In Cloud We Trust: A New Model

Everything is moving fast in the cloud. Public cloud services are delivered with the promise of instant provisioning and constant availability at a fraction of the cost of on-premises implementations. But given our mobile, cloud-enabled, data-driven environment, it is highly likely that in the near future there will be a significant data breach as a result of a compromised cloud account with excessive access to sensitive customer data.

While the prediction is that cloud service providers will be attacked, business consumers will have the most to lose if they are compromised. So how do we control the uncontrollable: the nexus of cloud, mobile, and a data-driven environment? CISO’s and their InfoSec teams need to simplify the problem, the solution, and how they speak with the rest of business operations.

A failure to communicate
Currently, most security teams don’t effectively communicate with business operations. However, by doing so, CISOs and InfoSec teams will have a much better chance to partner with the key internal influencers and enable smarter data security.
 
Back in the day, information security frameworks were created at a time when data was primarily housed behind a corporate firewall and the focus was on placing controls on the endpoints and key points in the infrastructure. As a result, existing frameworks do not cater to a cloud-enabled enterprise in which the traditional enterprise perimeter has evaporated, control of the endpoints is greatly diminished and there are new, higher levels of data fluidity than ever before.

Many public cloud-consuming organizations look for cloud providers to adhere to particular regulations and standards both because they are a business requirement and because they equate them with trustworthiness. However, even though the intent of these regulations and standards is to protect, the long rote of compliance has taken its toll. When faced with an overwhelming number of compliance obligations and audits, the tendency of CSPs is to treat compliance in a checkbox fashion versus taking a risk-based, sustained approach where controls are assessed on a continual basis rather than at just a single point in time. What that means is that while compliance requirements may prove that a particular scope of an environment was compliant at a specific state in time, it does not guarantee the protection of data.

Furthermore, the way that InfoSec teams classify and treat data has to align to the new business and usage contexts of cloud and mobile. In other words, data identification and classification schemes have to be intuitive and simple in order for the business to own up to protecting their data.

A new model for data sets
A simple-tiered model including three categories of regulated, commercial and collaborative could be an industry-agnostic model for tiering data and provide a construct for educating business users about to how to identify critical data assets and transactions. It also ensures that the business is able to have an intelligent discussion with the cloud service provider about protecting data.

Specific industries and verticals may want to tailor data classifications schemes to the values that data holds for their particular industry. However, in the baseline model we propose, regulated would classify any data that is subject to regulation including Personally Identifiable Information (PII), commercial would cover any business to business or business to consumer transactions, industry data or intellectual IP and collaborative would imply data such as document collaboration, DevOps, and also that which is publicly accessible.

We propose these tiers as a base model because they can be applied concretely to define data and minimize misinterpretation versus definitions such as "confidential" and "not confidential" which could be informed by personal values. They also allow for easy mapping of policy, process and technology controls based on standard definitions.

Data could be identified, classified, digitally tagged, or watermarked based on these tiers. User entitlement, access, authorization, authentication, logging monitoring, and other data protection controls map easily to this model. Encryption strength and type can be based on data type. For example, commercial cata would ideally require point-to-point encryption and tokenization while collaborative data may not always require encryption especially if already publicly accessible. Data residency controls in many cases will apply only to regulated data given data privacy and national regulations that restrict the flow of data geographically.

Looping CISOs and CSPs
Data classification schemes provide the means by which the CISO can be better looped in with cloud service providers. They enable the business to intelligently champion data security with the CSP without having to defer to a purely compliance-driven approach to security.

In this construct CISOs and InfoSec teams can start to focus on protecting the data, the users, the transactions and workstreams through digital watermarking technologies, encryption, strong user access controls and data residency controls. It is this overarching approach -- rather than having a secure and compliant infrastructure -- that will help retain customer trust, because in the cloud-enabled business, the most control that CISOs can have is over their data and users.

So, what can IT and InfoSec teams do to better protect data and reduce the risk of a data breach in the cloud?

By taking these steps, CISOs and InfoSec teams can ensure more effective collaborations with internal business units, while enabling smarter data security.