Skip to main content
FutureStack
Engineering Encyclopedia

AI & Systems Architecture Glossary

Precise, verified technical definitions written for developers and architects. No marketing fluff—just how things actually work under the hood.

Category:
Showing 18 of 18 verified terms
RAG (Retrieval-Augmented Generation) technical architecture diagram
AI & MLRead Deep Dive

RAG (Retrieval-Augmented Generation)

An architectural technique that enhances LLM responses by dynamically querying an external vector or keyword database for relevant context before generating an answer.

Related:#vector-embeddings#quantization#context-window
Quantization technical architecture diagram
AI & MLRead Deep Dive

Quantization

The mathematical process of reducing the precision of neural network weights (e.g., from 16-bit floating point to 8-bit or 4-bit integers) to shrink memory requirements and accelerate inference speed.

Related:#rag-retrieval-augmented-generation#speculative-decoding#context-window
Context Window technical architecture diagram
AI & MLRead Deep Dive

Context Window

The maximum number of tokens (words, characters, and punctuation) that a large language model can accept and reason across in a single conversational turn or API request.

Related:#rag-retrieval-augmented-generation#vector-embeddings#agentic-loop
Vector Embeddings technical architecture diagram
AI & MLRead Deep Dive

Vector Embeddings

High-dimensional mathematical array representations of text, audio, or images where semantic similarity translates directly to geometric proximity in vector space.

Related:#rag-retrieval-augmented-generation#context-window
Agentic Loop (ReAct Pattern) technical architecture diagram
AI & MLRead Deep Dive

Agentic Loop (ReAct Pattern)

An autonomous software cycle where an AI model reasons through a plan, selects and executes programmatic tools, inspects the tool output, and iterates until the objective is accomplished.

Related:#rag-retrieval-augmented-generation#context-window#speculative-decoding
Speculative Decoding technical architecture diagram
AI & MLRead Deep Dive

Speculative Decoding

An inference optimization where a small, fast 'draft' model rapidly predicts multiple tokens ahead, and a larger target model verifies them in parallel in a single forward pass.

Related:#quantization#context-window
React Server Components (RSC) technical architecture diagram
Web ArchitectureRead Deep Dive

React Server Components (RSC)

A React paradigm where components execute exclusively on the server, streaming lightweight serialized HTML to the client without sending any JavaScript bundle weight.

Related:#edge-computing#vector-embeddings
Edge Computing & Anycast technical architecture diagram
Cloud & DevOpsRead Deep Dive

Edge Computing & Anycast

A distributed cloud architecture that routes user requests to the nearest physical datacenter in a global network (e.g. Cloudflare, AWS CloudFront) to execute code within milliseconds of the user.

Related:#server-components-rsc#quantization
Non-contiguous virtual KV-cache memory blocks inside GPU VRAM mapped via glowing amber block tables
AI & ML11.4k words

PagedAttention & KV-Cache Memory Management

An LLM serving memory architecture inspired by OS virtual memory paging that stores attention Key and Value tensors in fixed-size, non-contiguous physical GPU VRAM blocks, eliminating up to 60%+ internal and external fragmentation.

Related:#quantization#speculative-decoding#context-window
Sparse gating router dispatching token vectors dynamically to top-k crystalline neural experts
AI & ML13.4k words

Sparse Mixture-of-Experts (MoE) & Auxiliary-Loss-Free Routing

A sparse neural architecture replacing dense Feed-Forward layers with multiple specialized sub-networks ('experts') where a lightweight gating router activates only the top-k experts per token, decoupling parameter knowledge capacity from per-token inference FLOPs.

Related:#paged-attention-kv-cache#quantization#speculative-decoding
Standardized MCP JSON-RPC bus connecting an autonomous core agent to sandboxed tool sockets and schema capsules
AI & ML11.7k words

Model Context Protocol (MCP)

An open JSON-RPC 2.0 client-host-server specification defining a universal, capability-scoped interface for AI models to discover tools, subscribe to structured data resources, and execute auditable external actions without bespoke glue integrations.

