Section 50: Acronyms — Overview
Purpose and Scope
This section is the complete acronym reference for the systems knowledge archive, covering 250+ acronyms across all domains of systems software: OS and kernel internals, CPU architecture, networking, security, cloud infrastructure, performance engineering, virtualization, containers, storage, distributed systems, and database internals. Each entry gives the full expansion, a one-sentence definition, and a cross-reference to the section where the concept is treated in depth. Where an acronym has multiple valid expansions in different contexts, all are listed with domain disambiguation.
The catalog serves two audiences: the reader encountering an unfamiliar acronym in technical literature, and the practitioner who wants to verify they are using a term correctly in conversation or writing. Systems literature assumes familiarity with hundreds of acronyms; this catalog makes that assumed knowledge explicit.
Structure
Acronyms are grouped by domain for browsability and listed alphabetically within each group. An alphabetical master index cross-references all entries regardless of domain.
Domain 1: OS and Kernel
| Acronym | Expansion | Definition | Section |
|---|---|---|---|
| ABI | Application Binary Interface | The contract between compiled code and the OS/library it runs on: calling convention, register usage, system call numbers, object file format | 30 |
| APIC | Advanced Programmable Interrupt Controller | Hardware component managing interrupt routing in SMP systems; Local APIC per CPU, I/O APIC for device interrupts | 06 |
| BKL | Big Kernel Lock | The coarse-grained lock that serialized Linux kernel execution; removed in Linux 2.6.39 | 10 |
| COW | Copy-On-Write | Technique deferring data copy until modification; used in fork(), filesystem snapshots, memory deduplication | 07, 13 |
| CFS | Completely Fair Scheduler | Linux's default process scheduler since 2.6.23; uses a red-black tree ordered by virtual runtime | 09 |
| DMA | Direct Memory Access | Hardware mechanism allowing peripherals to read/write main memory without CPU involvement; reduces CPU overhead for I/O | 06 |
| GDT | Global Descriptor Table | x86 table defining memory segments (code, data, TSS) accessible to all processes | 03 |
| IDT | Interrupt Descriptor Table | x86 table of 256 entries mapping interrupt/exception vectors to handler functions | 03 |
| IPC | Inter-Process Communication | Mechanisms for processes to exchange data: pipes, sockets, shared memory, message queues, signals | 07 |
| IRQ | Interrupt Request | Signal from hardware to CPU requesting service; each device is assigned an IRQ line or MSI vector | 06 |
| LDT | Local Descriptor Table | Per-process x86 segment descriptor table; rarely used in modern 64-bit systems | 03 |
| LKM | Loadable Kernel Module | Kernel code that can be loaded and unloaded at runtime without rebooting; used for device drivers, filesystems | 03 |
| LKML | Linux Kernel Mailing List | The primary development mailing list for the Linux kernel; all kernel patches reviewed here | 41 |
| MM | Memory Manager / Memory Management | The kernel subsystem responsible for virtual memory, paging, and physical memory allocation | 11 |
| MMU | Memory Management Unit | Hardware component translating virtual addresses to physical addresses via page table walks | 11 |
| NUMA | Non-Uniform Memory Access | Memory architecture where access latency depends on physical proximity of memory to CPU; important for SMP scalability | 06 |
| OOM | Out Of Memory | Condition where the kernel's OOM killer must terminate processes to free memory; triggered when physical + swap exhausted | 11 |
| PCI | Peripheral Component Interconnect | Hardware bus for connecting peripherals to the CPU; PCIe (PCI Express) is the current standard | 06 |
| PCB | Process Control Block | Kernel data structure holding all information about a process: PID, state, memory maps, open files, signal handlers | 07 |
| PID | Process Identifier | Integer identifier uniquely identifying a process to the kernel within a PID namespace | 07 |
| PTE | Page Table Entry | A single entry in a page table, holding the physical frame number and flags (present, writable, user-accessible, NX) | 11 |
