GigaOM

On the eve of Structure 08 in San Francisco, the folks behind many of today’s cloud computing initiatives will be gathering at CloudCamp. GigaOM is one of the evening’s sponsors:

CloudCamp was formed in order to provide a common ground for the introduction and advancement of cloud computing. Through a series of local CloudCamp events, attendees can exchange ideas, knowledge and information in a creative and supporting environment, advancing the current state of cloud computing and related technologies.

If you’re in the city, come check it out. The event, which will be held from 5-9 pm on June 24th, is community-backed and free to attend. And it will be a great primer on cloud computing.

Bribery, extortion and other con games have found new life online. Today, botnets threaten to take vendors down; scammers seduce the unsuspecting on dating sites; and new viruses encrypt your hard drive’s contents, then demand money in return for the keys.

Startups, unable to bear the brunt of criminal activity, might look to the clouds for salvation: After all, big cloud computing providers have the capacity and infrastructure to survive an attack. But the clouds need to step it up; otherwise, their single points of failure simply provide more appealing targets for the bad guys, letting them take out hundreds of sites at once.

Last Friday, Amazon’s U.S. site went off the air, and later some of its other properties were unavailable. Lots of folks who wouldn’t let me quote them, but should know, said that this was a denial-of-service attack aimed at the company’s load-balancing infrastructure. Amazon is designed to weather huge amounts of traffic, but it was no match for the onslaught.

When it comes to online crime, the hackers have the advantage. A simple Flash vulnerability nets them thousands of additional zombies, meaning attacks can come from anywhere. During Amazon’s attack, legitimate visitors were greeted with a message saying they were abusing Amazon’s terms of service, which could mean that those visitors were either using PCs that were part of the attack, or were on the same networks as infected attackers. The botnets are widespread, and you can’t block them without blocking your customers as well.

Other rackets give the attacker an unfair edge, too: It takes an army of machines to crack the 1024-bit encryption on a ransom virus, but only one developer to write it.

A brand like Amazon can weather a storm, because people will return once the storm has passed. But just look at the Twitter exodus to see how downtime from high traffic loads can tarnish a fledgling brand. Slideshare survived such an attack in April, and while many other sites admit to being threatened, they won’t go on the record as saying so.

Up-and-coming web sites are often great targets, as they often lack the firewalls, load-balancers and other infrastructure needed to fight back. And it’s not just criminals: In some cases, the attacker is a competitor; in others, it’s someone who just doesn’t like what you’re doing.

Fighting off hackers is expensive. Auren Hoffman calls this the Black Hat Tax, and points out that many top-tier Internet companies spend a quarter of their resources on security. No brick-and-mortar company devotes this much attention to battling fraud.

Wanting to survive an attack is yet another reason for startups to deploy atop cloud computing offerings from the likes of Amazon, Google, Joyent, XCalibre, Bungee, Enki and Heroku. But consolidation of the entire Internet onto only a few clouds may be its Achilles’ heel: Take down the cloud, and you take down all its sites. That’s one reason carriers like AT&T and CDNs like Akamai are betting that a distributed cloud will win out in the end.

Cloud operators need to find economies of scale in their security models that rival the efficiencies of hackers. Call it building a moat for the villagers to protect them from the barbarians at the gate. Otherwise, this will remain a one-sided battle that just gives hackers more appealing targets.

Could Dell or HP offer the next iPhone? Nvidia Co-founder and CEO Jen-Hsun Huang certainly isn’t ruling them out. In fact, his firm has launched a low-power computing platform called Tegra that’s specifically aimed at bringing more competitors to the mobile handset market.

But that’s long-term thinking; mobile currently accounts for less than 10 percent of Nvidia’s business. In the meantime, the $4.1 billion-a-year graphics chip maker is battling Intel to bring more focus and computing jobs to the graphics processor that in the past may have been handled by a computer’s main processor, also known as CPU.

Indeed, media-enamored consumers have pushed Nvidia’s sales up by 34 percent from 2007 through 2008 (Nvidia’s fiscal year ends in January). The Tegra platform, based on the company’s sexy new application processor, is one-tenth the size of Intel’s rival attempts to make smaller mobile computers — and runs at less than one-tenth of the power.

GigaOM: On the GPU side, how will you compete against Intel and AMD, who both have platform strategies for the PC market?

Huang: It’s the same way we compete selling graphics cards. Our GeForce chipsets sales are up 50 percent even though the overall PC market is up only 7 percent. So people who care about visual interfaces are buying our chips. We see ourselves as CPU-agnostic. We look at the vertical market we want to go into and we let the market decide which CPU it wants and then we partner with that CPU provider.

