GRAPH NAME COMPUTER SYSTEM ARCHITECTURE - MASTER TIMELINE *** ## NODE 1 NAME STORED-PROGRAM ARCHITECTURE DATE 1945-06-30 PLACE USA - Pennsylvania WHO John von Neumann, J. Presper Eckert, John Mauchly and EDVAC team BRIEF DESCRIPTION The First Draft of a Report on the EDVAC describes a computer architecture in which both instructions and data can be stored electronically in memory. This separates program logic from physical rewiring and establishes the architectural foundation of the programmable general-purpose computer. The stored-program principle becomes fundamental to processors, operating systems, applications, virtual machines and virtually every later software architecture. LINK *** ## NODE 2 NAME ENIAC DATE 1946-02-14 PLACE USA - Pennsylvania - Philadelphia WHO J. Presper Eckert, John Mauchly and University of Pennsylvania BRIEF DESCRIPTION ENIAC is publicly unveiled as a large-scale electronic general-purpose digital computer. Although its original programming model relies heavily on switches, cables and physical configuration, it demonstrates the enormous performance potential of electronic computation and exposes the architectural limitations that stored-program computers are designed to solve. LINK *** ## NODE 3 NAME TRANSISTOR DATE 1947-12-23 PLACE USA - New Jersey - Bell Laboratories WHO John Bardeen, Walter Brattain and William Shockley BRIEF DESCRIPTION The transistor demonstrates that electronic switching and amplification can be implemented without vacuum tubes. Smaller size, lower power consumption and greater reliability eventually make increasingly dense processors, memory systems and embedded computers possible. Modern software architecture ultimately depends on the massive hardware abstraction enabled by semiconductor switching. LINK *** ## NODE 4 NAME MANCHESTER BABY DATE 1948-06-21 PLACE United Kingdom - England - Manchester WHO Frederic Williams, Tom Kilburn and Geoff Toothill BRIEF DESCRIPTION The Manchester Small-Scale Experimental Machine successfully executes a program stored electronically in memory. It becomes the first digital electronic stored-program computer to execute a program and demonstrates that instructions can be represented, stored and manipulated in the same computational memory system as data. LINK *** ## NODE 5 NAME EDSAC DATE 1949-05-06 PLACE United Kingdom - England - Cambridge WHO Maurice Wilkes and University of Cambridge BRIEF DESCRIPTION EDSAC begins regular operation as one of the earliest practical stored-program computers. Its use of reusable subroutines and a library of program routines helps establish an important architectural idea: software systems should be constructed from reusable components rather than recreated entirely for each program. LINK *** ## NODE 6 NAME COMMERCIAL STORED-PROGRAM COMPUTER DATE 1951-02 PLACE United Kingdom - England - Manchester WHO FERRANTI AND UNIVERSITY OF MANCHESTER BRIEF DESCRIPTION The Ferranti Mark 1 becomes the first commercially available general-purpose stored-program computer. Computing architecture begins moving from experimental laboratories toward repeatable commercial platforms, creating the conditions for standardized hardware, reusable software, programming services and eventually independent software ecosystems. LINK *** ## NODE 7 NAME INTEGRATED CIRCUIT DATE 1958-09-12 PLACE USA - Texas - Dallas WHO Jack Kilby and TEXAS INSTRUMENTS BRIEF DESCRIPTION Jack Kilby demonstrates an integrated electronic circuit containing multiple components within a single semiconductor device. Integrated circuits enable increasing computational density and decreasing cost and eventually make processors, memory, controllers and entire computing systems compact enough to become ubiquitous architectural building blocks. LINK *** ## NODE 8 NAME CTSS TIME-SHARING DATE 1961 PLACE USA - Massachusetts - Cambridge WHO Fernando Corbató and MIT BRIEF DESCRIPTION The Compatible Time-Sharing System demonstrates that multiple interactive users can share a central computer while maintaining separate sessions, programs and files. Time-sharing introduces architectural concepts that later appear throughout operating systems and cloud computing: users, processes, scheduling, access control, persistent files and shared computational resources. LINK *** ## NODE 9 NAME IBM SYSTEM/360 DATE 1964-04-07 PLACE USA WHO Fred Brooks, Gene Amdahl, Bob Evans, Erich Bloch and IBM BRIEF DESCRIPTION IBM introduces System/360 as a compatible family of computers sharing a common instruction architecture. Software