Security and operations don't have to be at odds. You can turn security into an operations tool, creating a secure and agile environment that extends your tools and processes.
Statistics and trends support the notion that cloud computing has become a business requirement, rather than an optional capability. However, the adoption of cloud services continues to be inhibited by lack of trust and confidence related to concerns around the security of sensitive data and regulatory issues.
There appear to be two main culprits: the inability of traditional Data Center security tools to adapt to cloud services models and a lack of transparency and control. Both issues affect public, private, and hybrid clouds. The traditional Data Center security environment is static, slow to change, contained (e.g., DMZ), single tenant, and physical. Cloud security environments are dynamic (e.g., live-migration, burst), fast (e.g., automated, scale change), distributed (e.g., hybrid), multi-tenant, and abstracted. On top of these differences, managing external cloud service providers brings coordination costs and risks. Cloud service providers and users must agree on how to create confidence that implementations are trustworthy from top to bottom.
Within these challenges lies an opportunity. We are used to thinking of security as a trade-off -- better security is often considered synonymous with worse performance, agility, and user experience. As a result, there is a constant tug-of-war between operations and security. The opportunity is to turn security into a tool for operations, resulting in an environment that is more secure, more agile, and extends tools and processes.
To do this, instead of thinking in terms of managing security based on physical objects (that is, specific servers), you can substitute logical objects for them. A logical object attribute could be a "trusted server" that's established via measured boot and attestation process characteristics. We can associate other configuration information with the trustworthiness of the server (e.g., its physical location or tenant association). We can now create pools of trusted servers and manage against the pool association of a given server (e.g., deploy pools of workloads of certain types on predefined resource pools = "trusted compute pools"), as opposed to against that server's specific identity. Management applications can consume the trust attributes to execute policy and record event logs. This approach leverages measured boot -- a protection against attacks that target lower levels of the launch environment to form pools of trusted servers that can be used securely for mission-critical workloads.
What action can I take today? If you have an interest in implementing secure, trusted computing environments, either within your own infrastructure or working with cloud providers, the Trusted Computing Group offers resources for implementing standards-based architecture and solution stacks. More information is available from Intel at http://intel.com/TXT.
Tom Quillin is the Director, Cyber Security Technologies and Initiative at Intel Corporation