GigaOM: Intel is a big partner for you but you recently threatened to “open a can of whoop-ass” on them at an analyst meeting. What is the competition between the two companies like?

Huang: I think the “open a can of whoop-ass” response was really to dispel myths that Intel was out telling everybody. They’ve told people that GPUs are dead and that integrated graphics have taken over the world. I think that’s just really bad sportsmanship, frankly.

Looking toward the future, Tegra is a really fabulous computer and will increase in performance two to three times every year. And if Intel and AMD don’t continue to make the desktop and laptop PCs more and more magical every year, before you know it, mobile computing devices will be disruptive to the PC the same way PCs were disruptive to the mainframe industry.

GigaOM: So tell us about Tegra and mobile computing.

Huang: Five years ago I saw the convergence of a couple of technologies — particularly wireless technology and rich LCD displays that were eventually going to bring to consumers mobile computing devices. It will have elements of entertainment, elements of communications and elements of computing.

You have to deliver these elements with almost no power. If you boil it down to where the CPU, the GPU, and all the individual processors dissipate almost no energy so you could wind it up like a wristwatch or recharge it with the temperatures of your skin, you could make a mobile computing device that fits in your pocket. So we started with a blank sheet of paper and five years later we have Tegra.

GigaOM: What will these devices look like?

The iPhone is the world’s first legitimate mobile Internet device. There are different design decisions that can be made for the iPhone and devices like the iPhone. Some will have Wi-Fi, some touch, and some will have a slide-out keyboard, but that speaks to the orientation where the suppliers want to point their device.

GigaOM: How important will processors be for this type of device?

Huang: Inside the iPhone is a custom chip designed by Apple. Apple has a really great computer chip in there. It has a good graphics core, actually. For the rest of the computer industry who don’t have an internal chip design organization, they’re going to have to rely on someone else to do it. But that is a multi-, multiyear project. So the notion that the rest of the computer industry can quickly catch up with the iPhone really, really depends on someone else designing a chip that’s really low power but also leapfrogs the iPhone. That’s where we come in.

GigaOM: What about competition from Intel, Texas Instruments, Qualcomm, VIA and others also going after this mobile computing market?

Huang: I see the mobile computing space bifurcated in two basic categories where most of the suppliers will end up. There are people really good at communications and that’s where the baseband providers are. TI is really fabulous at communications and Qualcomm is really fabulous at communications.

What Nvidia and Intel are really good at is computing. The difference between us and Intel is we decided to start with a clean sheet of paper. The PC legacy is what causes the laptop to be so big. So we abandoned the PC legacy in favor of low power while retaining all the computing expertise.

GigaOM: So who are the end vendors for Tegra?

Huang: Anyone who wants an alternative like the iPhone.

GigaOm: Handset makers?

Huang: No. Every PC company in the world. I don’t think it’s unrealistic for Dell to offer a mobile computing device.

GigaOM: With a baseband processor?

Huang: Sure. And no, that’s not a product announcement for Dell. Handset and PC companies are becoming similar. So will Motorola be more successful at building a mobile computing device or will Dell be more successful at building a mobile computing device? It’s kind of hard to say.

From the company that spent $4.1 billion buying a tape company comes some cutting-edge storage news: Sun Microsystems said today that it will put solid-state Flash drives into a line of servers and other storage products, making access to stored data faster and more energy efficient. EMC made a similar announcement earlier this year.

The big vendors aren’t alone in their focus on speed. We’ve covered startups in the past whose entire existence is based on figuring out how to get to existing data faster, either through appliances or compression. With users storing more data and expecting continual access to that data, storage is no longer just about cramming as many bits and bytes in an archive as possible; it’s also about getting to them faster.

Flash, however, is a rather expensive way of solving the problem. Prices should drop as larger solid-state drives using Flash begin appearing in more consumer devices such as laptops, and the use of SSDs in the server world will only help prices fall, but it won’t be mainstream in the data center within the next year or two. Even as Flash gains ground, it will still be just one aspect of storage, for years — if not decades — to come.

While it may make sense for some companies to buy servers integrated with Flash, existing technologies will continue to flourish. Even tape is still in use today.

Amazon.com founder and CEO Jeff Bezos was one of the attendees at this week’s D6 Conference in Carlsbad, Calif., to be interviewed on stage, where he talked about Kindle at length. But right after his chat with Walt Mossberg and Kara Swisher, I caught up with him to discuss Amazon Web Services and his company’s efforts in cloud computing. Here is a short excerpt from that conversation, captured on my Sanyo Xacti. In particular, he talks about… Continue Reading