3.4 Resilience & Recovery

High availability (HA) refers to the implementation of system designs and configurations that ensure services remain accessible to authorized users even during hardware failures or spikes in demand.

Availability within the CIA Triad The goal of availability is to bridge the gap between security and functionality, ensuring that systems are both protected and usable. - Inversely Proportional: As security increases, functionality or ease of access often decreases; HA seeks to balance these alongside finite resources. - Redundancy: The primary method for achieving HA by eliminating Single Points of Failure (SPOF). - Resilience: The ability of a system to maintain operational status during an adverse event (e.g., hardware failure or a DDoS attack).

Load Balancing Load balancers distribute incoming network or application traffic across multiple servers to ensure no single device is overwhelmed. - Active-Active: All nodes in the cluster are simultaneously processing requests. This maximizes throughput and utilizes all available resources. - Active-Passive: One node handles all traffic while another remains in standby mode. The standby node only takes over if the primary node fails (failover). - Scheduling Algorithms: Methods used to distribute traffic, such as Round Robin (sequential distribution) or Least Connections (sending traffic to the least busy server). - Virtual IP (VIP): A single IP address presented to the client that represents the entire load-balanced farm.

Clustering Clustering connects multiple independent systems to work as a single unit, providing higher levels of availability and scalability than a single machine. - Failover Clusters: Designed strictly for redundancy; if one node fails, the service transitions to another node with minimal downtime. - Scalability: Clusters allow for Horizontal Scaling (adding more nodes) rather than just Vertical Scaling (upgrading hardware on one machine). - State Management: Clusters often share storage or state information to ensure user sessions are not lost during a transition between nodes.

Quantitative vs. Qualitative Risk in HA When deciding on HA investments, organizations use risk assessments: - Quantitative: Using numerical data (MTBF, MTTR) to calculate the cost of downtime against the cost of redundancy. - Qualitative: Assessing the impact of downtime based on reputation or "High/Medium/Low" severity scales when hard data is unavailable.

Quick recall - SPOF: Single Point of Failure; the main target HA seeks to eliminate. - Failover: The automatic switch to a redundant or standby computer, server, system, or network upon the failure of the previous application. - Load Balancer: A device that acts as a reverse proxy to distribute network traffic. - Redundancy: Having multiple "spare tires" for your IT infrastructure. - Active-Active: All systems working; highest performance. - Active-Passive: One system working, one waiting; saves on licensing or power but sacrifices performance capacity.