VNUNet: What if Google Mail had been your corporate IT system?

February 25 2009

VNUNet examines yesterday morning's prolonged Google gmail outage, and what it means for customers who are considering Google's "beta" corporate offering:

For consumers, losing access to email for a few hours is at most an annoyance. But for a business that moves critical applications to the cloud, such an outage -- especially when it is so out of its control --- could have dramatic consequences. ...

Peter Thomas, former head of enterprise IT at Chubb Insurance and now an independent consultant, said the way Google responded to the crash was also a worry.

"I thought Google had massive redundancy in its server farms -- it appears not -- and it wants to provide corporate mail?" he said.

"Maybe posting information other than 'try again in 30 seconds' would have been a good idea."

Freeform Dynamics analyst Jon Collins said the Google Mail failure shows that people should not simply assume that cloud computing is the way to go.
The cloud or a hosted approach can certainly be the right way to go, with the right product and right architecture.

Link: VNUNet: What if Google Mail had been your corporate IT system? >

Ed Brill | Highland Park, IL USA Category:  Messaging  Add/View Comments [21]

Post a Comment

  1. 1  Bruce Elgort http://elguji.com | 2/25/2009 9:25:51 AM

    Check out what Salesforce offers its cloud customers:

    { Link }

    Google needs something like this.

  1. 2  Scott Marchione http://scottmarchione.blogspot.com | 2/25/2009 9:57:02 AM

    I keep hearing about all these outages from GMail, but honestly I've never noticed. My usual morning routine is to come in to the office, fire up Notes, fire up Firefox (lots of fire around here) and log in to my Gmail, and I let that puppy go all day long. I'm not saying that it's a good corporate messaging solution, I'm just saying that I've never noticed an outage.

  1. 3  Ed Brill http://www.edbrill.com | 2/25/2009 10:01:37 AM

    The outage yesterday was around 5 AM to 5:30 AM EST... so you might have missed it, but folks in Europe certainly didn't.

  1. 4  Mikkel Heisterberg http://lekkimworld.com | 2/25/2009 10:16:46 AM

    Yeah I noticed it big time being in GMT+1 (EST + 6 hours)

  1. 5  Jim Casale http://www.jimcasale.net | 2/25/2009 10:17:18 AM

    for what it's worth...

    They are doing a trial of Gmail for some users where I work. They are supposedly not important users so they don't mind that the services will not be the same. I do wonder who will answer questions about issues users have. Will it be the company I work for or Gmail?

    I still have not been able to setup my school Gmail account to send through POP3. Again. who do I call? The instructions the school provides sends me to a Gmail page that has information on how to setup Gmail - but not for the school's Gmail. The real kicker is the school Gmail is to be used for official purposes meaning if I am submitting work to be graded - what happens when Gmail is down? Will they say too bad you didn't submit it on time?

  1. 6  Chris Hamoen http://edge4.blogspot.com | 2/25/2009 10:28:54 AM

    A commenter above mentioned salesforce.com. A couple years back, they had some serious outage issues, but overall they were still seeing 98%+ uptime. What is interesting is that most of their on-premise competitors don't see this amount of uptime - because of poor internal IT practices.

    I have seen many customers running on internal messaging systems go down for days...even weeks! If they are a traveling user - the dirty secret is that most auto-forward their work email to a gmail or hotmail account....because of the series of fail points in connecting to the internal network.

    I'm not making a pitch for cloud computing, but the assumption that on-premise solutions have 100% uptime is incorrect.

  1. 7  Ed Brill http://www.edbrill.com | 2/25/2009 10:55:29 AM

    @6 assumption? I definitely have customers who report 100% availability for their Notes/Domino environments, mail and applications. Yes they have outages but they have architected for reliability and have clustered servers take over etc.

  1. 8  Mike Johnston http://www.google.com | 2/25/2009 12:35:46 PM

    It only needs to happen once. You can rest assured Gmail won't be going down again after the publicity had yesterday.

    Despite the unexpected outage. Imagine the benefits of a corporate messaging system in the cloud.

    Not having to manage the servers, drive space, viruses, spam, security patching, etc. Not a single worry about lifecycling servers and upgrading. Cost savings by a reduction in your messaging infrastructure staff.

    I think the Pro's really outweigh the cons in this space. The writing is on the wall. Most home users are already using cloud mail scapes, It's about time corporate America joined the party.

  1. 9  Ed Brill http://www.edbrill.com | 2/25/2009 12:55:28 PM

    Wow, troll alert! Can't find any reference to a Mike Johnston at Google, but this guy from Ohio should know better. After all, it's not like this was the first such outage -- there was a big one six months ago, too. { Link }

    Most of the discussion we have on this topic is about how the consumer-grade e-mail doesn't necessary jump the shark to "corporate America" (his words not mine, I would never have limited scope to just the US)

    One admission that is in "Mike's" comment -- "reduction in your messaging infrastructure staff". This is important, because a lot of companies talk about cloud/SaaS as if their internal costs will drop to zero. Hardly the case.

  1. 10  Jim Casale http://www.jimcasale.net | 2/25/2009 2:08:51 PM

    @Ed - "This is important, because a lot of companies talk about cloud/SaaS as if their internal costs will drop to zero. Hardly the case."

    You must know something management here doesn't know

  1. 11  keith Brooks http://www.vanessabrooks.com | 2/25/2009 2:15:31 PM

    @8, sorry mike but your admin costs don't go away, Google is not going to add/drop your users for you or change passwords.