can operate across multiple models instead of being rewritten for every machine. The system strengthens the separation between hardware implementation and software platform and helps establish computer architecture as a distinct engineering discipline. LINK *** ## NODE 10 NAME MULTICS DATE 1965 PLACE USA - Massachusetts - Cambridge WHO MIT, BELL LABORATORIES AND GENERAL ELECTRIC BRIEF DESCRIPTION Development of Multics introduces an ambitious architecture for a continuously available multiuser computing utility. Important ideas include hierarchical filesystems, dynamic linking, segmentation, protection rings, shared libraries, virtual memory and fine-grained access control. Many concepts later influence Unix, operating-system security and cloud computing. LINK *** ## NODE 11 NAME VIRTUAL MACHINE ARCHITECTURE DATE 1967 PLACE USA WHO IBM CAMBRIDGE SCIENTIFIC CENTER BRIEF DESCRIPTION IBM's CP-67 and related virtualization work demonstrate that a physical mainframe can host multiple isolated virtual machines. Each environment behaves as if it controls its own computer. Hardware virtualization creates an abstraction layer that later becomes fundamental to data centers, infrastructure consolidation, cloud computing and secure workload isolation. LINK *** ## NODE 12 NAME UNIX DATE 1969 PLACE USA - New Jersey - Bell Laboratories WHO Ken Thompson, Dennis Ritchie and collaborators BRIEF DESCRIPTION Unix begins development as a comparatively small operating system organized around processes, files and composable programs. Its architectural philosophy favors simple components connected through well-defined interfaces. Unix becomes a foundational model for modern operating systems, servers, developer tools, containers and cloud infrastructure. LINK *** ## NODE 13 NAME ARPANET DATE 1969-09 PLACE USA - California WHO ARPA, UCLA, SRI, UCSB AND UNIVERSITY OF UTAH BRIEF DESCRIPTION ARPANET connects independently operated computers through a packet-switched network. Computing architecture begins expanding beyond a single machine into interconnected systems whose components communicate across unreliable physical infrastructure. The development creates a foundation for network services, distributed systems and ultimately Internet-scale software architecture. LINK *** ## NODE 14 NAME RELATIONAL DATA MODEL DATE 1970 PLACE USA - California - San Jose WHO Edgar F. Codd BRIEF DESCRIPTION Edgar Codd proposes the relational model for shared data banks. Data is represented through relations rather than application-specific navigation structures. The model separates logical data representation from many physical storage details and provides a foundation for database servers, repositories, SQL, transactions and modern persistence layers. LINK *** ## NODE 15 NAME MICROPROCESSOR DATE 1971-11-15 PLACE USA - California WHO Federico Faggin, Ted Hoff, Stan Mazor, Masatoshi Shima and INTEL BRIEF DESCRIPTION Intel announces the 4004 general-purpose programmable microprocessor. Processing logic that previously required many separate integrated circuits can increasingly be concentrated into a programmable chip. The microprocessor enables personal computers, embedded systems, controllers, sensors and eventually massive populations of distributed computing devices. LINK *** ## NODE 16 NAME ACTOR MODEL DATE 1973 PLACE USA - Massachusetts - Cambridge WHO Carl Hewitt, Peter Bishop and Richard Steiger BRIEF DESCRIPTION The Actor Model describes computation as independent actors that communicate through asynchronous messages. Actors encapsulate state and can create other actors, send messages and determine future behavior. The model becomes a major conceptual foundation for concurrency, distributed systems, event-driven architectures and fault-tolerant software. LINK *** ## NODE 17 NAME UNIX PIPES DATE 1973 PLACE USA - New Jersey - Bell Laboratories WHO Douglas McIlroy, Ken Thompson and UNIX team BRIEF DESCRIPTION Unix introduces practical pipelines that connect the output of one process directly to the input of another. Complex operations can be constructed by composing small independent programs rather than building a single monolithic program. Pipes become one of the clearest early demonstrations of software component composition. LINK *** ## NODE 18 NAME ETHERNET DATE 1973-05-22 PLACE USA - California - Palo Alto WHO Robert Metcalfe, David Boggs and XEROX PARC BRIEF DESCRIPTION Ethernet is proposed as a local-network architecture for connecting computers and peripherals. It provides a standardized communication