It’s time to stop talking about imaginary trolls under the cloud bridge and start talking about the real security challenges that exist in cloud computing


I’ve been watching with interest a Twitter stream rss of information coming out of the Gartner Data Center conference this week related to security.

There have been many interesting tidbits that, as expected, are primarily focused on cloud computing and virtualization. That’s no surprise as both are top of mind for IT practitioners, C-level execs, and the market in general.

Another unsurprise would be the response to a live poll conducted at the event indicating the imaginary “cloud security” troll is still a primary concern for attendees.

I say imaginary because “cloud security” is so vague as a descriptor that it has, at this point, no meaning.

Do you mean the security of the cloud management APIs? The security of the cloud infrastructure? Or the security of your applications when deployed in the cloud? Or maybe you mean the security of your data accessed by applications when deployed in the cloud? What “security” is it that’s cause for concern? And in what cloud environment?

See, “cloud security” doesn’t really exist any more than there are really trolls under bridges that eat little children.

Application, data, platform, and network security, however, do exist and are valid concerns regardless of whether such concerns are raised in the context of cloud computing or traditional data centers.


For those of you unfamiliar with the concept of a “godwin” I will summarize: the invocation of a comparison to Hitler in a discussion ends the debate.

Period. Mike Godwin introduced the “law” as a means to reduce the number of reductio ad Hitlerum arguments that added little to the conversation and were in fact likely borne of desperation or angst. What seems to be happening now in the cloud computing arena is the use of what amounts to a reductio ad Cloud Cautio argument that also adds little to the conversation and is often also borne of desperation or angst. Let’s call this a “cloudwin”, in honor of Mr. Godwin, shall we?

“Cloud security” is a misused term that implies some all-encompassing security that exists for a cloud computing environment. The cloud has no especial security that cannot be described as being as application, data, platform (hypervisor) or network related. This is important as the type of cloud we attach “security” to directly impacts both the provider and the consumer’s responsibility for providing that security. In a SaaS, for example, the application, data, platform, and network security onus is primarily on the provider. In an IaaS, however, application and data and even some network security lies solely in the lap of the consumer, with bare infrastructure and management framework security the responsibility of the provider.

I could spend an entire blog just making reference’s to Hoff’s twitterbird blogs on the topic. If you haven’t read his voluminous thoughts on the subject, do so

Suffice to say that the invocation of the term “cloud security” as a means to justify avoiding public cloud computing is equivalent to a cloudwin. Invoking the argument that “cloud security is missing” or “lax” or “a real concern” adds nothing to the conversation because the term itself means nothing without context, and even with context it still needs further exploration before one can get down to any kind of real discussion of value.


If what you’re really concerned with is data leakage (data security) then say so.

If what you’re really concerned with is network security, say so. If you were to ask me, “Hey, what can F5 do to address ‘cloud security’” my response would invariably be, “What are you trying to secure?” because we can’t have a meaningful conversation until I know what you want to secure. Securing a network is different than securing an application is different from securing a network in the data center than it is securing a network in the “cloud”.

We need to stop asking survey and research questions about “cloud security” and start breaking it down into at least the three core security demesnes: application, data, platform, and network. It’s fine to distinguish “application security” from “IaaS-deployed application security”.image  In fact we need to make that distinction because securing an application that is deployed in an IaaS environment is more challenging than security an application deployed in a traditional data center environment. Not because there is necessarily a difference in the technology and solutions leveraged, but because the architecture and thus topology are different and create challenges that must be addressed. 

Securing applications, data, and the network “in the cloud” is a slightly different game than securing the same in the local data center where IT organizations have complete control and ability to deploy solutions. We do need to differentiate between the two environments but we also need to differentiate between the different types of security so we can have meaningful conversations about how to address those issues. Yes, a web application firewall can enable data leak prevention regardless of whether it’s deployed in a cloud environment or a data center. But the deployment challenges are different based on the environment. These are the challenges we need to speak to but can’t under the all-encompassing and completely inaccurate “cloud security” moniker.

So let’s stop with the cloudwin arguments and start being more specific about the concerns we really have with cloud computing. Let’s talk about application, data, platform, and network security in a cloud environment and the challenges associated with such security rather than in broad, generalized terms of a “thing” that is as imaginary as trolls under the bridge.

Follow me on Twitter  View Lori's profile on SlideShare  friendfeed icon_facebook AddThis Feed Button Bookmark and Share

Related blogs & articles: