A colleague of mine recently posed the query: “What is the real difference between cloud security and data centre security?” Really it is all the same - servers, storage, networking and so on. I didn’t really think too much about it, however, after thinking about this for a while, I have changed my mind. There are a number of distinct differences which cover a range of security practice domains.
The first difference really breaks down the definition of what the primary differences are between the “cloud” and a “data-centre”. Many IT folk consider the “cloud” as a new buzz word for services delivered by data-centres. This is true in some instances, but in many cases for a cloud, the data centre could actually be many data centres, not necessarily located on the same continent. Essentially the cloud is the virtualisation of the data centre.
From a security practitioners perspective this emphasises location, the traditional fall-back for the security manager. Draw a circle around it and everything inside it is the security manager’s concern while everything outside the circle is someone else’s problem. So how does one circle a cloud? Don’t begin stressing too much, it gets worse yet, so save some stress.
The nature of cloud services is based on the fundamental and constant availability of processing power, memory, network performance and data. The key here is the last point, data. The primary action cloud services take to ensure ready availability of this data is via replication and distribution. In short, mirroring that data in several physical locations, across the country, or across the world. So what threats does the cross-jurisdictional copying of your organisation’s data pose to the compliance/governance teams? And how does one keep track of how many locations these copies are distributed to?
A number of years ago the OpenGroup’s Jericho forum (which ran from 2003 through 2010) came up with the term “de-perimeterization” which essentially meant pushing down the traditional network boundaries to ensure that security resources were not being complacent—only monitoring what was considered the circle or network boundary, ignoring prospective threats from the inside, and simply attempting to defend against those coming from the “outside”. Fundamentally one cannot trust the boundary, threats are proven to readily punch through defenses undetected.
Really the cloud is an extension of this. Can anyone be certain where each and every mirror of the organisation’s data resides? And how does one protect that, let alone ensure the prospective jurisdictions are “friendly” to cyber-security and their law enforcement/legal frameworks are as evolved as each other to ensure compliance with your business requirements.
Time to get locks on those rack cabinets? Maybe some electronic access control cards for the door to the cloud?