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-24×7 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.