The term “server” will…um…stay around
In a recent post, Jesse Stay discussed two trends that are affecting people in their homes.
First, people have more “computers” in their homes. I can remember a few short years ago when our family only had a single computer. That is no longer the case in our home, and it is certainly no longer the case in the Stay home: his post mentions Xboxes, a Nest, a Fitbit, a Sonos, and “other devices.” All of these devices have computing capability.
The other trend is the increased use of the cloud for storage. While the aforementioned Xboxes get data from “a server in the closet of my office,” many of the other devices use “the cloud” for storage. What this means is that Stay does not need a server in the home for these devices – the data is stored in a server farm in North Carolina or somewhere.
Based upon these two trends, Stay opens his post with the following statement:
I’m going to go on record – the name “server” is going extinct.
Now I should emphasize that Stay is talking about home use, not enterprise use. Obviously Google and Amazon and the like are still going to need servers. But Stay is saying that for the home user, he believes that the concept of a local server is going to become obsolete – all the data will be stored somewhere else.
Well, Jesse, while you may not want to use that six-letter “s” word in your home, I’m not quite ready to banish it from my home just yet.
Why? Because of my pendulum theory.
If you haven’t encountered my pendulum theory, I first proposed it back in 2009 (although I didn’t use the word “pendulum” in this initial statement of the theory). I used it in reference to Amazon’s mimicking of something that CompuServe did long ago.
Basically, ever since computers were invented in the 1940s or the 19th century or whenever, the computing industry has oscillated between two different models of computing:
* The Benevolent Model, in which a central service provides everything that the users need, including programs and processing power. All the user needs is a dumb terminal, something that acts as a dumb terminal, or something even dumber like a punch card reader. The central service takes care of everything for you. There is nothing to worry about. Dave?
* The Rugged Individualist Model, in which a computer user doesn’t need anybody else to do anything. A single computer, in the possession of the computer user him/herself, includes all of the power that the user needs. We don’t need no central service; we don’t need no thought control.
Now obviously these are the extremes, and there have been some computer trends (like client/server) that somehow combine the two. But it still seems like we alternate between the two models, and now the cloud computing model has us all leaning a little more toward the centralized model.
In the Stay household (ignoring the Xbox for the moment), data is stored by a central service – actually multiple central services. The important thing is that it’s not stored in the home itself.
But I don’t think that it will always remain that way.
A few years from now, the Stays may run into an instance in which the cloud rains on them. Perhaps some cloud provider will have a major security breach. Perhaps a cloud provider will jack up its storage rates. And the Stays, being of a technical bent, may end up saying to themselves, “It would be cheaper and more secure to store this stuff here in the house. I think there’s some room in the closet of Dad’s office where we can stick a server.”
And the six-letter “s” word may be spoken in the Stay household yet again.
P.S. I do want to correct one misstatement that I’ve made in my previous posts on this topic – specifically, in my November 5, 2011 post in which I said:
Now I’m sure that Google and other services put enough redundancy into their systems to minimize the occurrence of outages. But there is no “cloud” company – none – that can guarantee 100% availability to its users. It is literally impossible to do so.
Now the statement itself is accurate – but its implication is wrong. For while it is true that no cloud service can provide 100% availability, it is also true that no self-storage solution can provide 100% availability. As someone who has gone through a hard drive failure, I should have known this better than anyone. (Incidentally, that hard drive failure occurred at the same time that Facebook acquired FriendFeed. Go figure.) So the question remains – can cloud services provide higher availability (and accessability) than you can if you store things yourself?