layer beneath applications and eventually becomes dominant throughout local networks, servers, industrial systems and data centers. LINK *** ## NODE 19 NAME TCP INTERNETWORKING DATE 1974-05 PLACE USA WHO Vint Cerf and Bob Kahn BRIEF DESCRIPTION Cerf and Kahn publish a protocol architecture for packet network intercommunication. The design separates end systems from underlying network technologies and establishes a common architecture for connecting heterogeneous networks. This abstraction becomes the foundation of the Internet and network-based distributed software. LINK *** ## NODE 20 NAME SEQUEL AND SQL DATE 1974 PLACE USA - California - San Jose WHO Donald Chamberlin, Raymond Boyce and IBM BRIEF DESCRIPTION IBM researchers develop SEQUEL, which evolves into SQL, as a declarative language for relational databases. Applications can specify what data they require without encoding every physical retrieval operation. SQL becomes a durable contract between application logic and persistent relational data services. LINK *** ## NODE 21 NAME MULTI-NETWORK INTERNET DEMONSTRATION DATE 1977-11-22 PLACE USA WHO Vint Cerf, Bob Kahn, ARPA and collaborators BRIEF DESCRIPTION A major internetworking demonstration successfully connects heterogeneous packet networks using TCP technology. The experiment validates the principle that independently designed networks can participate in a larger network of networks, an architectural idea central to the modern Internet and distributed systems. LINK *** ## NODE 22 NAME MODEL VIEW CONTROLLER DATE 1979 PLACE USA - California - Palo Alto WHO Trygve Reenskaug and XEROX PARC BRIEF DESCRIPTION The Model-View-Controller architectural pattern emerges from Smalltalk work at Xerox PARC. Application state, user presentation and interaction logic are separated into cooperating components. MVC becomes one of the most influential patterns for graphical, desktop and later Web application architecture. LINK *** ## NODE 23 NAME TCP/IP ADOPTION DATE 1983-01-01 PLACE USA WHO ARPANET AND INTERNET RESEARCH COMMUNITY BRIEF DESCRIPTION ARPANET formally transitions to TCP/IP. A common layered communication architecture allows heterogeneous machines and networks to interoperate. Applications can increasingly rely on standardized transport and addressing rather than understanding the underlying network implementation. LINK *** ## NODE 24 NAME DOMAIN NAME SYSTEM DATE 1983-11 PLACE USA WHO Paul Mockapetris and INTERNET ENGINEERING COMMUNITY BRIEF DESCRIPTION The Domain Name System replaces an increasingly unmanageable centralized host-name list with a hierarchical distributed naming architecture. DNS separates logical service names from changing network addresses and becomes one of the most important examples of globally distributed infrastructure. LINK *** ## NODE 25 NAME REMOTE PROCEDURE CALL DATE 1984 PLACE USA - California - Palo Alto WHO Andrew Birrell and Bruce Nelson BRIEF DESCRIPTION Remote Procedure Call architecture formalizes a programming model in which software invokes operations in another process or machine through procedure-like interfaces. RPC introduces stubs, serialization, request-response communication and interface contracts and becomes a major ancestor of CORBA, distributed objects, gRPC and modern service APIs. LINK *** ## NODE 26 NAME NETWORK FILE SYSTEM DATE 1984 PLACE USA - California WHO SUN MICROSYSTEMS BRIEF DESCRIPTION Sun develops Network File System so applications can access remote files through filesystem-like abstractions. NFS demonstrates architectural transparency across machine boundaries and becomes an important distributed-storage primitive for Unix systems, clusters and networked infrastructure. LINK *** ## NODE 27 NAME MACH MICROKERNEL DATE 1985 PLACE USA - Pennsylvania - Pittsburgh WHO Richard Rashid and CARNEGIE MELLON UNIVERSITY BRIEF DESCRIPTION The Mach project explores an operating-system architecture based on a smaller kernel providing fundamental primitives while higher-level services can operate outside the kernel. Microkernel research strengthens architectural ideas involving message passing, isolation, modularity and replacement of operating-system services. LINK *** ## NODE 28 NAME ERLANG DATE 1986 PLACE Sweden WHO Joe Armstrong, Robert Virding, Mike Williams and ERICSSON BRIEF DESCRIPTION Erlang is developed for highly concurrent telecommunications systems. Lightweight processes communicate through messages and avoid shared mutable state. Built-in mechanisms for failure isolation and recovery influence later