them[6].

      So, just about every service on the Internet is in the cloud.


  • thebillionaire

    Jhorra – its probably hosted on a cloud architecture. Cloud is just starting out why does everyone have to piss their pants when every their is a small outage, their still working out the kinks.

  • thebillionaire


  • Anonymously

    How many hours has your personal computer been down in sum this year? Add up the 5-mins it takes to turn on, during reboot, a crash, a disk failure, virus, etc.

    And I beat you’ve got a huge downtime issue during core hours of use.

    Gmail was down while I was sleep, I could care less.

    • Anony, The point is well taken, but 100 million people do not use my PC.


  • randywehrs

    I agree that the downtime isn’t a sign of anything. Email isn’t like the postal service, its prone to so many errors. It’s like when the power goes out; it’s rare, but when it happens it should make us more appreciative and grateful that we have these conveniences. Even if the cloud crashes all the time, I’d still use it if it worked some of the time, and I think everyone else would too.

  • Sciamachy

    So far it’s proved way more reliable than Outlook plus Exchange. The only difference has been that it hasn’t been a company who is contractually obligated to provide the service taking care of it. That said it’s still in Google’s interest to fix these things quickly because of their advertising business model – downtime means less clicks which in many cases means less money for Google.

  • Anyone who would give up running their own mail server and hand over control to a third party (and advertising company no less) deserves what they get.

    This is how your company communicates! Take back your mail server!

    • The point exactly Steve. This is just email, what about other proprietary and more crucial data hosted on servers not supported by an Internet monopoly?


  • Anonymous

    Those who complain about cloud services seem to be blessed with perfect in-house mail servers, which of course has yet to be invented.

  • Neil Bradley

    It was minimal disruption for myself, and had gmail open when it went down. It didn’t last too long for me.

    I heard that for people using Gmail professionally they promise a 99.9% uptime. What does this mean? As I didn’t realise you there was a capacity for Google to offer this pro? Do you have to pay for it? And if you do, have they broken their promise?

    • Neil, There is in fact a Google professional variant. You can check it out here Here are the selected advanced features –, and it looks like it costs $50 per user – per year. You can try it for free in a 30 day trial also. Evidently the 99.9 is guaranteed, but I think this may involve simply a rebate etc. I expect when the servers crash and burn, they take the Pro variant with them. I hope this helped, but I guess one would have to do the math on the 99.9 for the recent event.


  • Thomas C

    Twitter is exponentiating every fail into a horrible disaster.

    In my company we have more power outages than gmail drop outs.

  • Failing isn’t always a bad thing. Hopefully the error reports generated will give cloud developers enough information to fix this problem from happening in future cases.

    As a side note, it is slightly amusing that the all-powerful google-machine failed.. if only for a short while.

  • Ryan

    Recently I see a lot of negative posts on sitepoint against could computing. Could computing is fine. The problem is the huge user base of gmail.

    If a problem like this occurs to a smaller company that offers SaaS services, it won’t have such a big impact.

    If you should always have access to your emails, backup your emails or go offline with google gears and stop complaining :)

    • Hello, Thomas, Joe and Ryan,

      i hope everyone understands that this post was not meant to be negative. pointing out failures, as Joe points out, is part of the process of refinement (hopefully). The message of this post is that Cloud computing has a ways to go before it is universally acceptable. It may never be universally acceptable for some tasks any way. This is not a bad thing, it does not impact developers other than some particulars of when where and how they build things. Just throwing up our hands and yelling “Yeeeehaaaa!” because someone else likes a philosophy is not exactly the best thing to do when all our livelyhoods may depend on it.

      Computing in the cloud is actually little different in most respects to simply using the Web aas a conduit or communication device. It is still like a copper or more appropriately fiber optic link between us all. The costs inherent for businesses can be much lower, but if you tjhink about it, softwarer supported correctly, and in modeern machines, can do the job much faster. At the end of the day, there has to be someone on the other end of these machines every time. There is no getting around that fact until we create Hal 9000 ( 2001 and 2010), and then, there are another set of problems to consider.

      I use Gmail as my primary email. I was not mad because they went POOF the other day, the other 100 million people are the ones I thought would like some clarity and thought on the matter. As Ryan noted, there are several posts lately on cloud computing, they are not here because I chose to select them, but because they all popped up as news. Think about that for a second. Everything from Spam to security and outage issues coming up one-day-after-another as news. Just some thoughts guys.


  • Sueblimely

    If Gmail starts to crash regularly I would be concerned but not over a one off incidence. I am more used to my ISP email and the internet being down. We do put a lot of trust in Google and expect them to perform perfectly. When something like this occurs we are surprised, because it is unusual.

    If there is a problem with their cloud computing and this starts happening regularly then it is time to move on to another email application.

    Twitter does tend to stir up strong reactions be it Google crashing or a delay in the morning train.

  • sharmendra

    If everyone is buzzing over the Cloud computing just because of one incident, then its really not appreciable. Gmail was not working for few Hours , it doesn’t mean that people shouldn’t think about the cloud.

    • Thanks for your comments Sue and Sharmendra, your points are well taken. There is a much larger picture to paint here. I will attempt to do a post about this. For now, consider the problem as being one of 100 million people being a target with regard to attacks or even down time. The problem is the massive nature of Gmail, and secondarily where it resides. We are talking about collective data versus shared from singular sources.


  • David Mulder

    Couldn’t read through all comments, but it would be more appropriate to say how extremely little any of the big web-apps are offline essentially. Aside from that I am a big fan of cloud computing where the data can be stored by a fourth party. Say google would host the web interface, but a company could set their own servers up where the mails would be stored. This would spare google the storage cost (even though 90% would still have it hosted at google) and would give users the option to quite easily keep there data more private.