Livin’ on the Edge: What Comes Next After Cloud Computing
It’s too early to say whether or not associations should take edge computing, the successor to cloud computing, seriously. But what should be taken seriously is its potential. Any good association IT department should research potential disruptors ahead of time.
Cloud computing isn’t dead yet. Far from it.
But it is becoming more mature. As Gartner recently pointed out, the technology is now in its second decade, at a point where if you’re not using it for mainstream enterprise applications yet, you should really get on that.
But maturity implies a point of limitation—that things have reached their peak. And that’s leading some to wonder what’s next. Even as the cloud defines the modern technology stack—and further, the way we interact with our vendors—there’s always a good chance that it may not stick with us forever.
That sounds more than a bit premature, but some of the technology world’s most prominent thinkers are already making this case.
What’s Edge Computing?
Last year, Andreessen Horowitz analyst Peter Levine made this very point in a 25-minute presentation that contextualized the need for offsite storage in the context of historic technology trends.
“Everything that’s popular in technology always gets replaced by something else. It always goes away,” he said. “And that’s, you know, that’s either the beauty of the opportunity, or the beauty of the business—that these things actually go away.”
This claim is bold, but he has a larger point that he’s reaching at. Levine makes the case for something called “edge computing,” a phenomenon which he suggests will come about as our computing devices become so sophisticated that it no longer makes sense to offload computing ability to an exterior source—as, for example, devices that rely on the Internet of Things (IoT) become increasingly sophisticated.
Here’s the gist of what that means in real-world terms: Once, computing power was so resource-intensive that we needed a server room. Cloud computing moved that server room off the premises. Eventually, however, the devices we use on a daily basis will be sophisticated enough that they’ll have as much computing power as the machines in the server room—without all the extra space.
There will still be a need to offload data elsewhere, but most of the computing capabilities that the cloud once managed can now be handled on the premises—which is where you want them, because it’ll save money and time to manage that data onsite.
In comments to Business Insider last year, Levine noted that as a greater number of sophisticated computing devices become more common, we’re going to reach a theoretical limit on what a remote computing platform can do.
“You will never have enough bandwidth and speed on the network between [artificially intelligent devices] for that,” Levine told the website.
The result is that the cloud will still be there, but it will take on a management role—orchestrating the execution, as the technology on the ground does all of the work.
Why Associations Should Care
The problem with something new, like edge computing, is that it often feels like an idea without a use case. But the use cases, over time, will show themselves, even in spaces like associations.
Amazon Go is probably a great example of where something like edge computing would come into play. The concept, announced late last year, effectively would allow grocery stores to work without human input—automatically charging customers as they pick up items, walk through the store, and walk out.
But, according to The Wall Street Journal, the problem with the approach is that, when too many people are in a store, it starts to have trouble tracking all the activity. At some point, managing all this data via an offsite cloud system is simply going to complicate the experience.
I have no clue what Amazon is doing to solve this problem, but edge computing would probably come in handy in a case like this. The cloud needs a connection to enable the transaction, but it’s a case where so much data needs to be pushed around the store that adding an outside element like a cloud connection could gum up the works. Sure, the cloud needs to be there at some point of the transaction—say, during the actual payment—but if not properly contextualized, it’ll get in the way.
Likewise, there are other real-world situations where a lot of data is involved—and as a result would probably benefit from an edge computing approach. In terms of associations, I could see the immediate benefit with events that have embraced beacon or IoT technologies. As those approaches become more sophisticated, we’ll ultimately want platforms that can not only collect data but also react to it in real time.
What would be the benefit? Ernest Sampera, the chief marketing officer for vXchnge, recently explained in a Data Center Knowledge op-ed that the benefits of this approach will best be seen in real-time use cases.
“Today’s data management systems require the most immediate information to support ‘in the moment’ decisions that can have an impact of millions of dollars to the bottom line,” he wrote. “By bringing processing to the edge of the network, businesses reduce latency by prioritizing processing and lightening the load on the primary network, supporting better, faster decision-making.”
Know What’s Coming
Let’s be clear: I’m throwing this out there knowing that the answers aren’t all laid out. They won’t be for a while. But opening up this discussion so early is a great way to contextualize how IT departments make critical decisions.
There was a time when the things being said about edge computing could be said about cloud computing, where the approach wasn’t ready for prime time … and then suddenly, it was. (Going back a little further, it’s worth keeping in mind that the internet was once the same way, and just think of how obsessed we are with that.)
But if associations are going to eventually gain a reputation for innovation, they need to be thinking a couple of steps ahead.
A cloud computing model is great now, but will it be forever? Depending on the use case, not necessarily.