actor frameworks, distributed services and architectures designed under the assumption that components will eventually fail. LINK *** ## NODE 29 NAME POSIX DATE 1988 PLACE GLOBAL WHO IEEE BRIEF DESCRIPTION The first POSIX standard establishes standardized operating-system interfaces for Unix-like environments. Programs become less dependent on individual vendor implementations. Standard APIs strengthen software portability and reinforce the architectural separation between applications and underlying operating systems. LINK *** ## NODE 30 NAME WORLD WIDE WEB PROPOSAL DATE 1989-03 PLACE Switzerland - Geneva WHO Tim Berners-Lee and CERN BRIEF DESCRIPTION Tim Berners-Lee proposes a distributed hypertext information system based on network-addressable resources, hyperlinks and standardized protocols. The architecture combines clients, servers, identifiers and documents and establishes the conceptual foundation of the World Wide Web. LINK *** ## NODE 31 NAME PUBLIC WORLD WIDE WEB DATE 1991-08-06 PLACE Switzerland - Geneva WHO Tim Berners-Lee and CERN BRIEF DESCRIPTION Information about the World Wide Web becomes publicly available on the Internet. Web architecture creates a universal application-delivery model in which clients retrieve resources from independently deployed servers through common protocols and identifiers. LINK *** ## NODE 32 NAME LINUX DATE 1991-08-25 PLACE Finland WHO Linus Torvalds and open source community BRIEF DESCRIPTION The Linux kernel project begins and develops into a portable Unix-like operating-system kernel. Linux later becomes the dominant execution foundation for Internet servers, cloud virtual machines, containers, Kubernetes nodes, embedded devices and many critical distributed systems. LINK *** ## NODE 33 NAME CORBA DATE 1991-10 PLACE GLOBAL WHO OBJECT MANAGEMENT GROUP BRIEF DESCRIPTION CORBA 1.0 defines an Object Request Broker architecture, Interface Definition Language and APIs for distributed object invocation. It represents a major attempt to make software components written in different languages and running on different machines interact through standardized contracts. LINK *** ## NODE 34 NAME JAVA VIRTUAL MACHINE DATE 1995-05-23 PLACE USA - California WHO SUN MICROSYSTEMS BRIEF DESCRIPTION Java introduces a managed execution architecture in which compiled bytecode runs inside a virtual machine rather than directly against one hardware platform. The JVM provides memory management, security mechanisms, portability and runtime services and becomes a major application platform for server and enterprise systems. LINK *** ## NODE 35 NAME HTTP/1.1 DATE 1997-01 PLACE GLOBAL WHO IETF AND WEB COMMUNITY BRIEF DESCRIPTION HTTP/1.1 standardizes persistent connections, host-based virtual hosting, caching controls and numerous improvements required by the rapidly expanding Web. HTTP becomes not only a document-transfer protocol but an increasingly general communication layer for distributed applications. LINK *** ## NODE 36 NAME PAXOS DATE 1998 PLACE USA WHO Leslie Lamport BRIEF DESCRIPTION Paxos provides a rigorous protocol for reaching consensus among distributed participants despite failures and uncertain message timing. Consensus becomes a foundational mechanism for replicated databases, distributed coordination, configuration systems and highly available infrastructure. LINK *** ## NODE 37 NAME ERLANG OTP DATE 1998 PLACE Sweden WHO ERICSSON AND ERLANG COMMUNITY BRIEF DESCRIPTION Erlang and the Open Telecom Platform become openly available. OTP formalizes reusable architectural behaviors including supervisors, workers, servers and restart strategies. Supervision trees make fault recovery part of system architecture rather than an exceptional afterthought. LINK *** ## NODE 38 NAME X86 VIRTUALIZATION DATE 1999 PLACE USA - California WHO VMWARE BRIEF DESCRIPTION VMware introduces practical virtualization for commodity x86 computers. Multiple isolated operating systems can share ordinary server hardware. Virtual machines become increasingly practical outside mainframes and establish the infrastructure abstraction later used throughout enterprise virtualization and public clouds. LINK *** ## NODE 39 NAME SOAP DATE 2000 PLACE GLOBAL WHO MICROSOFT, IBM AND WEB STANDARDS COMMUNITY BRIEF DESCRIPTION SOAP develops as a standardized XML-based protocol for exchanging structured messages between network services. Together with WSDL and related specifications, it helps establish service-oriented architecture based on explicit machine-readable contracts