Bill Hewlett and Dave Packard worked their magic in a garage. Apple’s Macintosh development team flew a pirate flag over the Bandley 3 building. Now Sun Microsystems hopes a young team that toiled in an unmarked — and reportedly unkempt —  San Francisco loft can spark a turnaround in a tough economy, and build a new billion-dollar business.

The pizza-fueled group, led by engineering whizzes Mike Shapiro, Bryan Cantrill and Jeff Bonwick, spent about three years developing a new type of data storage box that uses flash technology, off-the-shelf parts and open-source software to help companies store and manage information more effectively. Don’t doze off, now — data storage may sound boring, but it’s also lucrative. With the explosion in online video, social networking, Web-based software and online commerce, it’s about as close to a recession-proof business as you can find.

Just think: Every stock transaction, YouTube video, and Facebook friend request gets stored somewhere. That’s why market research firm IDC has network storage pegged as a $4 billion business in the second quarter alone, up 22% over a year before. Tech heavyweights like Hewlett-Packard and Dell have recently bought storage outfits to gain a competitive edge.

For Sun, a respected Silicon Valley data center supplier that never completely recovered from the dot-com bust, this growing pot of money is too big an opportunity to pass up. Investment firm Southeastern Asset Management has recently accumulated a fifth of Sun’s shares, and has hinted that it might start hunting for suitors whether Sun’s board likes it or not. A hit product could buy Sun some time — or even convince investors that it can do just fine on its own.

Shapiro and Cantrill certainly had the brains to come up with something game-changing. The two college buddies, who had risen through Sun’s ranks to become Distinguished Engineers, believed that by tapping software they have developed over the years, they could build a new breed of networked storage box that’s faster, cheaper, and simpler to operate than the mainstream fare from companies like EMC and Network Appliance . They decided to bring the idea to top executives. “I honestly though we’d get laughed out of the room and just get told to go back and do our jobs,” Shapiro said.

One day in the cafeteria the two cornered Greg Papadopoulos, Sun’s chief technology officer, and made an unlikely pitch for resources to pursue their vision. They continued the conversation in his office — and to their surprise, he said yes on the spot. A project code-named Amber Road was born. “Some of our top engineers came and they said this is profoundly important, the time has come, and we want to go do it,” Papadopoulos recalled. “I’d better have a good reason to say no.”

Actually, he could have come up with several good reasons. Unlike high-fliers like Google and Apple, which have billions in cash and zero debt, Sun doesn’t have money to burn on every engineer’s pet project: Recently the company reported it had $2.6 billion in cash, $1.2 billion in debt, and enough potholes in its balance sheet that it has announced plans to lay off 350 employees in January. (That alone doesn’t sound like much, but Sun has let go of 1,500 employees in recent months.) In the most recent quarter, sales slumped 7% to $3 billion. Sun stock also reflects the harsh reality: it’s trading near $4 per share, where it was in early 1995. That’s down 80% from a year ago.

If that weren’t enough, the Amber Road team asked for its own office space, away from Sun’s big-company atmosphere and product development protocols. Shapiro in particular felt that to design an easy-to-use storage system, the team would need to escape Sun’s culture. “We build for engineers — lots of knobs and dials and gauges — and that’s really anathema for ease of use,” Papadopoulos admitted.

So he and systems group chief John Fowler agreed to the group’s pleas for independence, under a few conditions. The space couldn’t be as nice as Sun’s official offices, and there would be no elaborate decorating budget; the team of about a dozen engineers painted it themselves, and found furniture at flea markets. “They love it, it’s right across the street from the bus terminal,” Papadopoulos said. “You just sort of slide pizzas under the door.”

If it sounds like a startup, that’s the idea. There’s some history of this at Sun; legend has it that James Gosling left the company’s offices to work on the Java programming language in rented space on University Avenue in Palo Alto — not far from where Facebook’s headquarters is today. But the question is whether this infusion of outside-the-box thinking will be enough to make a difference for Sun’s bottom line. “How fast it can grow to be a billion dollars plus, which is what it needs to be, is not clear,” admitted Fowler, the systems group chief. But he’s optimistic; customers who have seen the new storage product in action can’t wait to get their hands on one, he says.

Analysts so far are skeptical. Sun will have to prove that its Sun Storage 7000 appliances are powerful enough to serve the needs of large customers, who tend to be comfortable with companies like EMC whose storage products have longer track records. But the fact that Sun has elegantly built flash-based solid-state storage technology into these boxes should help. IT managers are eyeing solid-state storage as a way to boost performance, and while all of the major storage companies are promising to include it in future systems, Sun is among the first to deliver.

“Can Sun get out there quickly enough and grab market share so they can grow their base before others arrive to the party?” asked Gene Ruth, analyst with Burton Group. “A year from now I suspect we’re going to see all the majors out there with significant solid-state products. I always look at Sun as being a technically strong company. Whether they can execute is the question. I hope they do.”

This may be one of Sun’s last chances to prove it can get something going. For years, rumors have swirled around Silicon Valley that someone will swoop in and buy out the company. Granted, that seems mighty unlikely in this environment –- with scads of plucky startups for sale at bargain-bin prices, why buy a struggling giant? But if the stock stays this cheap, valuing the company at just $3 billion, it’s a safe bet that antsy investors will make a move.