Related:#agentic-loop#rag-retrieval-augmented-generation#vector-embeddings
NIC RX ring packet intercepted cleanly at XDP driver hook inside Linux kernel strata with eBPF verifier gates
Cloud & DevOps10.4k words

eBPF & XDP (eXpress Data Path)

An in-kernel virtual machine and safety verifier permitting sandboxed bytecode programs to attach directly to Linux NIC drivers (XDP), syscalls, and LSM hooks—enabling multi-million-packet-per-second DDoS drops and zero-overhead distributed observability.

Related:#io-uring-zero-copy-io#edge-computing#post-quantum-cryptography-ml-kem
Concentric Submission Queue and Completion Queue ring buffers bridging userspace and Linux kernel via DMA
Cloud & DevOps12.2k words

Linux io_uring & Zero-Copy Ring Buffers

A Linux kernel asynchronous I/O interface built around lockless shared Submission Queue (SQ) and Completion Queue (CQ) ring buffers in memory mapped between userspace and kernel—eliminating syscall overhead for high-IOPS NVMe and socket networking.

Related:#ebpf-xdp-kernel-telemetry#lsm-tree-write-amplification#edge-computing
Geometric high-dimensional Module-LWE cryptographic lattice defending data against quantum interference
Cloud & DevOps10.4k words

Post-Quantum Cryptography (ML-KEM & SLH-DSA)

NIST-standardized cryptographic algorithms (FIPS 203 Module-Lattice Key Encapsulation and FIPS 205 Stateless Hash-Based Signatures) engineered to withstand Shor's quantum integer-factorization and discrete-logarithm attacks.

Related:#edge-computing#ebpf-xdp-kernel-telemetry#model-context-protocol-mcp
Deterministic discrete-event simulation cube replaying distributed network splits from a golden PRNG seed
Web Architecture13.8k words

Deterministic Simulation Testing (DST)

A verification architecture pioneered by FoundationDB and TigerBeetle that executes entire multi-node distributed systems inside a single-process, seed-driven virtualized runtime—controlling every clock tick, network packet drop, and disk fsync fault for 100% reproducible bug reproduction.

Related:#crdt-local-first-sync#lsm-tree-write-amplification#agentic-loop
Decentralized mesh of local devices merging diverging state graphs into identical crystalline DAGs
Web Architecture12.2k words

CRDTs (Conflict-Free Replicated Data Types) & Local-First Sync

Algebraically structured replicated data types whose merge operation satisfies Associativity, Commutativity, and Idempotency (ACI)—guaranteeing all disconnected replicas mathematically converge to identical state once updates arrive in any order.

Related:#deterministic-simulation-testing#edge-computing#server-components-rsc
Sequential write-ahead log flowing into compacted leveled SSTable tiers with Bloom filter gates
Data Engineering11.5k words

LSM-Trees, Compaction & Write Amplification

A disk-optimized storage engine architecture that buffers mutations sequentially in an in-memory MemTable + WAL and flushes immutable Sorted String Tables (SSTables) merged asynchronously—delivering massive write throughput governed by leveled vs tiered compaction tradeoffs.

Related:#hnsw-diskann-vector-indexing#io-uring-zero-copy-io#deterministic-simulation-testing
Multi-layer HNSW proximity graph suspended over an NVMe DiskANN Vamana storage slab
Data Engineering12.4k words

HNSW & DiskANN Vector Indexing

The two dominant Approximate Nearest Neighbor (ANN) graph architectures: HNSW keeps multi-layered proximity edges entirely in RAM for sub-2ms recall, while DiskANN combines Product-Quantized codes in memory with a single-layer Vamana graph on NVMe SSDs for billion-scale cost efficiency.

Related:#vector-embeddings#rag-retrieval-augmented-generation#lsm-tree-write-amplification