Barron’s Tiernan Ray recently published a great blog about how the ‘Old Guard’ including Cisco, Dell, HPE are taking advantage of the backlash against the surprisingly high cost of cloud computing. As it turns out the race to zero initiated by the cloud leaders (AWS, Microsoft, and Google) was really a strategy to get organizations to host applications or store data in the cloud for almost nothing and then entice those customers with rich and powerful services to augment data and applications. Those additional services are far from free. I’ll be the first to admit that these are compelling offerings and they are growing exponentially. Have you seen the list of AWS services available today? It’s a periodic table of cloud technologies.
I don’t see the shift back to the old guard as real backlash as much as I do a more rational distribution of computing resources - cost, processing, hosting and decisions - between the edge and the cloud. There is an ironic brilliance in one of my favorite executives in technology today - AWS CTO Werner Vogels. If you have a chance to hear him speak - especially someplace other than an AWS event - do so. The man is both brilliant and articulate. So it was a surprise, actually a shock to me, when I learned the title of his blog. Hint - it’s not “All Things Cloud”. Vogels’ blog is called “All Things Distributed”. Even as he/they were building out the most powerful cloud company in the world, he knew that organizations must and will rely on highly distributed computing architectures. AWS has introduced more than one service that helps organizations decentralize their computing capabilities.
In one of his earlier blogs, Vogels talks about some of the reasons that edge computing is important and necessary. The laws of economics certainly are on this list. To Tiernan Ray’s point - why would an organization unnecessarily pay to store all its operational data in the cloud or process it in the cloud when it can just as effectively be done at the edge. Vogels also points out that the laws of physics are at play here too. Edge, or operational, data is heavy. Devices can be chatty, generating lots of data very frequently that is actually superfluous to decision-makers. Edge computing allows organizations to see and analyze the data where and when it is generated.
The fact is that connected devices are almost always endowed with computing power, not just connectivity. That means that these ‘smart devices’ have the ability to act as an intelligent agent, acting in real time as events occur. It used to be that sensor and machines that produce data had to be hardwired into broader systems that did the actual thinking for them and then triggered a series of actions and reactions. But the smarts at the edge is another reason why you just don’t need to send data to the cloud for massive numbers of auto-scaling servers to receive, translate, analyze and act upon events. There is too much latency, risk and cost in this model to use it for everything.
I don’t see this so much as a question of the old guard or new guard, but rather new thinking. In the new world, computing will be highly distributed with expertise and technologies at the edge and cloud each serving unique and important purposes. Of course the titans - whether new or old - will need stories and offerings for both domains, but that’s great for customers and the marketplace as a whole. So avoid going all-in with anyone or anything. Balance is more than a hedge - in this case, it’s just the right thing to do.