| RAII | Resource Acquisition Is Initialization | C++ / Rust idiom: resource acquisition in constructor, release in destructor; guarantees cleanup on scope exit | 44 |
| RCU | Read-Copy-Update | Synchronization mechanism for read-mostly data; readers never block; writers publish new version atomically | 10 |
| RSS | Resident Set Size | Amount of physical memory currently held by a process (pages in RAM, not swapped out) | 11 |
| SMP | Symmetric Multiprocessing | Architecture where multiple CPUs share a single memory system and OS instance; all CPUs have equal access to memory | 06 |
| TCB | Thread Control Block | Per-thread data structure: thread state, register context, stack pointer; may be part of the PCB or separate | 07 |
| TID | Thread Identifier | Integer identifying a thread within a process or (in Linux) across the system | 07 |
| TLB | Translation Lookaside Buffer | Cache for virtual-to-physical address translations; TLB miss triggers page table walk | 11 |
| TSS | Task State Segment | x86 structure storing CPU state for task switching; in modern OSes mainly used to hold kernel stack pointer for ring transition | 03 |
| UTS | Unix Timesharing System | Namespace type controlling hostname and NIS domain name isolation | 20 |
| VFS | Virtual Filesystem Switch | Linux abstraction layer presenting a uniform filesystem API regardless of underlying filesystem type | 13 |
| VMM | Virtual Memory Manager | Kernel component managing virtual address spaces; also used for Virtual Machine Monitor (see Virtualization) | 11, 19 |
| VSZ | Virtual Set Size | Total virtual address space size of a process, including unmapped and lazily-allocated regions | 11 |
Domain 2: CPU Architecture and Hardware
| Acronym | Expansion | Definition | Section |
|---|---|---|---|
| ALU | Arithmetic Logic Unit | CPU component performing integer arithmetic and bitwise logic operations | 06 |
| ASID | Address Space Identifier | CPU tag attached to TLB entries to avoid full TLB flush on context switch | 11 |
| BMI | Bit Manipulation Instructions | x86 ISA extension for efficient bit manipulation; BMI1, BMI2 | 06 |
| BTB | Branch Target Buffer | CPU structure caching predicted targets of indirect branches; exploited by Spectre variant 2 | 06 |
| CPI | Cycles Per Instruction | Performance metric: IPC = Instructions Per Cycle = 1/CPI; higher IPC = better CPU utilization | 25 |
| CXL | Compute Express Link | PCIe-based coherent interconnect for CPU-to-accelerator and CPU-to-memory attachment | 41 |
| ECC | Error-Correcting Code | Memory technology that detects and corrects single-bit errors and detects double-bit errors | 06 |
| FPGA | Field-Programmable Gate Array | Reconfigurable logic device; used for hardware acceleration and prototyping | 06 |
| FPU | Floating-Point Unit | CPU component for floating-point arithmetic; IEEE 754 operations | 06 |
| HBM | High Bandwidth Memory | Stacked DRAM technology with very high bandwidth; used in GPUs and HPC accelerators | 06 |
| HPC | High-Performance Computing | Cluster computing for scientific simulation; MPI-based parallelism; Infiniband networking | 31 |
| HTT | Hyper-Threading Technology | Intel's Simultaneous Multi-Threading (SMT) implementation; two logical CPUs per physical core | 06 |
| IOMMU | Input-Output Memory Management Unit | Hardware translating device DMA addresses to physical addresses; enables device isolation and passthrough | 06 |
| IPC | Instructions Per Cycle | CPU efficiency metric (also Inter-Process Communication; disambiguate by context) | 25, 07 |
| ISA | Instruction Set Architecture | The programmer-visible CPU interface: instructions, registers, memory model, calling convention | 06 |
| LLC | Last-Level Cache | The largest, slowest on-die cache (typically L3); shared across cores; miss goes to DRAM | 06 |
| MCE | Machine Check Exception | CPU error reporting mechanism for hardware faults: ECC errors, bus errors, TLB corruption | 06 |
| MSR | Model-Specific Register | CPU register accessible only in kernel mode; used for microcode updates, performance counters, feature control | 06 |
| NVMe | Non-Volatile Memory Express | PCIe-attached SSD protocol; very low latency (sub-100 microsecond); replaces SATA/SAS for high-performance storage | 12 |
