What's the big deal with Cloud VPS?

Question

What makes a "cloud" VPS far more superior to an ordinary Xen VPS?

Answer 1

Cloud is a nebulous marketing term. There are as many different definitions of "cloud" as there are addresses in the IPv6 space (a lot).

Generically speaking, cloud VPSes (like Amazon's EC2) are generally meant to be used in somewhat of a dynamically-scaled architecture. Providers of these systems make it very easy to programmatically start and stop large (or small) numbers of instances at need.

Standard VPSes (like those that, say Linode provides), are more of a replacement for dedicated servers. They're good for long-running services that most likely don't need to scale as much nor as quickly as others.

Like I said, though, there's no one definition for "cloud", so the above statements are very general, and there are exceptions to both of them.

Answer 2

Cloud computing is nothing new (contrary to what any modern web marketers might want you to believe). It's synonymous to (or at least, similar to) parallel computing, distributed computing, grid computing, etc.

My interpretation of the basic idea of the model is that it is a system which allows for a dynamic, on-demand distribution of computing resources. The Xen hypervisor has the ability to do this with Xen Cloud. However, that does not mean that a "Xen VPS" is a "cloud vps".

The term "cloud computing" has become so convoluted in marketing gibberish lately, that it's even become impossible to compare "cloud" services, since no technical implementation or requirements have been defined, and nobody can seem to agree what the heck it's supposed to mean.

That being said, assuming that by cloud vps you mean a virtual-private server sitting on a grid-based hosting system compared to a single machine using the xen hypervisor to run a virtual machine that is your web server, the major advantages are two-fold:

• depending on the provider, your application isn't limited to a static amount of processing power and memory. the more power your application needs at any given time, the grid-system can supply to your application, on-demand, automatically.
• depending on the provider, your application may suffer less from geographic distribution, since in some cloud implementations, your application can be distributed through the use of content-distribution networks (CDNs).

These benefits however, greatly depend on the "cloud" provider and their definition and implementation of the "cloud".

Confused yet? ;-)