Google's Gmail outage on Monday was the latest stumble for nascent cloud computing services, which are becoming the lifeblood for small businesses and startups. The Gmail outage--along with Amazon's stumbles of late--raises a few key questions: Where's the offline synchronization capability? Can we depend solely on the Web? Is Microsoft's software and services mantra the best path forward?
Those questions don't have easy answers at the moment. But one thing is clear: If you're depending on these Web-based applications you need a backup plan. To wit:
These growing pains, which are more evident each day that we rely more on service-based software efforts, indicate that you can't really trust the cloud at this juncture. It's too early and providers are learning as they go. Phil Wainewright noted last week that the uptime issue really comes down to trust. A customer trades in dependency only if a vendor earns trust. In the meantime, Phil outlines some common sense backup plans:
Don’t trust the cloud at this early stage in its evolution. Use the SMTP capability that Gmail provides to keep a local copy of your email inbox. If that’s not possible, use a third-party Gmail backup service. Use your own domain and host the DNS somewhere else so that you can switch email providers without having to change your email address. Keep a back-up copy of important documents and other vital online assets. Finally, make sure you know what recourse you have when things go wrong.
However, there's another angle to this: These cloud providers should be offering offline synchronization capabilities. In Gmail's case, Google could use its Gears code to work offline. This offline synchronization is the Holy Grail and should be a common feature. Ryan Stewart notes:
Systems are going to go down, it’s a fact of life. What’s important is to be prepared when those systems go down which is a major reason that some kind of offline access should be built into systems like email. In theory we’ll reach a time when the cloud really is always on, but we’re not close and it may never happen. When a company like Google, which has a ton of redundancy built in, or Amazon S3, has issues, we’ve got to have applications and systems that let us continue to work.
Ryan adds that synchronization is a really hard problem. But is it really that difficult? Or is it just not a priority right now? I'd argue the latter.
See all cloud computing posts.Add it up and it's clear that the cloud just isn't ready for primetime. You depend solely on outside cloud services for your apps, but there will be times when that choice will look like a bad one. What's needed is a hybrid approach to act as a bridge to the future.
Is Microsoft on to something?Microsoft's has widely pitched its software and services vision, but to date the effort has been mocked in cloud puritan quarters. It's not true cloud computing argue these cloud computing puritans. Microsoft is just trying to protect its software dominance (duh) they argue. Here's Microsoft's vision:
The future is a combination of local software and Internet services interacting with one another. Software makes services better and services make software better. And by bringing together the best of both worlds, we maximize choice, flexibility and capabilities for our customers. We describe this evolutionary path in our industry as Software + Services.
Last month, Microsoft's chief software architect Ray Ozzie foreshadowed the company's elastic cloud computing effort dubbed Zurich. Mary Jo Foley wrote:
Microsoft is known (at least by some of us) to be building a cloud platform atop which it will allow ISVs to build their business applications. That platform is codenamed “Zurich.” Microsoft has described Project Zurich publicly — to the very limited extent done so — an initiative to “extend Microsoft’s .NET application development technologies to the Internet ‘cloud.’”
“Many software vendors and VARs (value-added resellers) want to move their solutions to the cloud,” Ozzie told FAM attendees. Ozzie also said that Microsoft hoped to build a hosted-developer solution that would appeal not just to commercial vendors, but to open-source companies, as well. Ozzie said Microsoft was developing a solution that would have a “pay-as-you-go” model.
Two things to note here. First, Microsoft is late to the party. However, being late may not be such a bad thing and it's possible the market developments may play right into Microsoft's hand. Rest assured Microsoft's cloud will hook into its software on the desktop and enterprise (that may be the only part that the software giant gets right out of the gate). That desktop synchronization will be key and probably the biggest selling point for Microsoft's brand of the cloud.
And the more cloud outages we have the more customers are going to look at Microsoft's outlook and say "hey that makes sense." In Silicon Valley, they'd like you to believe we'll all be in the clouds. The reality is different. You'll be taking a hybrid approach and the cloud barbarians at Microsoft's gate better come up with better offline synchronization efforts.
Update: There's some good discussion about backing up Gmail and Google Apps. Google has a list of services and third party providers worth a look, but I'm referring to an offline synch that just happens. Many of these services in Google's list focus on one area instead of a bulk offline backup.
Other suggestions via the Talkbacks worth noting:
Also see: Cloud computing: A look at the myths