August 2008 Archives

Cloud computing has been getting an increasing amount of press recently, because it offers attractive solutions to a number of business problems. Cloud computing translates to "hosted applications;" a hosting company lets you build and deploy applications on its servers, delivers those applications through users' browsers, and generally promises near-24x7 uptime and unlimited scalability. To businesses, that translates to hard dollars: servers the company won't have to buy and maintain, databases for which the company won't have to buy licenses, free storage, backup, and disaster-recovery services—all of which mean IT personnel that the company doesn't have to hire and pay salaries and health care for.

Most hosting companies offer infrastructure for development as well. Common types of applications can be developed more quickly, deployed more easily, and scale immediately. That's an attractive proposition; however, there are downsides.

The first problem? Despite the promises availability  has been less than stellar. Many of the major hosting companies have had outages recently, including Google's Gmail, Amazon's S3, and salesforce.com. Web site outages aren't usually terribly important to businesses, because most sites don't host applications that the business depends on. But when businesses begin to host mission-critical applications in the cloud, such as email, customer/sales data, and financial services, they need and expect near-perfect service levels.

Second, because hosting companies own the hardware, it's unclear how difficult or expensive it might be to retrieve the data generated by all these cloud-computing applications. But in today's world, these applications and data are essential to the business. What happens if a service outage can't be fixed quickly? What happens if the hosting company goes out of business altogether? What recourse does a business have if a hosting company fails to back up their data and loses it? Hosting companies essentially hold businesses hostage to their own data.

Third, so far, each system is unique—there's little to no ability to move applications from one hosting company to another without rewriting and retesting the application. Sure, businesses could write applications that copy the data to internal servers—but that puts them back in the position of maintaining their own servers, losing one of the major advantages that cloud computing offers.

Finally, there's a question about what the hosting companies could do with the data that gets collected on their servers. Obviously, they have access to that data, and obviously, they must be allowed to have some level of interaction with it: encrypting and decrypting it, backing it up, copying it to improve throughput, moving it between servers, etc. This raises questions about exactly what rights the hosting company has. If a hosting company employee looks at your data, what's the penalty? What if a hosting company decides to sell one company's data to a competing company? What if they outsource the data-hosting portion of the business? What if the hosting company decides to mine the data for email addresses and user interests? Whose data is it, anyhow?

Unfortunately, the real-life answers to these questions have been disappointing. So far, court decisions about who "owns" stored electronic data have trended firmly in favor of the owner of the hardware on which that data is stored, or the business that generated the data, regardless of who generated it. For example, many employees have discovered (to their surprise) that anything they write or any data they generate using a company computer (such as email, documents, or Internet access logs) belongs to their employer, and that the employer is perfectly within its rights to analyze that data or use it however they see fit. One recent UK court case found that an employee who simply uploaded his Outlook contacts list to LinkedIn did so illegally, because—even though the employee insisted he had built that contact list personally—it belonged to the employer.

All in all, it's no surprise that a recent survey of top CIOs by Goldman Sachs & Co. found cloud computing was the last item on their priority lists. One analyst posited that the reason cloud computing placed so low is because "They [hosted services] require a technical understanding to get to their importance. I don't think C-level executives and managers have that understanding." But it's rather obvious from the simple advantages stated at the beginning of this blog that gauging the benefits of cloud computing does not, in fact, require much technical understanding. It's far more likely that executives and managers are avoiding cloud computing because they do understand the implications, and are simply making logical decisions to protect their business data.

At a lunch conference yesterday, Yahoo! released Fire Eagle out of beta. This API allows you to broadcast your location to whomever you want, automatically or manually via your phone. You can post your location on your blog or other various sites. To me, and certain segments of the population, this sounds like a nightmare. To other segments of the population, this sounds way-cool. The latter segment is the same population that thinks email is ancient technology. How secure does today’s youth feel that they want people to know exactly—to the address—of their current location?

At yesterday’s press conference, I saw three different demos of this type of application. The first was from Pownce. The 25-year old co-founder presented Pownce as a means to “send stuff to your friends” because who has time for email and all that spam? Pownce uses Fire Eagle and tells all your friends where you are, in case you want them to join you, without courtesy of an invite or a phone call. Because, you know, phone calls and invitations take so much time what with all the dialing, ringing, and having to leave a voice mail… and then you have to wait for someone to call you back. Who has time for that?

Moveable Type, Six Apart’s flagship product, was the second demo and also featured Fire Eagle. You guessed it, this application allows you to post your current location on your blog. Just in case you want anyone who happens to read your blog to stop by and say hello, or mug you, or something…

The last demo was the most useful looking, to me. Outside.in Radar brings you local blogging news. Based on the concept that you are concerned about the pot hole on your street and not concerned about the pot hole in the next city, Outside.in Radar tells you who is blogging what within 1000 feet of your zip code (you manually provide your zip code). I gave it a test drive and I learned that someone wrote a review of a restaurant that is within walking distance of my home. The site provided more links to more information about this restaurant—and several others in the area—that I don’t think I’ll ever have time to click on them all. It also allows you to track all future news items about this restaurant. It is information overload at it’s finest. Talk about a time suck. I could seriously get lost on this site.

As for me using Fire Eagle, I think I’ll put it on hold. I’m very comfortable without people knowing exactly where I am at all times. Unexpected visits do not sound all that appealing to me. But then again, I think email is still useful, so I am probably not the targeted audience for this API.

Amidst a flurry of announcements about the foundation at last week's LinuxWorld, LiMo's Morgan Gillis held a kind of "state of the foundation" session on Wed., describing the foundation's seemingly realistic goal: to strike a balance between the open source ideals of collaboration and community and proprietary software's insistence on competition and innovation. “The community delivers the platform," Gillis said, "and commercial innovation thrives on top of it."

Founded in 2007, the LiMo Foundation was begun by mobile industry heavy hitters (Motorola, NEC, NTT COCOMO, Orange, Panasonic, Samsung, and Vodafone), to "create an open, Linux-based software platform for use by the whole global industry to produce mobile devices through a balanced and transparent contribution process enabling a rich ecosystem of diffrentiated products, applications, and services from device manufacturers, operators, ISVs, and integrators."

Make no mistake: LiMo is NOT a standards body, but a group of collaborators, developing a free, device-agnostic middleware upon which you can build using almost any development environment you'd like. What if you run into patent trouble? LiMo's members hold over 300,000 patents—and all of them have agreed to mutual patent non-assertion. But if a non-member should try to enforce their patent on your code, LiMo can help—if their subtantial weight can't stave the assertion off altogether.

Perhaps the most interesting facet of LiMo's strategy is its angle on solving fragmentation. Their solution lies in "reciprocation of fixes and optimization"—the user community, consisting of representatives from the entire mobile ecosystem, agree on changes to the OS. This means no forking, and—if you stay within the foundation's framework—no worries about device- or carrier-specific snags to porting your app. The more people that join the foundation, the more widespread the agreement, and consequently, the less fragmentation.

And since the platform is open, there’s nothing to stop LiMo’s users from creating their own market spaces.

Sounds reasonable, right? I guess only time will tell.