| PCIe | PCI Express | Current generation PCI bus; serial lanes; PCIe 5.0 at 32 GT/s per lane | 06 |
| PMU | Performance Monitoring Unit | CPU hardware for counting hardware events: instructions, cache misses, branch mispredictions | 25 |
| RDMA | Remote Direct Memory Access | Network technology allowing direct memory access between hosts without CPU involvement; Infiniband, RoCE | 15 |
| SMT | Simultaneous Multi-Threading | CPU technique running two hardware threads on one physical core; shares execution units | 06 |
| SIMD | Single Instruction, Multiple Data | CPU instruction class operating on vectors: SSE, AVX, AVX-512, NEON | 06 |
| SSD | Solid-State Drive | Flash-based storage device; NAND flash cells; FTL manages wear leveling and bad block management | 12 |
| TDP | Thermal Design Power | Maximum sustained power dissipation a CPU cooling solution must handle | 06 |
| UEFI | Unified Extensible Firmware Interface | Replacement for BIOS; provides pre-OS environment, secure boot, firmware services | 05 |
Domain 3: Memory Management
| Acronym | Expansion | Definition | Section |
|---|---|---|---|
| ASLR | Address Space Layout Randomization | Security mitigation randomizing base addresses of stack, heap, libraries; defeats hardcoded address exploits | 26 |
| DAX | Direct Access | Filesystem mode bypassing the page cache for PMEM devices; CPU accesses persistent memory directly | 41 |
| GC | Garbage Collector | Automatic memory reclamation; tracing GC (reachability), reference counting, generational; not used in kernel | 29 |
| HMA | Heterogeneous Memory Architecture | Linux subsystem managing PMEM, HBM, and CXL memory tiers with different performance characteristics | 11 |
| KPTI | Kernel Page Table Isolation | Meltdown mitigation: separate page tables for user/kernel mode; eliminates kernel mapping from user-mode TLB | 26 |
| MMIO | Memory-Mapped I/O | Technique mapping device registers into the CPU's physical address space; accessed with normal load/store | 06 |
| OOM | Out Of Memory | See Domain 1 | 11 |
| PMEM | Persistent Memory | Byte-addressable non-volatile memory (Intel Optane); persistence survives power loss; DAX mode | 41 |
| THP | Transparent Huge Pages | Linux kernel feature automatically backing eligible anonymous memory with 2MB huge pages | 11 |
| UAF | Use-After-Free | Memory safety bug: accessing memory after it has been freed; a primary source of kernel exploits | 26, 44 |
Domain 4: Networking
| Acronym | Expansion | Definition | Section |
|---|---|---|---|
| ARP | Address Resolution Protocol | Maps IPv4 addresses to MAC addresses on a local network segment | 15 |
| BGP | Border Gateway Protocol | Inter-autonomous system routing protocol; path vector; the routing protocol of the internet | 15 |
| CIDR | Classless Inter-Domain Routing | IPv4/IPv6 addressing with variable-length subnet masks (e.g., /24); replaced classful addressing | 15 |
| DPDK | Data Plane Development Kit | User-space framework for high-performance packet processing; PMD bypasses kernel networking stack | 41 |
| DSCP | Differentiated Services Code Point | IPv4/IPv6 header field for QoS marking; used in traffic prioritization | 15 |
| GRE | Generic Routing Encapsulation | Tunneling protocol encapsulating arbitrary network packets inside IP | 15 |
| GRO | Generic Receive Offload | Kernel networking optimization coalescing multiple received packets into a single skb | 15 |
| GSO | Generic Segmentation Offload | Kernel networking optimization deferring TCP segmentation to the NIC | 15 |
| ICMP | Internet Control Message Protocol | Network layer diagnostic and error reporting protocol; ping uses ICMP echo request/reply | 15 |
| LACP | Link Aggregation Control Protocol | IEEE 802.3ad protocol for negotiating bonded network link groups | 15 |
| MAC | Media Access Control | Layer 2 hardware address (48-bit); unique per network interface | 15 |
| MTU | Maximum Transmission Unit | Largest IP packet size a network path can carry without fragmentation; typically 1500 bytes on Ethernet | 15 |
| NAT | Network Address Translation | Mapping private IP addresses to public IPs; hides internal topology; stateful connection tracking | 15 |
