Governance and risk management principles remain the same, but there are changes to the risk picture as well as available controls in the cloud. We need in particular take into account the following:
- Cloud risk trade-offs and tools
- Effects of service and deployment models
- Risk management in the cloud
- Tools of cloud governance
A key aspect to remember when deploying services or data to the cloud is that even if security controls are delegated to a third-party, the responsibility for corporate governance cannot be delegated; it remains within the cloud consumer organization.
Cloud providers aim to streamline and standardize their offerings as much as possible to achieve economies of scale. This is different from a dedicated third-party provider where contractual terms can often be negotiated. This means that governance frameworks should not treat cloud providers with the same approach as those dedicated service providers allowing for custom governance structures to be agreed on.
Responsibilities and mechanisms for governance is regulated in the contract. If a governance need is not described in the contract, there exists a governance gap. This does not mean that the provider should be excluded directly, but it does mean that the consumer should consider how that governance gap can be closed.
Cloud governance tools
The key tools of governance in the cloud are contracts, assessments and reporting.
Contracts are the primary tools for extending governance into a third party such as a cloud provider. For public clouds this would typically mean the terms and conditions of the provider. They are the guarantee of a given service level, and also describes requirements for governance support through audits.
Supplier assessments are important as governance tools, especially during provider selection. Performing regular assessments can discover if changes to the offerings of the cloud provider has changed the governance situation, in particular with regard to any governance gaps.
Compliance reporting includes audit reports. They may also include automatically generated compliance data in a dashboard, such as patch level status on software, or some other defined KPI. Audit reports may be internal reports but most often these are made by an accredited third party. Common compliance frameworks are provided by ISO 27017, ISO 38500, COBIT.
Risk management
Enterprise risk management (ERM) in the cloud is based on the shared responsibility model. The provider will take responsibility for certain risk controls, whereas the consumer is responsible for others. Where the split is depends on the service model.
The division of responsibilities should be clearly regulated in the contract. Lack of such regulation can lead to hidden implementation gaps, leaving services vulnerable to abuse.
Service models
IaaS mostly resembles traditional IT as most controls remain under direct management of the cloud consumer. Thus, policies and controls do to a large degree remain under control of the cloud consumer too. There is one primary change and that is the orchestration/management plane. Managing the risk of the management plane becomes a core governance and risk management activity – basically moving responsibilities from on-prem activities to the management plane.
SaaS providers vary greatly in competence and the tools offered for compliance management. It is often possible to negotiate custom contracts with smaller SaaS providers, whereas the more mature or bigger players will have more standardized contracts but also more tools appropriate to governance needs of the enterprise. The SaaS model can be less transparent than desired, and establishing an acceptable contract is important in order to have good control over governance and risk management.
Public cloud providers often allow for less negotiation than private cloud. Hybrid and community governance can easily become complicated because the opinions of several parties will have to be weighed against each other.
Risk trade-offs
Using cloud services will typically result in more trust put in third-parties and less direct access to security controls. Whether this increases or decreases the overall risk level depends on the threat model, as well as political risk.
The key issue is that governance is changed from internal policy and auditing to contracts and audit reports; it is a less hands-on approach and can result in lower transparency and trust in the governance model.
CSA recommendations
- Identify the shared responsibilities. Use accepted standards to build a cloud governance framework.
- Understand and manage how contracts affect risk and governance. Consider alternative controls if a contract leaves governance gaps and cannot be changed.
- Develop a process with criteria for provider selection. Re-assessments should be regular, and preferably automated.
- Align risks to risk tolerances per asset as different assets may have different tolerance levels.
#2cents
Let us start with the contract side: most cloud deployments will be in a public cloud, and our ability to negotiate custom contracts will be very limited, or non-existing. What we will have to play with is the control options in the management plane.
The first thing we should perhaps take note of, is not really cloud related. We need to have a regulatory compliance matrix in order to make sure our governance framework and risk management processes actually will help us achieve compliance and acceptable risk levels. One practical way to set up a regulatory compliance matrix is to map applicable regulations and governacne requirements to the governance tools we have at our disposal to see if the tools can help achieve compliance.
| Regulatory source | Contractual impact | Supplier assessments | Audits | Configuration management |
| GDPR | Data processing agreement | Security requirements GDPR compliance | Data processing acitvities audits | Data retention Backups Discoverability Encryption |
| Customer SLA | SLA guarantees |
| Uptime reporting |
|
| ISO 27001 |
| Certifications | Audit reports for certifications | Extension of company policies to management plane |
Based on the regulatory compliance matrix, a more detailed governance matrix can be developed based on applicable guidance. Then governance and risk management gaps can be identified, and closing plans created.
Traditionally cloud deployments have been seen as higher risk than on-premise deployments due to less hands-on risk controls. For many organizations the use of cloud services with proper monitoring will lead to better security because many organizations have insufficient security controls and logging in their on-premise tools. There are thus situations where a shift from hands-on to contractual controls is a good thing for security. One could probably claim that this is the case for most cloud consumers.
One aspect that is critical to security is planning of incident response. To some degree the ability to do incidence response on cloud deployments depends on configurations set in the management plane; especially the use of logging and alerting functionality. It should also be clarified up front where the shared responsibility model puts the responsibility for performing incident response actions throughout all phases (preparation, identification, containment, eradication, recovery and lessons learned).
The best way to take cloud into account in risk management and governance is to make sure policies, procedures and standards cover cloud, and that cloud is not seen as an “add-on” to on-premise services. Only integrated governance systems will achieve transparency and managed regulatory compliance.
safecontrols 