In the “not really surprised” category, we have recent news that a forthcoming HTC tablet will use a quad-core processor. Given that dual-core has been more-or-less a common part of mobile devices in the past year or so, moving to quad core was a logical next step. As mobile devices are called upon to do more and more complex and compute-intensive tasks, more power will be needed. So, like the desktop and notebook performance competition that continues unabated, the same will also happen in the mobile space: more cores, more memory, more storage. Of course, this has an impact in terms of physical size and space, dealing with heat and the most noticeable impact, on power consumption.
Will it ever change? Possibly, but some of that will depend on how successful Amazon’s approach to browser support becomes. Amazon is taking the step, with the Kindle Fire, of doing a lot of the heavy-lifting of parsing and rendering web sites on their cloud servers, and uploading the resulting data to the tablet. While most of the focus of cloud computing has been for server-side enterprise functions (like websites, databases, heavy computation, etc), the cloud may actually be more useful to support end-user devices by providing computing-on-demand, particularly for common tasks. That, in combination with cloud storage, would allow devices to have more modest hardware, and focus on extending battery life, because the raw computing can be done on servers with lots of available power, cooling, etc. On the computing side, availability can be easier to address than with enterprise services, because the computing requests are more ephemeral, much like a search in Google. When you want to perform some business transaction on a database, or perform some kind of data analysis, it means that the infrastructure has to be there, all day, every day, without fail. When the servers supporting your enterprise are down, so is your enterprise.
But, for common tasks like rendering web sites, any server will do, and you don’t have to use the same server every time. The data connection between the end-user device and the servers has to be fast and reliable, but again, each connection only has to last for a brief period of time. This isn’t a long-lived database connection exchanging data with a database. It s more akin to a Google search: render this site, give me the results. In most cases, that will take a few seconds to complete. The data itself doesn’t have to be preserved, so there isn’t any data that has to be protected for any longer than the duration of the request. This makes security simpler, because you aren’t protecting something for long periods of time.
When looking at cloud storage for mobile devices, though, the same issues that exist for enterprise services apply here. The services must be available at all times. The data must be protected from damage or theft. Right now, cloud computing is in its infancy, and these problems have not really been solved in any permanent sense. There are still issues with availability, reliability and security with all cloud services providers, and your average consumer isn’t going to understand the details, but they will understand the implications when these services fail, and they won’t like it.
Until cloud computing can become about as reliable and predictable as a dial tone, we can expect to see more and better hardware continue to appear in mobile devices. The pressure for more will continue to increase as mobile devices, particularly tablets, become the primary computing platform for more people. We are still a few years away from cloud services to reliably and predictable augment mobile devices.