| NIC | Network Interface Card | Hardware device providing physical network connectivity and packet send/receive | 15 |
| OSPF | Open Shortest Path First | Link-state interior gateway routing protocol; Dijkstra's algorithm on link-state database | 15 |
| RSS | Receive Side Scaling | NIC feature distributing incoming packets across multiple CPU cores via hash-based steering | 15 |
| SCTP | Stream Control Transmission Protocol | Transport protocol combining TCP reliability with UDP message boundaries; multi-homing support | 15 |
| SPDK | Storage Performance Development Kit | User-space NVMe driver framework; analogous to DPDK for storage; sub-100 microsecond latency | 41 |
| SR-IOV | Single Root I/O Virtualization | PCIe feature enabling physical NIC to present multiple virtual functions (VFs) directly to VMs | 19 |
| TCP | Transmission Control Protocol | Reliable, ordered, connection-oriented transport protocol; stream semantics; congestion control | 16 |
| TLS | Transport Layer Security | Cryptographic protocol for authenticated and encrypted communication; successor to SSL | 26 |
| UDP | User Datagram Protocol | Unreliable, connectionless transport protocol; no flow control; used for latency-sensitive applications | 15 |
| VLAN | Virtual Local Area Network | IEEE 802.1Q standard for logical network segmentation within a physical network | 15 |
| VPC | Virtual Private Cloud | Cloud provider's isolated virtual network; VXLAN or similar overlay | 21 |
| VTEP | VXLAN Tunnel Endpoint | Device terminating a VXLAN tunnel; performs encap/decap of VXLAN-wrapped Ethernet frames | 21 |
| VXLAN | Virtual Extensible LAN | Layer 2 overlay over Layer 3 networks; 24-bit VNI allows 16M virtual networks | 21 |
| XDP | eXpress Data Path | eBPF hook in the NIC driver receive path; processes packets before skb allocation; near line-rate | 41 |
Domain 5: Security
| Acronym | Expansion | Definition | Section |
|---|---|---|---|
| ACL | Access Control List | List of permissions attached to an object specifying which subjects can perform which operations | 26 |
| ASLR | Address Space Layout Randomization | See Domain 3 | 26 |
| CFI | Control Flow Integrity | Mitigation restricting indirect control flow to valid targets determined at compile time | 26 |
| CVE | Common Vulnerabilities and Exposures | MITRE-managed database of publicly known security vulnerabilities; each assigned a CVE-YYYY-NNNNN identifier | 26 |
| DAC | Discretionary Access Control | Access control where resource owners set permissions; traditional Unix permissions model | 26 |
| DEP | Data Execution Prevention | Windows term for W^X / NX: marking data pages non-executable | 26 |
| EPID | Enhanced Privacy ID | Intel anonymous attestation scheme for SGX; proof of genuine Intel hardware | 26 |
| HSM | Hardware Security Module | Physical device for cryptographic key storage and operations; tamper-resistant | 26 |
| IDS | Intrusion Detection System | System monitoring for signs of attack; NIDS (network), HIDS (host-based) | 26 |
| JOP | Jump-Oriented Programming | Variant of ROP using indirect jump gadgets instead of returns | 26 |
| KASLR | Kernel Address Space Layout Randomization | Randomizes kernel base address at boot; mitigates kernel exploit hardcoded addresses | 26 |
| LSM | Linux Security Module | Framework allowing kernel security policy enforcement; SELinux, AppArmor, TOMOYO built on LSM | 26 |
| MAC | Mandatory Access Control | Security model where a central authority defines access policy; SELinux, AppArmor | 26 |
| NX | No-Execute | CPU page table flag marking a page non-executable; prevents shellcode execution on the stack/heap | 26 |
| PAC | Pointer Authentication Code | ARM hardware feature embedding a cryptographic signature in unused pointer bits; mitigates ROP | 26 |
| RBAC | Role-Based Access Control | Access control model assigning permissions to roles, roles to users | 26 |
| ROP | Return-Oriented Programming | Code reuse attack chaining existing instruction sequences (gadgets) ending in RET | 26 |
| SMEP | Supervisor Mode Execution Prevention | CPU feature preventing kernel from executing user-space pages; blocks ret2user | 26 |
| SMAP | Supervisor Mode Access Prevention | CPU feature preventing kernel from accessing user-space memory without explicit annotation | 26 |
