I am forever running into people who are confused about what “the cloud” is. Sadly, some of these are the same people who are promoting and using “the cloud” or “cloud-based services”. What has always confused me is the fact that people who work in IT are confused about the cloud. To be fair, I don’t blame those who are confused (at least not if they have made any attempt at all to resolve their confusion). I blame those who have overused, overloaded, and sensationalized the phrase, “the cloud.”
From a technical perspective, “the cloud” is very simple. It’s just a big pool of hardware resources hosted by “not you” (there is an exception I’ll explain later, but stick with me for a moment). Think about it this way.
We used to have “server rooms” and “datacenters” where some IT department managed physical machines and, for the most part, applications were deployed to one or more identifiable, physical servers. Of course, these “server rooms” or “datacenters” lived on a network behind a firewall that people with the proper credentials could access via some other network… frequently, “the internet”.
Along comes “virtualization,” and instead of thinking in terms of physical servers, we started talking about “virtual servers” or “virtual machines”. At first, this primarily meant that applications were deployed to pre-defined, dedicated, fractional portions of a physical server. Of course, this evolved to a point where a “virtual server” was not necessarily a fractional portion of a single physical machine, but could be either a fractional machine OR several physical machines operating as a single physical machine. (For those of you who have been around for a while, this probably sounds like a “cluster”… because for all practical purposes, that’s what it is). Basically, virtualization changed the “unit of measure” from physical servers to sub-components of a server (like CPU, Memory, Disk, etc.) and enabled those sub-components, generally called “hardware resources” to be allocated without regard for what physical machine those resources are plugged into.
Then someone got the idea that housing and managing the physical servers is a lot of overhead, so why not outsource it – and great big “server rooms” were built that anyone could use (often for a fee) in lieu of maintaining their own “server room”. This was a wildly popular idea. IT departments shrunk. Energy costs went down. Hardware purchase and maintenance time and costs plummeted. And thus “the cloud” was born.
If you’re still a bit confused, maybe this will help. Not so long ago, I took an epic journey “to the cloud and back again.” Ok, I admit, it wasn’t as fanciful as it sounds, but I did get a tour of a cloud hosting facility and it was really cool. According to the NDA, I cannot tell you which provider it was, or the location of the facility until 2017, but bear with me while I share with you some of its attributes.
The unmarked, fenced, barbed wired, security-camera-peppered facility was a converted water bottling plant with armed and imposing looking security guards at the sole gate. Security to get into the visitors area was roughly equivalent to airport security. Access beyond that point to offices and meeting rooms involved a badge, a code and a fingerprint scan. Access to the actual “cloud” required an additional card, code and retinal scan. Of course, there were security cameras pretty much everywhere, the windows did not open and composed of “high-caliber bullet proof” material. What had appeared to be emergency exits from the outside had been bricked over from the inside. Basically, the physical security was amazing, even when viewed from the eyes of an ex-Army Officer like myself.
Once we entered “the cloud” area, we were briefed on how the entire facility had been equipped with triple redundant power and cooling. In both cases, over 48 hours of continuous power and cooling was “stored” on site, so both the entire power grid and the multi-source emergency fuel delivery mechanisms could fail for 48 hours without service disruption. All in all, it would take nothing short of a full scale zombie apocalypse or skynet becoming self-aware to take the cloud down.
But what you are interested in is what was in there, right? Exactly what you’d expect. Aisle after aisle after aisle of floor to ceiling racks of servers. Yup, it was nothing more or less than the biggest server room you’ve ever imagined. I don’t recall the exact amount of total storage available in the facility, but I do recall that it was in was a large number of Exabytes “and growing exponentially annually”.
But what about this “private cloud” thing? There are actually 2 types of private cloud, self-hosted and remote hosted. Self-hosted private clouds are nothing more than your “old school” server room being managed as hardware resources instead of being managed as physical servers. A remotely hosted private cloud is simply taking all that equipment out of your server room, locking it away in a secure container and shipping it to the cloud hosting facility. Nothing changes except the location and (in most cases) a dramatic increase in physical security.
So you see, “the cloud” is nothing to be confused about. It’s just a big pool of hardware resources that is usually hosted by “not you”. What is confusing is how people talk about it, and all the acronyms that go with it, like SaaS, PaaS, and IaaS. If you’re unsure about those things refer to Greg Mooney’s recent blog “Confusion in the Cloud”.
- Confusion in the Cloud
- Application Performance Monitoring: A First Line of Defense Against Attacks
- How In-Memory Databases May Remake Your Database
- Why the Cloud Can Mean Storms for Quality Control