Just because IT hasn’t gotten the memo, doesn’t mean they should kid themselves into thinking cloud computing hasn’t arrived. Two simultaneous events – the launch of substantive online coverage of the Olympics and the massive Gmail outage – signal the coming of age of on-demand.
First the one that’s working, making money, and replacing a prior generation of technology. No, not the Olympics coverage. The numbers are climbing – 4 million on Saturday and a whole new demographic starting with the beginning of the workweek. But the lion’s share of revenue continues to be made by NBC here and other big media companies overseas the old fashioned way – TV.
No, the real replacement strategy is the death of Office and the rerouting of email from corporate to global servers. Gmail’s first substantial outage as a mature service produced a firestorm of panic on the next generation communications hub: Twitter. Just as the Blackberry survived the early hours of 9-11, Twitter and its family of social media nets easily managed the outage by rapidly signalling the scope and speed of the event.
For the first time, companies who rely on Gmail and Gchat found 2 of the 3 channels they rely on for day-to-day operations shut down on a system-wide basis. Like a virtual earthquake, you could watch the spread across the globe, and then track the aftershocks as they continued to ripple through the system for more than an hour. A rare profile of small businesses with workers spread across local borders and timezones emerged first in the new media, where Gmail is widely used in much the same way that Macs dominate tech conferences.
To those of us who’ve grown comfortable with Gmail’s relative reliability and wide penetration of the burgeoning iPhone culture, the outage was initially written off as something temporary that would resolve with enough clicks of Refresh. But the familiar signature of a partial outage in one area and no sign of trouble in another was clearly not present. Twitter immediately contradicted that notion, and raised some interesting next questions.
First, IM was down, and for those of us who naturally gravitate toward single sign on and aggregation of services, that meant AIM and the Jabber XMPP gateways to Twitter, Identi.ca, and the IMfeeds mashup of Techmeme Firehose and Dave Winer’s NewsJunk services. I have a redundant IM service in Skype, so the next thought was to realize that I had no email replacement immediately available. I’ve never been a fan of Yahoo Mail or Hotmail, though I believe I have accounts on both as a result of signing up for services that persistently bait you into establishing email. The only way I was able to launch my Live Mesh preview account was with a Live mail account (which may already be Hotmail for all I know.)
Next I sent a Twitter direct message to a friend at Google, which probably wasn’t a very good test of its availability since said friend was probably running around trying to fix things along with everybody else there. A call to his cell phone rang for 10 or more rings without an answer or voicemail. At approximately the same time Twhirl started spewing some interesting messages including this from Scobleizer: “Just because Gmail is down doesn’t mean you should DM me. I hate DM’s. Hate. Hate. Hate. Is that clear enough? :-)” Others had the same idea, and now I knew from Robert that the DM pipe was open.
Net: I had IM-level service via Twitter, XMPP realtime via Twhirl’s Identi.ca feed, and no easy way to get to email. I forward most of my email from my home domain to Gmail, so I thought briefly about accessing that Web interface, only to remember that the password and log-on details are stored – yes, in Gmail. I had thought some weeks earlier about signing up for MobileMe to test push iPhone mail, which I would have enabled by forwarding my Gmail feed; that may not have worked if Gmail stopped forwarding during the outage. Besides, MobileMe was also down today.
What I really needed was a redundant mirror service, perhaps something that was receiving the same type of forward from my home domain. Or perhaps something like MeshMail, which would synchronize incoming messages from all relevant services: email, Twitter, FriendFeed, Facebook Mail/CHat, AIM, MobileMe, Skype, etc. That service could even route messages across whatever individual services messages came from, allowing me to build and maintain a message cloud according to the strengths of each of these services. Of course, then Google would need to build a comparable service to create redundancy with MeshMail.
MeshMail may not be available today, but the speed with which Silverlight went from a flashy demo at MIXO8 in March to millions of deployed seats on Friday’s opening day ceremonies suggests at the very least the possibility of an equally rapid roll out of just such a flashy MeshMail demo at the PDC in October. On the Gillmor Gang today, Dan Farber said “The network has a remarkable lack of intelligence” and Robert Scoble said “I go where my friends are.” Between those two points lies the MeshMail opportunity – an intelligent message router that gives IT a scalable solution to the problem of cloud failover.