    Then again they aren't going to police your security or usage either, unless you pay them to find out your employees are emailing illegal items or not corporate emails.

  1. 12  Ben Poole http://benpoole.com | 2/25/2009 2:57:58 PM

    @9 how is that post "trolling"?

    OK, so Mike takes a contrary position to yours, but he's not being offensive about it, and that's what healthy debate is about, right?

  1. 13  Ed Brill http://www.edbrill.com | 2/25/2009 3:11:39 PM

    @12 he used a fake name and, despite using the Google URL in his comment, there's nobody at Google by that name that I can find. I don't mind the partisan comment, it's easy to shoot down, but at least have the courage to use one's real name.

  1. 14  Chris Hamoen http://edge4.blogspot.com | 2/25/2009 4:22:16 PM

    Hey Ed - my only point really is that whenever an On-Demand app goes down, there is a lot of news made re: uptime. However, many on-premise solutions (messaging, apps, etc) have much less than 100% uptime - nothing to do with the application, just with internal IT.

    I do agree that an app like Lotus Domino should NEVER go down - if it does, IT failed. We have a small customer (under 30 users) that blackboxed an old linux server with Domino - they don't even know what it does and have never touched it....and it's been up and running for 4 years. Then we have other customers who probably have 80% uptime because, simply put, of IT issues that have nothing to do with technology. (Hence the allure of On-Demand - business users can skip over the IT people who they don't understand)

  1. 15  Mike Brown  | 2/25/2009 5:29:16 PM

    A previous company of mine had a 100% business hours up-time record for mail and apps servers during the six years that I was there. That's 100%. Not "close to". Not "99.9999 recurring". 100%.

    Sure, servers crashed at times; they were running on Windows, after all! But clustering made those crashes all but invisible to the users.

    In the early days, we would get complaints when users' unread marks changed. That was, in fact, the only way that they knew their home mail server had gone down at all. After that problem was fixed in ND6.5, barely a murmur was heard.

    It's interesting how all of the wise-after-the-event IT press articles are basically saying "this sort of thing has to be expected". Reminds of the old E.M. Forster short story, "The Machine Stops". (When the machine starts going wrong, everybody minimises then ignores its problems, because they're too scared to admit just how dependent upon it they've become).

  1. 16  Richard Schwartz http://www.poweroftheschwartz.com | 2/26/2009 12:40:05 AM

    @6: 98% uptime? Let's do some math. There are ~250 work days per year. 2% of that is 5. So 98% uptime implies that you lose 5 work days of productivity per year. Doesn't sound all that good to me.

  1. 17  Ben Poole http://benpoole.com | 2/26/2009 1:00:27 AM

    "So 98% uptime implies that you lose 5 work days of productivity per year"

    … if your productivity relies on email. And this is where I part with received logic; I just can't get that excited about an email outage when there are so many other ways to work, communicate, and be productive!

    Isn't collaborative software like Lotus Notes partly about getting people *out* of the inbox?

  1. 18  Mike McP http://www.openntf.com/mPortal | 2/26/2009 6:54:20 AM

    @17. I'm not sure how we work, as tech geeks, compared to how businesses function. Where I work, if email goes down, we grind to a halt. People can't pass around files, meeting reminders don't happen, people can't book meetings, etc. Granted, all of this is avoidable if you train folks not to use email for that stuff, but it not an easy culture to change.

    Luckily, we use Notes in a cluster, so it just doesn't happen! I'm not even sure how you'd do some of the basic productivity stuff in the cloud. How do you book a meeting? What sends your reminder and alerts if you're not always connected? How do you prevent confidential email from leaking out of the environment? What stops someone at home from accessing their work Gmail account unknowingly with a malware/sniffer installed? I think folks who jump on this bandwagon early will eventually realize that it was not a wise move. Of course, bean-counters aren't always the sharpest tools in the shed when it comes to critical analysis and long-term risk.

    Not even mentioning apps, I think remotely hosted email has a way to go to do all that Notes does. I'm sure it will handle the basic Outlook Express or Exchange customers. That doesn't mean I don't see this remote sector growing...eventually it will mature.

  1. 19  Ben Poole http://benpoole.com | 2/26/2009 7:12:16 AM

    @18 meetings != productivity ;o)

  1. 20  Chris Hamoen http://edge4.blogspot.com | 2/26/2009 8:56:59 AM

    @16 You are correct - my point however is that many IT groups can't keep on-premise solutions up more than that.

    @17 If email goes down for non-tech geeks (ie manufacturing/industrial), the world stops. No sending quotes, no receiving POs, etc.

  1. 21  Ferdy Christant http://www.ferdychristant.com | 2/26/2009 12:03:15 PM

    "But for a business that moves critical applications to the cloud, such an outage -- especially when it is so out of its control --- could have dramatic consequences. ..."

    I do not see how an email outage of a few hours have dramatic consequences in most businesses. In fact, I dare to question whether the reliability of the average internal email system at "corporate america" is that much better. Google has been down, what...twice? In how many years? This does not happen in business environments?

    Sure, I agree it can be an annoyance when email is down. Sometimes even painful. Dramatic? I think the few Google mistakes are blown out of proportion.

    By the way, this other cloud service, Amazon S3, was down for multiple days once. Pretty bad for a storage server. The world didn't end though, and I don't many people went out of business. Nothing dramatic about it, just sheer bad luck, incompetence, or any other factor.