Enterprise Cloud Computing
Mastering the Balancing Act of #Cloud Security and Business Agility
There are three clear control capability areas needed for IT to effectively manage financial, reputation and legal risk
Sep. 3, 2014 12:00 PM
In 2012, an IDG survey of enterprise cloud computing adoption showed that 70 percent of respondents said security was among their top three concerns, and two years later, not much has changed. The Everest Group Enterprise Cloud Adoption Survey released in March of 2014 shows that 70 percent of enterprises prefer private cloud because it offers higher security - a clear indication that security concerns still weigh heavily on the minds of enterprise leaders. Centralizing cloud resource access could prove to be the path through, addressing security concerns while providing the agility cloud computing promises.
It is understandable how cloud security presents itself as a chief IT concern when you consider that cloud computing transfers control from IT to business users and developers. And that adopting cloud entails replacing numerous IT processes with self-service portals.
While there are innumerous benefits to adopting cloud computing, transferring control away from IT does open the business to risk as it diminishes IT's ability to protect the organization's resources and data against unauthorized access and misuse. It also ties IT's hands when it comes to identifying and resolving security issues, and enforcing compliance with industry regulations. These are critical functions that have direct impact on business risk.
Addressing Business Risk via Security Controls
Cloud computing transforms the way infrastructure is provisioned in an organization. It replaces the centralized IT-controlled infrastructure provisioning model where developers make an infrastructure request that IT reviews and then fulfills, with a new, distributed developer-centric infrastructure provisioning process where developers effectively bypass IT. As a result, enterprises adopting cloud find themselves in a paradoxical situation where IT is responsible for the infrastructure security that developers now control.
There are three clear control capability areas needed for IT to effectively manage financial, reputation and legal risk.
- Preventive capabilities: IT must be able to prevent insecure provisioning requests from being fulfilled, on both a per-user-role and per-environment basis. For example, IT must be able to enforce specific firewall policies for production infrastructure.
In order to satisfy developer requirements, it is obvious that IT cannot change the way cloud infrastructure is accessed: provisioning must remain self-service. As a result, IT needs transparent and automated policy enforcement. Provisioning requests made to the organization's cloud need to be inspected in real-time and checked against governance policies that are in place. When approved, requests must be forwarded to the relevant cloud API; when denied, the developer that made the request must be immediately informed. Ideally, the developer should be provided with an explanation and an alternate course of action should be suggested.
- Detective capabilities: IT must have a centralized view of infrastructure to identify vulnerabilities and intrusions; IT must be able to understand the purpose of every resource provisioned by the business. For example, IT must be able to identify the configuration of every deployed resource and the environment to which it belongs. That knowledge can then be used to decide whether an unusual firewall configuration or activity pattern should trigger an alert.
To satisfy these requirements, IT needs a federated view and understanding of all of the business's cloud resources, ensuring visibility over the organization's cloud resources. To do so, IT must ensure that every provisioning request is associated with a legitimate owner and use case (ideally in an automated fashion); that all provisioned resources remain visible throughout their lifecycle; and that metadata regarding their purpose remains accessible.
- Corrective capabilities: IT must control access to the business's cloud infrastructure. For example, IT must be able to revoke access for employees that leave the company, and be able to centrally identify and patch affected resources when a vulnerability is identified.
To do so, IT needs centralized credential management to govern access to cloud resources. IT must ensure that access to cloud resources is controlled by the organization's existing identity management infrastructure, and not by ad-hoc SSH keys or RDP passwords created by developers. Naturally, in order to not hinder developer productivity, IT must ensure that developers can still access the resources for which they have a legitimate use.
Where the Rubber Meets the Road
Cloud security has been an issue since 2006 when cloud emerged with the release of AWS EC2. Back then, all cloud instances were exposed to the Internet, and access was only available with root keys. To address these respective problems, Amazon announced AWS Virtual Private Cloud and AWS Identity and Access Management. Some AWS competitors have also issued access control management, though they remain somewhat limited. Yet, these controls only address IT's preventive needs, are only available on AWS as of this writing, and are often complex to use.
As a result, IT is frequently opting to deploy a cloud management platform (CMP), an often on-premise, web-based application, that sits between end-users and the multiple cloud platforms that they may use. CMPs are extensible platforms that let IT departments customize the CMP's behavior to fit their organization's workflows and policies. In turn, CMPs enforce those IT policies in a fully transparent and automated fashion, so that developers aren't slowed down by red tape when getting work done. As a result, CMPs ensure that IT is provided the security capabilities it requires, while ensuring developers retain the agility they need.
Most importantly, CMPs play a critical role in addressing all three control capability areas:
- Preventive: CMPs can provide IT with governance and role-based access control capabilities, and empower IT to secure and control access to cloud resources on a per-user or per-user-group basis. Using a CMP, these policies can be enforced in real-time, so that developers are not slowed by their enforcement. IT can, for example, ensure that specific firewall rules are automatically added for every single instance that is launched, and that instances are automatically launched in secure networks (e.g. a specific AWS Virtual Private Cloud, or VPC).
- Detective: Because CMPs are used for the provisioning of all the organization's resources, they may automatically keep a precise account of the resources that were provisioned, by whom, and for what purpose. As a result, resource tracking can be performed automatically, and developers won't have to perform extra effort to comply with IT policies.
- Corrective: CMPs may centralize the creation and use of CMP-controlled credentials and make those available to dev and IT, or automatically configure cloud resources to leverage the company's existing identity management framework instead. For example, with a CMP, IT can enforce developer use of Active Directory credentials to login to their instances.
While cloud momentum continues to grow, so does concern - rightfully so - for cloud security. While IaaS providers have taken steps to address these concerns within their systems, they do not currently address the spectrum of capabilities needed to fully address business risk. CMPs are an effective option that can be deployed in a way that addresses IT, business, and developer needs.