In a sense, yes If you suspect that cloud is an over-used marketing term, you’d be right, but to only to an extent. Executed properly, the technologies underlying modern cloud based services provide new and always improving levels of reliability and scale. These technologies drop the barriers to benefit, and in business terms, this opens up new opportunities by leveling the playing field and allowing for new levels of scale and reliability.

With this philosophy, one can craft all sorts of online services whose components are massively distributed and replicated, all working in unison and tolerating all sorts of failures and spikes in demand. Availability, capacity, and a single point of management. Well, at least that’s the way it is meant to work.

In an ideal world, an online service (say a website, ERP app or other “workload” or “service”) should be able to hop and exist across the boundaries of individual physical servers and even geographical boundaries in order to tolerate failures, ramp up resources when needed, all through a really easy interface (or perhaps even semi-automagically).

This raises the question of just how all this can actually be achieved. So we may ask the ‘what if’ question of ‘what if it just worked and we didn’t have to depend on hardware?’ That’s a great question. And so, let’s dispense with our reliance on single pieces of hardware. This would imply that the computing work isn’t being done on any one piece of hardware and there’s some phantom worker that does the computing. It may be a silly way to phrase things, but that’s precisely where great ideas begin.

Virtualized reality

So, instead of a phantom computer, the industry standard term is actually called a ‘virtual’ machine – the concept remains the same where computers become “virtualized”. One of the benefits of virtualization is that many machines can run on a single powerful machine. An additional benefit is that virtual machines can be moved between the hardware that hosts them, just as easily as copying a file. Resources like storage and networks are shared, utilizing the computing power and space to the maximum extent possible. This yields a high degree of utility and cost effectiveness that cloud (or really, “virtualization”) providers pass onto their customers. There are many other ways to further virtualize storage, databases, and websites.

Keeping the lights on

Services online these days must satisfy the contract of “it just works”. The lights need to stay on and customers and staff need to be served. The dependence on technology like internet connectivity and computers to run business processes are now so high, that upwards of 75% of an average business’ processes depend on them. Being down is not an option.

Since virtual machine workloads can move between hardware, it means hardware failures can be tolerated and services become “highly available”. Therefore, virtualization and its so-called high availability benefits can endow a workload with certain guarantees of reliability. Virtualization technologies allow infrastructures to be built such that if a single physical server goes down, there’s always another one to take its place. More powerful setups are built such that if a whole bank of servers go down, there’s another one…and so forth. Now imagine scaling that concept to a planetary scale. That’s quite a fantastic concept because it means services remain available, and they can scale up and allow businesses to grow.

The downside is these setups can become very expensive to build, at least for the average company. Cloud providers, however, are really great at creating and maintaining large cloud setups that can handle large numbers of workloads. Cloud providers can do this cost-effectively and at scales that aren’t tenable for the average company. But the bottom line is that the dependence on single pieces of hardware are removed. You can run your own cloud in your office, but it won’t have the massive levels of scalability or reliability from all the multiple levels of redundancies offered by cloud providers.

Small steps & hybrids

Running a local office cloud is a great first step. It also leads into the interesting possibility of combining that local office cloud with that of a cloud provider. There are many reasons to do this: compliance, control, and sometimes simply plain comfort level – not everyone is comfortable giving up control of their data and assets and uploading it somewhere into the ether. This is an understandable fear, but hopefully this article has managed to allay some of the concerns. Combining clouds is both possible and encouraged. Not surprisingly, these are called “hybrid” clouds.

Hybrid clouds and cloud-based disaster protection and recovery services use virtualization to deliver on that contract of “it just works” by letting a large number of potentially geographically disparate resources act as one. If your office file store crashes, you can still access your files. How? Your file share ‘cloud’ spans both your local resources and those of your cloud provider’s, acting as one virtual resource.

Part 1: Cloud, what is it Good for?

First part in the series make the concept of cloud crystal clear

Part 3: Your identity dictates what you need

This series brings up the matter – probably the most important one in this discussion – of identifying internal needs and therefore who can help.

Part 4: The money questions

This series ties up the final matter of price and ties the series together into a whole.