Seybold's Take: Do you trust the cloud to store your mission-critical data?

Everywhere I turn it seems the world of computing and the Internet, both wired and wireless, is looking to the cloud as the next big gold rush. Cloud storage companies are coming online almost daily. AT&T and Verizon Wireless have cloud plans, and other companies are touting cloud-based applications. The premise of all this is that we will carry lighter, less expensive devices because all of our applications and data will be stored in the cloud. This certainly sounds like a wonderful future for all-but there are a few issues with cloud computing, and when you add wireless broadband into the mix, the problems associated with cloud computing are magnified.

First, the Internet is not a mission-critical network, and neither are any of its on-ramps, wired or wireless. Monitoring traffic on the Internet in the United States indicates that we are currently at about 80 percent of capacity most of the time and we have not yet seen the true impact of streaming video on this traffic load. We have already seen issues with wireless broadband access. IPhone users in major metropolitan areas complained about slow or no access to AT&T's broadband services.

Wireless broadband services are shared broadband services. That is, within each cell sector there is a finite amount of broadband capability available. If you are the only user in that cell sector, you will have all of the available bandwidth. However, if there are a number of customers within the same cell sector, you will be sharing the total data capacity with all of the other users. If all you are doing is checking email or surfing the web, the capacity in a given sector will probably suffice. However, if some of the customers are downloading streaming video or sending and receiving large files to and from the cloud, chances are that data rates will drop dramatically. Add to that the fact that the further you are from the center of a cell site, the less actual data speed and capacity you have, and you begin to see the problem.

It is interesting that both AT&T and Verizon are embracing the cloud as a service they want to offer. Store your data in the cloud, or even your applications, and use a lighter, less capable wireless device because you won't need the storage for your applications and your files. However, using cloud-based applications and files puts an even greater strain on a wireless network. The FCC says that by 2012 we will be 95 MHz of spectrum short for the demand for wireless broadband services, and by 2014 that shortfall will be more than 200 MHz of spectrum. That is why the FCC has committed to "find" an additional 300 MHz of spectrum over the next five years and 200 MHz more in years 5-10 of the broadband plan.

But finding the spectrum and making it available to operators are two different things. Finding spectrum means identifying spectrum that could be used for broadband wireless, and then finding a new portion of spectrum to relocate the existing license holders (all available spectrum is used by someone). Then the spectrum must be auctioned and the incumbents moved to another portion of the  spectrum by the willing bidders. The bottom line is that it will be 5-7 years before more spectrum really becomes available for us to use, and, by then, demand for wireless broadband services will have doubled and doubled again.

Given this shortfall of spectrum vs. the demand for wireless broadband services, and given the fact that AT&T and Verizon  (and other operators) are trying their best to manage the spectrum they now have, including doing away with all-you-can-eat pricing, why would wireless network operators even consider pushing cloud computing, which will drive up demand that much more? One of the answers is that network operators are searching for ways to increase their revenue stream. Voice and data pricing are falling and apps stores are not contributing to the network operators' income, so they are all searching for ways to increase their revenue and cloud computing is a hot topic.

There seems to be a disconnect inside the network operators' companies when it comes to data services. Marketing and sales are charged with increasing the number of customers, and offering the latest buzz in the form of cloud computing would seem like a great idea. Not only will customers use more data (and pay for it), if the cloud is within the network, the customer is more captive on that network than ever. If the cloud is not part of the network, operators still gain data revenue by pushing up the average customer's data usage.

Meanwhile, in the network operations and management side of the business, they are trying to find ways to make their networks more efficient, how to serve all of their customers at an acceptable level, and how to deal with data hogs. Perhaps the two sides of the business are not comparing notes about the effects of cloud computing, or those on the network side of the business have not fully realized the effect of cloud computing on their networks if it does, in fact, become the "next big thing."

It seems logical to me that with today's faster processors, better battery life, inexpensive memory, and bandwidth limitations that the networks should be encouraging the opposite: Store all of your files and applications on your devices and access the network only when you need to send your work to others.

My concern is that when I really need access to my data and applications I won't be able to make a connection. It appears as though many cloud believers take connectivity for granted, which can be risky. In both the wired and wireless worlds, network capacity is shared capacity, and Murphy's Law says that if something can go wrong, it will.

So cloud systems rely on connections that are not 100 percent reliable, a wired network that is not managed and is not mission-critical, wireless networks with shared bandwidth, and a site that its designers believe is hacker-proof. I don't know about you, but I sure don't like these odds. But like all things Internet and wireless, there are those who sit up at night trying to discover the next big thing to make money. I guess it shouldn't, but it always surprises me how often these ideas are reborn out of older ideas tried in a different time.

Yes, we live in different times, times when wired and wireless connectivity are better than they ever have been, when access is faster, devices are smarter, and networks are smarter, but that does not mean we live in a world where connectivity is 5-9's reliable (99.999%). Will we ever be there? Probably not. In the meantime, I prefer the old fashioned way of keeping my applications and data safe and with me at all times. I carry a device that can store it all locally, I sync it with my network when I get back to my office, and if there is an Internet attack I can disconnect my network from the Internet in less than a minute.

Wireless operators can also disconnect their networks from the Internet instantly if needed. Yes, you would lose Internet connectivity, but you would still have voice, text, and data over the network until the threat from the Internet subsided, but you probably would not have access to the Cloud.

What does this have to do with the wireless developer community? I believe developers should not get caught up in the cloud hype and should take a more realistic approach to applications and data services. I store my data in the cloud as a backup, and it is encrypted, yet I have to concede that someday someone might hack it, so there is some data I store on a removable disk off-site, not connected to anything. Perhaps someone working on cloud applications will give me an option of where I want to store my data-some in the cloud, some on a removable device-and all encrypted for added security.

Data encryption and compression are two areas where I would like to see more progress, not only for corporate customers but for small and medium business customers as well. Unfortunately, we live in a world where all of our data needs to be encrypted, both on our devices and when it is transmitted. This is a problem because the intended recipient would need to use the same encryption. But years ago when encryption was "the next big thing," there were many packages that used public/private key systems, which work well.

If, like me, you don't fully trust the cloud, look for opportunities to help users more effectively sync their data, keep it up to date on multiple platforms automatically, encrypt it, compress it, and make all of this seamless and in the background. The cloud is our future? If you believe that, good luck. Ask the IT professionals at Walgreens and they will tell you that in spite of every precaution they took, they were still hacked. If you are a Walgreens' customer, your data is now in the hands of someone else.

Andrew M. Seybold is an authority on technology and trends shaping the world of wireless mobility. A respected analyst, consultant, commentator, author and active participant in industry trade organizations, his views have influenced strategies and shaped initiatives for telecom, mobile computing and wireless industry leaders worldwide. www.andrewseybold.com.