technologybriefs
11:03in productionCh. 1 · A term, not a chip/ 11:03 · ceiling 15 min
Systems · Hardware

RAID

RAID doesn’t make storage reliable—it makes unreliability legible, quantifiable, and partially containable.

RAID is a storage architecture that distributes data across multiple physical drives using techniques like striping, mirroring, and parity to improve performance, reliability, or both. It was formally defined in 1987–1988 by Berkeley researchers who demonstrated that arrays of inexpensive drives could surpass expensive single drives in performance while achieving higher reliability through redundancy. Its core principles—striping, mirroring, and parity—are implemented in standardized configurations (RAID levels), with SNIA’s Common RAID DDF serving as the defining specification.

SystemsHardwareAbstractionComputingEngineeringAbstraction across the network →
Chapters & takeaways4
  1. 1:04
    A term, not a chip

    RAID was coined in 1987 by Patterson, Gibson, and Katz—not invented as a product, but proposed as a counterintuitive architectural argument.

  2. 2:50
    Coordination, not copying

    RAID writes to multiple drives at once—not for duplication alone, but to turn independent devices into a single coordinated unit.

  3. 4:48
    Trade-offs baked in

    Striping improves speed; mirroring or parity enables recovery—but each level forces a specific, non-negotiable compromise.

  4. 6:29
    Performance first, redundancy second

    The original claim wasn’t about safety—it was about beating mainframe drives using PC-grade hardware, through architecture, not materials.

Worth your time?

Yes. Study the whole thing.

4/ 5
What works
  • delivers measurable throughput gains via striping
  • enables recovery via mirroring or parity within defined failure boundaries
  • provides documented, level-specific trade-offs
What does not
  • prevent silent corruption
  • guarantee availability during rebuilds
  • scale linearly with drive count
  • eliminate need for backups
Study it if
  • storage administrators
  • infrastructure engineers
  • system architects
Skip it if
  • end users expecting automatic resilience
  • teams without monitoring or operational discipline
The written brief1 min read

What it is and the problem it solves

RAID is a storage architecture that writes data to more than one secondary storage device. It solves the problem of single-drive fragility and performance ceilings by distributing data across arrays using striping, mirroring, or parity.

How it works

RAID coordinates two or more storage devices into a disk array. It distributes data across the array during write operations. It uses data striping for efficiency and disk mirroring or parity drives for recovery. All standard RAID levels except RAID 1 use storage virtualization to abstract multiple devices into one logical volume.

What works

RAID delivers measurable improvements in throughput via striping and recoverability via mirroring or parity—provided the array remains within its design constraints. Its levels provide consistent, documented trade-offs: redundancy, minimum drive count, fault tolerance, and performance.

What does not

RAID does not prevent data loss from silent corruption, controller failure, or operator error. It does not guarantee availability during rebuilds. It does not scale linearly with drive count. It does not eliminate the need for backups.

What it changes

RAID changes how reliability and performance are engineered into storage: by treating multiple physical drives as a single coordinated unit, it decouples fault tolerance from single-drive engineering. It shifts failure mode from catastrophic drive loss to partial degradation with defined recovery paths.

Is it worth your time

Yes—if you manage storage infrastructure and need predictable trade-offs between performance, redundancy, and drive count. No—if you expect plug-and-play resilience or assume modern SSDs eliminate its relevance. RAID does not replace backups, monitoring, or human intervention.

Same field · Systems4 of 177
9:36
Apache ActiveMQ2006ActiveMQ is a Java-based open source message broker implementing JMS 1.1. It enables asynchronous communication across Java and cross-language clients. It supports clustering, multiple persistence options, and several protocols. It originated in 2004 at LogicBlaze, moved to Apache in 2007, and later expanded to include Artemis—a next-generation broker donated in 2015. It is used in enterprise service buses and SOA infrastructure.
10:28
Unmanned surface vehicleUSVs are operational—but not systemic. They deliver real results in niche applications. They lack standardisation, interoperability, and regulatory grounding. Their value lies in removing humans from risk—not in replacing captains with code.
10:32
Self-driving truckSelf-driving trucks are a systems-level adaptation of autonomous technology to freight logistics. They rely on multi-sensor fusion and AI navigation, but their real-world deployment is bounded—not by capability, but by self-imposed safety thresholds and infrastructural control. Kodiak’s December 2024 launch on private lease roads is the first commercial driverless operation in the U.S., yet no autonomous truck has hauled freight without a human on public highways. What works is geofenced, industrial, or military convoy logic—not open-road autonomy.
10:35
China Mobile1997China Mobile is a state-owned telecommunications company formed in 1997. It delivers mobile voice and multimedia services across mainland China and Hong Kong. It is the largest wireless carrier in China, with 945.50 million subscribers as of June 2021. It is listed on the Shanghai and Hong Kong stock exchanges. It is the world's largest mobile network operator by subscriber count and largest telecom company by revenue. Its market value stood at US$240 billion as of 8 July 2025. It originated from the 1999 break-up of China Telecommunications Corporation. In May 2008, it acquired China Tietong to add fixed-line and broadband Internet services.
Up next in Technology

Random-access memory

· 11:08

Random access wasn’t invented to speed up computers—it was built to prove memory could be both volatile and trustworthy enough to hold a program.

Play nextBack to Systems
11:08