Read a sample: 1. A Map of Cloud and Virtualization Security
A company moves a customer portal to a cloud provider. The provider maintains the physical data center, the virtual machine runs a web application, and an administrator uses a cloud console to change network settings. A week later, the portal’s database is exposed to the internet.
Who is responsible?
The useful answer begins with another question: What exactly failed, and who controlled it? The provider might have operated the underlying infrastructure as intended while the customer configured an overly permissive network rule. Alternatively, a failure in a provider-managed component could have contributed to the exposure. The service agreement, configuration, and available evidence matter more than the simple fact that the database was “in the cloud.”
That distinction is the foundation of this book. Cloud and virtualization security is not a contest to memorize the largest number of terms. It is the practice of identifying valuable assets, locating the boundaries that protect them, understanding how those boundaries can fail, and assigning controls to the parties able to operate them.
This chapter gives you a map for doing that work. Use it to orient yourself before exploring individual cloud models, virtual machines, containers, and their supporting controls.
An asset is something worth protecting. In a cloud environment, the obvious example is data: customer records, financial information, source code, or encryption keys. But assets also include the services that process data and the identities that can access them.
Consider a customer portal. Its assets may include:
- The database containing customer records.
- The application and its source code.
- Administrator accounts and service credentials.
- Backups and virtual machine snapshots.
- Logs needed to investigate suspicious activity.
- The portal’s availability and the integrity of its transactions.
The same asset can have several security needs. Customer records require confidentiality so unauthorized people cannot read them. They require integrity so records cannot be altered improperly. The portal requires availability so authorized users can reach it when needed.
These three familiar goals—confidentiality, integrity, and availability—are useful because they make vague statements specific. “Protect the database” is a starting point. “Prevent unauthorized reads, detect unauthorized changes, and restore the database after a failure” points toward controls you can evaluate.
An asset’s location matters, but it does not determine its sensitivity. A backup stored separately from a production database may contain the same confidential information. A machine image may contain embedded credentials. A log intended for troubleshooting may capture sensitive requests. When tracing a workload, look for copies and supporting resources, not just the primary application.
A trust boundary is a point at which the level of trust, control, or authority changes. Data crossing such a boundary may need authentication, authorization, inspection, encryption, or other protection appropriate to the risk.
Cloud and virtualized environments contain several overlapping boundaries:
- Identity boundaries: A user or service requests access to a resource.
- Network boundaries: Traffic moves between the internet, cloud networks, application tiers, or on-premises systems.
- Virtualization boundaries: Workloads share physical resources while being separated by virtualization technology.
- Administrative boundaries: A provider and a customer control different parts of a service.
- Data boundaries: Information moves between systems, accounts, environments, or organizations.
A boundary is not automatically secure because a diagram shows a line around it. A private network, for example, can reduce exposure to the public internet, but an authorized user with excessive permissions may still reach sensitive data. Likewise, strong account authentication does not correct an application that accepts malicious input from an already authenticated user.
For exam scenarios, trace a request in plain language. A browser connects to a public endpoint; the endpoint passes a request to an application; the application accesses a database using a service identity. At each step, ask what permits the connection, what checks the request, and what could happen if a component is compromised.
A threat is a potential cause of harm. A vulnerability is a weakness that a threat could exploit. Risk concerns the possibility and consequences of that harm. A control is a measure intended to prevent, detect, correct, or otherwise reduce risk.
Suppose an attacker is scanning for exposed services. A database listening on a public interface with weak authentication presents weaknesses the attacker may exploit. Restricting network access and strengthening authentication reduce different parts of the risk. Monitoring connection attempts can help detect suspicious activity, while tested backups support recovery if data is damaged.
Controls are often described by what they do:
- Preventive controls attempt to stop an unwanted event, such as an access policy denying an unauthorized request.
- Detective controls identify events or conditions, such as an alert for an unusual privileged login.
- Corrective and recovery controls help restore an acceptable state, such as removing an unauthorized rule or restoring a verified backup.
The categories are useful, but do not let the label replace the reasoning. A backup can help recover data; it does not prevent an attacker from reading an exposed copy. Encryption can reduce the value of stolen data under the right conditions; it does not replace access control, and its effectiveness depends partly on who can use the keys.
A strong answer usually connects a control to a particular failure. If the problem is an unnecessarily public database, restricting the database’s network path is more direct than adding a dashboard. If the problem is stolen administrator credentials, multifactor authentication and tighter privileged access are relevant; a…