| TEE | Trusted Execution Environment | Isolated secure computing environment: ARM TrustZone, Intel SGX, AMD SEV | 26 |
| TPM | Trusted Platform Module | Hardware chip storing cryptographic keys and performing secure boot measurement; TCG standard | 26 |
| SGX | Software Guard Extensions | Intel hardware feature creating encrypted enclaves isolated from OS and hypervisor | 26 |
Domain 6: Distributed Systems
| Acronym | Expansion | Definition | Section |
|---|---|---|---|
| CAP | Consistency, Availability, Partition tolerance | Brewer's theorem: distributed systems can satisfy at most two of these three properties simultaneously | 17 |
| CRDT | Conflict-free Replicated Data Type | Data structure that can be replicated across nodes and merged without conflicts; enables strong eventual consistency | 17 |
| DCS | Distributed Control System | Industrial control system with distributed computing nodes; not to be confused with the CS usage | 17 |
| FLP | Fischer-Lynch-Paterson | FLP impossibility result: no deterministic consensus algorithm can tolerate even one crash in an asynchronous system | 17 |
| HLC | Hybrid Logical Clock | Combines physical clock and logical clock to provide both causality tracking and approximate real-time | 17 |
| NTP | Network Time Protocol | Protocol for synchronizing clocks across networks; not sufficient for distributed transactions | 17 |
| PTP | Precision Time Protocol | IEEE 1588; microsecond-accurate hardware-timestamped clock synchronization | 17 |
| RPO | Recovery Point Objective | Maximum acceptable data loss in time; defines required backup/replication frequency | 28 |
| RTO | Recovery Time Objective | Maximum acceptable downtime; defines required failover speed | 28 |
| SLA | Service Level Agreement | Contractual uptime/performance commitment; typically expressed as nines (99.9%, 99.99%) | 28 |
| SLI | Service Level Indicator | Quantitative measure of service behavior: latency P99, error rate, availability | 28 |
| SLO | Service Level Objective | Target value for an SLI; internal performance target, narrower than SLA | 28 |
| WAL | Write-Ahead Log | See Domain 8; also applies to distributed logs (Kafka) | 13, 18 |
| ZAB | ZooKeeper Atomic Broadcast | ZooKeeper's consensus protocol; similar to Paxos; primary-backup ordering | 17 |
Domain 7: Performance Engineering
| Acronym | Expansion | Definition | Section |
|---|---|---|---|
| BPF | Berkeley Packet Filter | In-kernel virtual machine for packet filtering; extended to eBPF for general-purpose use | 15, 41 |
| eBPF | Extended Berkeley Packet Filter | General-purpose in-kernel VM; sandboxed, JIT-compiled programs; hooks throughout kernel | 41 |
| EDP | Energy-Delay Product | Performance metric combining energy and execution time; used in power-constrained design | 25 |
| FLOPS | Floating-Point Operations Per Second | Measure of floating-point computation throughput | 31 |
| IPC | See Domain 1 and Domain 2 | Disambiguate by context | 07, 25 |
| IOPS | Input/Output Operations Per Second | Storage throughput metric counting discrete I/O operations | 12 |
| LLC | Last-Level Cache | See Domain 2 | 06 |
| P50/P95/P99 | 50th/95th/99th Percentile Latency | Latency distribution metrics; P99 captures tail latency affecting worst-case user experience | 25 |
| PMU | Performance Monitoring Unit | See Domain 2 | 25 |
| QPS | Queries Per Second | Throughput metric for request-serving systems | 25 |
| RDMA | Remote Direct Memory Access | See Domain 4 | 15 |
| TSC | Time Stamp Counter | x86 64-bit counter incremented at CPU frequency; used for high-resolution timing | 25 |
| TTFB | Time To First Byte | Web performance metric measuring server response latency | 25 |
Domain 8: Virtualization and Containers
| Acronym | Expansion | Definition | Section |
|---|---|---|---|
| EPT | Extended Page Tables | Intel VT-x hardware for second-level address translation (guest virtual → guest physical → host physical) | 19 |
| GPA | Guest Physical Address | Address as seen by the guest OS; translated to HPA by EPT/NPT | 19 |
| HPA | Host Physical Address | Actual physical address on the host system | 19 |
