For Some, Amazon Takes the Cloud Out of Its Cloud

We can add another item to Amazon Web Services’ recent major improvements, as Monday morning, it announced dedicated EC2 instances for customers using AWS’s newly upgraded Virtual Private Cloud (VPC) service. Instead of customers sharing their servers with other customers, Amazon is bowing to the security-conscious or performance-minded, and offering them their own dedicated servers. Because sharing multiple virtual machines in a single server in a multitenant environment is a key definition of cloud computing and how clouds can provide economies of scale, Amazon’s move is noteworthy and points to the maturation of the industry and its service.

Further, this is yet another example of AWS pushing the innovation envelope despite its customer lead, although Dedicated Instances are a bit pricier than the recent free tools like CloudFormation and VPC.

Amazon could have continued growing its business without offering dedicated servers, but the company appears to have been on a mission lately to improve its image among all constituents — developers and CIOs alike. However, Dedicated Instances will probably give the most comfort to those higher up the command chain simply aren’t comfortable sharing physical servers with other customers. Running on dedicated servers certainly doesn’t mean customers can rest assured on security, but it does eliminate some of the dangers — both inherent and perceived — of running in a multitenant, virtualized environment. Here’s how Amazon’s Jeff Barr describes the new feature:

Each Virtual Private Cloud (VPC) and each EC2 instance running in a VPC now has an associated tenancy attribute. Leaving the attribute set to the value “default” specifies the existing behavior: a single physical machine may run instances launched by several different AWS customers.

Setting the tenancy of a VPC to “dedicated” when the VPC is created will ensure that all instances launched in the VPC will run on single-tenant hardware. The tenancy of a VPC cannot be changed after it has been created.

You can also launch Dedicated Instances in a non-dedicated VPC by setting the instance tenancy to “dedicated” when you call RunInstances. This gives you a lot of flexibility; you can continue to use the default tenancy for most of your instances, reserving dedicated tenancy for the subset of instances that have special needs.

It’s worth noting, though, that running a dedicated server in Amazon EC2 is not akin to running a dedicated managed server with another cloud provider such as Rackspace. Probably the most noteworthy difference is that AWS instances aren’t managed, which means customers still rely on themselves and the AWS support channels to keep applications up and running. Further, customers don’t have complete access to the server, meaning they can’t use it to whatever ends they please, such as maximizing value by running multiple applications in a single dedicated server. In fact, writes Barr:

It is important to note that launching a set of instances with dedicated tenancy does not in any way guarantee that they’ll share the same hardware (they might, but you have no control over it). We actually go to some trouble to spread them out across several machines in order to minimize the effects of a hardware failure.

AWS isn’t alone in offering dedicated cloud infrastructure. GoGrid and SoftLayer are among the other cloud providers offering dedicated hardware with pay-per-use billing and on-demand provisioning rather than as managed servers.

Per-hour prices are higher than their multitenant counterparts, and users will incur a $10 per hour “opportunity cost” fee for each Availability Zone in which they have a Dedicated Instance running. AWS justifies this by pointing out the lost business on its end by rendering the majority of any dedicated server unusable by other customers, and Barr notes that this fee is flat regardless how many dedicated servers are running in that data center. I don’t know how that $87,600 per year (assuming at least one Dedicated Instance is always running) compares with what AWS actually loses by not using that space for multitenant instances, but it probably comes a lot closer to even — and looks a lot less daunting to customers — when customers are running hundreds or thousands of dedicated instances versus just a handful. One could argue this is a prime opportunity for AWS to invest in micro servers running ARM, Atom or other low-power server processors once they hit the market, but AWS hasn’t indicated any plans along that line.

It’s far from perfect, but as with most things AWS, it’s important that the company rolled out the feature in the first place. Certainly, some large businesses with tight security needs will see real value in Dedicated Instances, and others might over time. If AWS has proven anything over the past year, it’s that it isn’t afraid to regularly drop prices and improve capabilities, something that might be necessary to make Dedicated Instances financially feasible for run-of-the-mill users.

Image courtesy of Wikipedia Commons contributor Abrahami.