# Availability Availability is a measure of a system's operational uptime and accessibility to users over a period. It is a key metric in reliability engineering, often expressed as a percentage (e.g., 99.999% or 'five nines'). It quantifies the probability that a system will be functioning and able to deliver its services at any given time. High availability is a primary goal in designing robust and dependable systems, especially for critical services. **Pronunciation:** uh-vey-luh-bil-i-tee **Difficulty:** Beginner **Synonyms:** Uptime, Serviceability, Operational Readiness **Categories:** System Design, Reliability Engineering, Cloud Computing **Tags:** Reliability, Uptime, SLA, Resilience, Fault Tolerance, System Design Canonical: https://scaleengineer.com/glossaries/availability --- ## Definitions - **Core Concept in System Design:** **Availability** is a fundamental quality attribute that measures the percentage of time a system or service is operational and accessible to perform its required function. It is a primary metric for assessing the resilience and dependability of a system. It is mathematically expressed as: `Availability = Uptime / (Uptime + Downtime)` This is often communicated using the 'nines' system, which provides a clear understanding of the expected downtime for a service: * **99% (Two nines):** ~7.3 hours of downtime per month. * **99.9% (Three nines):** ~43.8 minutes of downtime per month. * **99.99% (Four nines):** ~4.38 minutes of downtime per month. * **99.999% (Five nines):** ~26.3 seconds of downtime per month. Achieving higher levels of **availability** involves significant trade-offs. While a five-nines system offers exceptional **uptime**, it requires greater complexity, redundancy, and cost compared to a three-nines system. Factors impacting **availability** include hardware failures, software bugs, network outages, deployment errors, and external dependencies. - **Strategies for High Availability (HA):** High Availability (HA) refers to the architectural principles and practices aimed at ensuring a system achieves a desired level of **operational readiness**. The goal is to minimize or eliminate downtime by building resilience against failures. Key strategies include: * **Redundancy:** This is the core principle of HA. It involves duplicating critical components of the system (e.g., servers, databases, power supplies, network links). If one component fails, a redundant one can take its place, ensuring continuous **serviceability**. * **Failover:** This is the mechanism that automatically switches from a failing component to a redundant one. Failover can be active-passive (a standby component is activated upon failure) or active-active (all components are operational and share the load). * **Fault Detection:** Systems must be able to detect failures promptly to initiate a failover. This is often achieved through health checks or 'heartbeat' mechanisms where components regularly signal their operational status. * **Load Balancing:** Distributes incoming requests across a cluster of redundant servers. This not only improves performance and scalability but also enhances **availability** by routing traffic away from servers that have failed or are unresponsive. ## Etymology Derived from the word 'available', meaning 'capable of being used or obtained'. In a technical context, it refers to the state of a system being ready and accessible for its intended function. ## First used 1960s-1970s ## Historical context The concept of **availability** became a significant concern with the advent of mainframe computers in the 1960s and 1970s. Businesses relied on these systems for critical operations, making their **uptime** essential. A major milestone was the introduction of Tandem Computers' NonStop systems in 1976, which were specifically designed for fault tolerance and high **availability**. The rise of the internet in the 1990s transformed **availability** from a business-centric concern to a public expectation. E-commerce sites and online services needed to be accessible 24/7, making **operational readiness** a competitive advantage. This era saw the popularization of techniques like load balancing and server clustering. With the emergence of cloud computing in the 2000s, high **availability** became a commoditized feature. Cloud providers like AWS, Azure, and Google Cloud Platform built global infrastructure with built-in redundancy, such as Availability Zones and Regions, allowing developers to build highly available systems more easily and cost-effectively than ever before. ## Q&A - **How is availability typically measured and expressed?:** Availability is typically measured as a percentage of uptime over a total period of time. The formula is: `Availability = (Total Time - Downtime) / Total Time * 100%` It is commonly expressed using the 'nines' notation: * **99% (Two Nines):** Allows for about 3.65 days of downtime per year. * **99.9% (Three Nines):** Allows for about 8.77 hours of downtime per year. * **99.99% (Four Nines):** Allows for about 52.6 minutes of downtime per year. * **99.999% (Five Nines):** Allows for about 5.26 minutes of downtime per year. This metric is a critical component of Service Level Agreements (SLAs). - **What is the difference between availability and reliability?:** While related, they measure different aspects of a system's performance. **Availability** is the probability that a system is operational and accessible at a specific point in time. It answers the question, 'Is the system up right now?' A system can have high availability even if it fails frequently, as long as it recovers very quickly. **Reliability** is the probability that a system will perform its intended function without failure for a specified period. It answers the question, 'How long can the system run without failing?' It is often measured by Mean Time Between Failures (MTBF). ## Usage examples - Our Service Level Agreement (SLA) promises customers 99.99% **availability**, which dictates our entire operational strategy. - To increase the application's **availability**, the engineering team implemented a multi-region failover system. - The recent hardware failure severely impacted the database's **uptime**, leading to a breach in our guaranteed **serviceability** for the quarter. ## Related terms - Reliability - Fault Tolerance - Disaster Recovery - Scalability - Redundancy - Mean Time Between Failures (MTBF) - Mean Time To Recovery (MTTR) ## Popular related terms - High Availability (HA) - Service Level Agreement (SLA) - Uptime