| KVM | Kernel-based Virtual Machine | Linux kernel module implementing hypervisor using hardware virtualization (VT-x, AMD-V) | 19 |
| NPT | Nested Page Tables | AMD-V hardware for second-level address translation; equivalent to Intel EPT | 19 |
| OCI | Open Container Initiative | Standard defining container image format and runtime interface; Docker, containerd, podman comply | 20 |
| QEMU | Quick EMUlator | Machine emulator and virtualizer; used with KVM for hardware-accelerated virtualization | 19 |
| VCPU | Virtual CPU | Software abstraction of a CPU presented to a guest VM; scheduled on physical CPUs by the VMM | 19 |
| VFIO | Virtual Function I/O | Linux kernel framework for device passthrough to VMs; safe user-space DMA via IOMMU | 19 |
| VMCS | Virtual Machine Control Structure | Intel data structure controlling VM behavior: entry/exit conditions, control register values | 19 |
| VMX | Virtual Machine Extensions | Intel's hardware virtualization instructions; VMXON, VMLAUNCH, VMEXIT, VMREAD, VMWRITE | 19 |
Domain 9: Cloud and Orchestration
| Acronym | Expansion | Definition | Section |
|---|---|---|---|
| AWS | Amazon Web Services | Amazon's public cloud platform | 21 |
| AZ | Availability Zone | Physically isolated datacenter within a cloud region; fault isolation boundary | 21 |
| CDN | Content Delivery Network | Geographically distributed cache network; serves static content from edge nodes near users | 21 |
| CRD | Custom Resource Definition | Kubernetes mechanism for extending the API with new resource types | 22 |
| CSI | Container Storage Interface | Kubernetes standard API for storage plugins | 22 |
| CNI | Container Network Interface | Standard API for Kubernetes network plugins; Cilium, Flannel, Calico | 22 |
| ECS | Elastic Container Service | AWS managed container orchestration service | 21 |
| EKS | Elastic Kubernetes Service | AWS managed Kubernetes service | 22 |
| GKE | Google Kubernetes Engine | Google Cloud managed Kubernetes service | 22 |
| IAM | Identity and Access Management | Cloud access control system; maps identities to permitted API operations | 21 |
| IaC | Infrastructure as Code | Managing infrastructure through declarative configuration files; Terraform, Pulumi | 21 |
| OPA | Open Policy Agent | General-purpose policy engine for cloud-native authorization | 26 |
| RBAC | Role-Based Access Control | See Security domain; also used in Kubernetes for API access control | 26 |
| S3 | Simple Storage Service | AWS object storage service; REST API; durability through replication | 21 |
| VPC | Virtual Private Cloud | See Networking domain | 21 |
Alphabetical Master Index (Sampled)
A complete alphabetical index spanning all domains is maintained in 50-acronyms/01-master-index.md. Example entries:
- ASLR → Security (also Memory Management)
- BPF → Performance / Networking
- CAP → Distributed Systems
- CFS → OS/Kernel
- CXL → CPU Architecture / Modern Kernel
- DMA → CPU Architecture
- eBPF → Performance / Modern Kernel
- EPT → Virtualization
- FPGA → CPU Architecture
- IPC → OS/Kernel (also Performance — see disambiguation note)
- KPTI → Security / Memory Management
- LLC → CPU Architecture / Performance
- MAC → Networking (layer 2) and Security — always disambiguate
- MMU → OS/Kernel / Memory Management
- NIC → Networking
- NUMA → CPU Architecture
- OCI → Containers
- PCIe → CPU Architecture
- RDMA → Networking / Performance
- RCU → OS/Kernel
- SGX → Security
- SIMD → CPU Architecture
- TLB → Memory Management
- VMX → Virtualization
- WAL → Storage / Database
- XDP → Networking / Performance
File Map
50-acronyms/
├── 00-overview.md ← This file
├── 01-master-index.md (alphabetical, all 250+ entries)
├── 02-os-kernel-acronyms.md
├── 03-cpu-hardware-acronyms.md
├── 04-memory-acronyms.md
├── 05-networking-acronyms.md
├── 06-security-acronyms.md
├── 07-distributed-systems-acronyms.md
├── 08-performance-acronyms.md
├── 09-virtualization-container-acronyms.md
└── 10-cloud-orchestration-acronyms.md
Cross-References
- Section 49 (Glossary): full definitions for the terms these acronyms abbreviate
- All Sections 00-48: the source sections where each concept is explained in depth
- Section 45 (Learning Roadmaps): acronym fluency is a prerequisite for reading primary literature in each track