For every win that seems to promise more enterprises will at least consider adopting Google Apps, there's a worry that seems just as likely to hold them back.

Last week I wrote about Google's announcement that it had signed a 30,000-seat deal with automotive supplier Valeo, in what it called "one of the largest enterprise deployments of Google Apps to date." Yet also last week, millions of users experienced problems accessing Google's search engine, e-mail and other services in an hour-long outage that almost certainly gave potential Apps customers pause. (I don't know about other users, but I've been having continued, intermittent issues with the Google News service.)

Way back in 2007, ZDNet's Phil Wainewright tapped performance outages and Google's inability to respond to them as one of three "mega-traps" that could hinder adoption of Google Apps and hurt the overall software-as-a-service market. (The other two traps: unsustainable pricing and a lack of real love for a subscription-based model that doesn't rely on the company's acknowledged expertise in advertising.) Google isn't the only company that's suffered outages in the cloud. Other members of the crash club include Amazon's S3 cloud storage service and a beta offering of Microsoft's Azure cloud computing platform.

Google's commitment to subscription-based software sales also still remains a challenge, according to a CIO Today story about Google's challenges in the enterprise. A Google engineering director involved with Apps said it's difficult to engage executive attention, perhaps because Apps is still such a small part of the company's business. Google doesn't reveal revenue numbers from Apps sales, but a Sanford C. Bernstein analyst estimates that Apps will generate $273 million in sales in 2009, 1.3 percent of Google's expected total sales of $21.05 billion.

Google has also failed to adequately address security and compliance issues. As IT Business Edge blogger Lora Bentley wrote last month, folks at the Computer Security Institute have been asking, "How do you prove compliance in the cloud?" Burton Group analyst Guy Creese raised the issue when I interviewed him earlier this year for a story on whether Google Apps were ready for the enterprise.

Compliance/security concerns prompted General Electric, a company that has been using Google Apps for two years, to test AdventNet's Zoho Web-based productivity software. While Zoho's applications aren't as robust as Google Apps, GE can host them in-house on its own servers, GE CTO Greg Simpson told CIO Today. (Zoho also announced a pilot with Swisscom in July.) The CTO called the inability to keep Google Apps data on internal servers "probably our biggest stumbling block to going bigger with Google."

Dave Girouard, president of Google's enterprise business, professes not to be worried about Apps adoption. According to CIO Today, he said the company has another decade or two to grow into the enterprise market. But does it? GE's Simpson said the company will also consider Web-based versions of Microsoft's Office software. In another story I wrote about changes to Google Apps' pricing, Creese told me:

If Google had come out with a full-featured enterprise package, I think they’d be doing massive damage (to Microsoft), but they didn’t do that. In a large corporation, if you fire the starting gun now, you might be able to get a conversion project going in nine months. So if Microsoft can come out with something in mid- to late 2009, a lot of IT people will just be inclined to wait.

Microsoft's Office 14, which willl reportedly feature at least some cloud-based components, is currently slated for late 2009 or early 2010. Based on past Microsoft releases, however, it’s possible that Office 14 may be late or riddled with bugs. And just as folks have questioned Google's commitment to the enterprise, others have questioned Microsoft's commitment to the cloud.

One thing seems fairly certain. There won't be a "one size fits all" (or even most) cloud model. As IT Business Edge blogger Art Cole wrote in March, there will be a mix of public, private and hybrid clouds.