A cloud of confusion
“Just put it in the cloud, and it’ll be safe. Right?”
Well, it’s not nearly that simple. While cloud applications are often viewed as a great way to offload security concerns to a third-party that act alone does not necessarily make your data more secure. On the contrary, a poor choice of cloud service provider might instead make your data less secure.
The cloud really isn’t a specific thing or place, nor is it defined by a new or emerging technology. It’s a metaphor and a buzz-word at the same time. All “cloud” actually means is that the resource in question is remotely located and Internet-accessible. This definition includes many rather mundane concepts such as offsite data centers, hosted web and/or application servers, as well as web-based email such as Yahoo. Google Docs and Dropbox, which many users take for granted (both in terms of accessibility and security) are located in the cloud.
In summary, if you’re using it, and you or your organization does not support it onsite using your own resources, then it’s probably cloud.
Cloud security characteristics
What characteristics make the cloud more secure than your environment? Scale is most likely one such characteristic, as cloud service providers are typically large enterprises that bring almost immeasurable resources to the table. Unless your organization utilizes all of the latest technologies and techniques to manage threats, supported by an elite stable of IT security professionals, chances are that a malicious attacker will get through your defenses before, for instance, Amazon’s.
Cloud providers also have to meet at least a minimum of security standards to stay in operation, based on whatever laws are pertinent to the service in question. These standards vary greatly, from Federal laws to unenforceable industry recommendations, but they do exist. As a cloud customer, your Service Level Agreement should specify that data security is an integral part of the services being performed; therefore, liability (in the event of a breach) may to some extent fall upon the third party. However, the concept of cyber indemnity is still somewhat a grey area.
Now, scale can work the other way as well. This is due to two factors: one, cloud data centers hold much more data from a variety of end users when compared to private onsite storage facilities. Consequently the risk/reward ratio leans much more heavily towards the “reward” side as far as hackers are concerned; the fact that large enterprises handle so many customers’ data means they can, for lack of a better way of putting it, generally afford to get hit a couple times before consequences are felt. Imagine it from a criminal’s perspective: would you rather infiltrate a single company, or get the data from a dozen companies all in one campaign?
There’s also the issue of data-in-transit; many administrators forget about this topic, focusing on the more obvious issue of data-at-rest. But there needs to be a discussion about how your data gets to the cloud (Internet connection), what the ramifications are for supporting that connection, and what might go wrong even if the connection is configured properly.
Suppose your environment is not intended to be connected to the web—it is an enclosed network segment with no publicly visible IP addresses, nor any gateway between that LAN and the outside world. Then, suppose we add a router and a public IP (which is Network Address Translated to a bunch of private IPs)—now, we have Internet access and can access our cloud storage and backup services, but there’s also an additional access point for bad guys that did not exist before. Was that a net security increase, or decrease?
Let’s also suppose that your data is encrypted while in transit, as it should be. Can it be intercepted while traversing public Internet infrastructure? Technically, yes, although the encryption would be hard (not impossible) to break. And, although you could potentially get a VPN from your gateway to the other endpoint, that adds additional cost, more overhead such as key management, etc. while still not achieving 100% security.
As an IT professional, I don’t believe in security through obscurity. However, I also must concede that some cloud providers are extremely attractive to attackers due to the sheer volume of data contained. While I can entertain the thought of navigating this quandary by selecting a less well-known service provider to store my data, at that point am I simply compromising on the advantages of scale and resources that made cloud so compelling to begin with?
Companies that offer cloud services of all types from SaaS to IaaS want to sell you on those services. If “reducing the size of your IT workforce by 70% by switching to cloud” sounds too good to be true, that’s because it is. Similarly, it sounds like a great idea to transfer risk to a third party—until it becomes evident that the scope of risk being transferred is only a small portion of your organization’s true risk expectancy.