For quite some time, I have been struggling with coming to grips with the whole notion of the Cloud Computing industry. I had trouble trying to figure out why anyone would want to use the Cloud. My reasoning was this: Pretty soon, everyone would have machine with an 8 core processor running at a clock speed of a few GHz, many Gigabytes of RAM and Terabytes of memory on their hard disks. With so much computing power available so cheaply, why would someone turn over all their computing tasks to a remote, unreliable Cloud? Why torture all that computing power to work through a bottleneck of around 1 Mbps? In most cases, people would be lucky to get an Internet connection of 1 Mbps.

The second related issue that confounded me was the ever increasing number of cores on processors. For the average PC user, what is the benefit of multi-core processors? At home, I use a 2004 Dell Inspiron 600m, with an Intel Pentium 1.6 Ghz processor, and 512 MB RAM. At work, I use a 2008 Dell Desktop with a Core Duo processor in it. And guess what? I barely notice the difference between the two machines. Granted I no longer do any compute intensive tasks on my home laptop but for all other tasks such as Word processing, spreadsheets, Web, video etc, there is no noticeable difference between the two machines. Even Don Knuth, expressed his unhappiness with multi-core architectures. And if the Don himself had doubts, what could I say? I reasoned that multi-cores were just the next Marketing fad for Intel. Back in the 1990’s, the microprocessor clock-speed was the most important feature in a computer. Every 18 months the frequency kept doubling and Moore’s Law relentlessly marched forward. By the early 2000’s the clock speeds ran into the power ceiling and the focus changed to low-power computing. With CMOS technology rapidly reaching the physical limits of miniaturization, multi-cores seemed to be the best way forward. Forget the fact that there were hardly any software applications to take advantage of these cores, the chip vendors kept making them and marketing them. It looked as though they did this for one simple reason: because they could. The number of cores then took over as the USP. And now, there is plenty of talk that “cores are the new transistors”.

Fast forward to 2011 and the tech-world seems to be forming a different picture, one that ignores the traditional big players (Microsoft, Intel) and has many new players. The problem with my analysis was that I had turned a blind eye towards the oncoming Mobile revolution.

My questions were simply the wrong question to ask. The question is not why people would waste their computing resources. The question is: Do people even want that computing power? It might be true, that all that computing power would be available cheaply, but people seem to be increasingly picking mobility over power. The ability to access information from anywhere at anytime has become more important that computing power. This is the basic explanation for the rise of the tablet/mobile revolution. It turns out that the average PC user does not even need a PC let alone a high-performance, multi-core machine. All the average user needs is a tablet or a smart-phone to surf the web, read tidbits of information, watch videos and interact with friends and family on social media. Currently the most important feature in computing devices is connectivity. Is it 3G or 4G enabled? Does it have WiFi? Does it have Bluetooth? No one seems to care much about what processor is in the mobile device and how much memory and storage it contains. If the device has a good connectivity then the Cloud offers unlimited storage.

Look around in your workplace and ask yourself how many of the employees’ PCs could be easily replaced by a less powerful and more connected device. My guess is that apart from the Engineering team, most other people can afford to replace their PCs. It is only the engineers who run complex applications and tools that require high performance machines. For everyone else, the computing can easily be outsourced to the cloud.

Now, where do Multi-cores fit in this picture? Multi-cores are going to rule the data centers that power the Cloud.  One of the big revelations at the 2011 CES was the seeming death of Wintel and the rise of Armdroid. However the Armdroid is only going to power the consumer’s handheld device. What is going to power the other end of the connection? For this, we need to look at the developments in the multi-core industry.  Not surprisingly, many of these companies have been bought by the big players. The surprising element here is that multi-core technology might not end up being a disruptive technology. It might simply help reduce costs and improve performance at data centers. It seems like the hard problems of parallel computing might after all not have to be solved to take advantage of multi-core processors. The technology might serve very well to simply shrink the size of the data center. Instead of worrying about solving the parallel programming problems, data center architects could simply apply the principles of Queuing Theory and Load Balancing to take full advantage of the multi-core processors.

The Cloud, it turns out, is a necessity for Mobile Computing. And, Multi-core processors will form the backbone of the Cloud.