between independently deployed systems. LINK *** ## NODE 40 NAME REST ARCHITECTURE DATE 2000 PLACE USA - California WHO Roy Fielding BRIEF DESCRIPTION Roy Fielding formally describes Representational State Transfer as an architectural style for network-based software. REST emphasizes resources, representations, stateless interactions, uniform interfaces, caching and layered intermediaries. It becomes one of the dominant foundations of Web API architecture. LINK *** ## NODE 41 NAME CAP CONJECTURE DATE 2000 PLACE USA - California WHO Eric Brewer BRIEF DESCRIPTION Eric Brewer articulates the trade-off among consistency, availability and tolerance of network partitions in distributed systems. CAP helps architects reason explicitly about failure, replication and consistency rather than designing distributed systems as though network communication were perfectly reliable. LINK *** ## NODE 42 NAME XEN HYPERVISOR DATE 2003 PLACE United Kingdom - England - Cambridge WHO Ian Pratt, Keir Fraser, University of Cambridge and collaborators BRIEF DESCRIPTION Xen demonstrates an efficient open-source hypervisor architecture based initially on paravirtualization. It becomes influential in large-scale server virtualization and early cloud infrastructure and reinforces the hypervisor as an independent architectural layer between physical machines and operating systems. LINK *** ## NODE 43 NAME MAPREDUCE DATE 2004 PLACE USA - California WHO Jeffrey Dean, Sanjay Ghemawat and GOOGLE BRIEF DESCRIPTION Google describes MapReduce as a programming architecture for processing extremely large datasets across clusters of commodity machines. Developers express map and reduce functions while the execution framework handles partitioning, scheduling, data movement and failures. The architecture demonstrates how distributed-system complexity can be hidden behind a higher-level programming model. LINK *** ## NODE 44 NAME GIT DATE 2005 PLACE GLOBAL WHO Linus Torvalds and Linux community BRIEF DESCRIPTION Git introduces a distributed version-control architecture designed for large and highly parallel software projects. Every repository can contain complete project history, branches are inexpensive and development no longer depends on a single central version-control server. Git becomes foundational infrastructure for modern software collaboration, CI/CD and open-source development. LINK *** ## NODE 45 NAME AMAZON S3 DATE 2006-03-14 PLACE USA WHO AMAZON WEB SERVICES BRIEF DESCRIPTION Amazon S3 exposes durable object storage as a programmable Internet service. Applications can acquire storage through an API rather than provisioning storage hardware directly. Infrastructure begins increasingly behaving as a remotely accessible software component. LINK *** ## NODE 46 NAME AMAZON EC2 DATE 2006-08-24 PLACE USA WHO AMAZON WEB SERVICES BRIEF DESCRIPTION Amazon EC2 introduces resizable virtual computing capacity through Web APIs. Developers can create and destroy virtual servers programmatically within minutes rather than purchasing and configuring physical machines. Compute infrastructure becomes elastic, metered and automatable, establishing a central architectural foundation of public cloud computing. LINK *** ## NODE 47 NAME AMAZON DYNAMO DATE 2007 PLACE USA WHO AMAZON BRIEF DESCRIPTION Amazon describes Dynamo, a distributed key-value storage architecture designed for high availability under failures. Techniques involving partitioning, replication, eventual consistency, consistent hashing and quorum-like mechanisms influence a generation of distributed NoSQL databases. LINK *** ## NODE 48 NAME LINUX CONTROL GROUPS DATE 2007 PLACE GLOBAL WHO GOOGLE AND LINUX KERNEL COMMUNITY BRIEF DESCRIPTION Control groups add mechanisms for organizing processes and limiting or accounting for their CPU, memory and other resource usage. Combined with Linux namespaces, cgroups provide critical kernel-level primitives from which modern application containers are constructed. LINK *** ## NODE 49 NAME NODE.JS DATE 2009 PLACE USA WHO Ryan Dahl and open source community BRIEF DESCRIPTION Node.js popularizes event-driven, non-blocking server programming using JavaScript outside the browser. A single event loop can coordinate large numbers of network operations without creating a traditional thread for every request. The model influences modern asynchronous APIs, gateways and real-time services. LINK *** ## NODE 50 NAME DEVOPS DATE 2009 PLACE Belgium and Global WHO Patrick Debois and DevOps community BRIEF DESCRIPTION The DevOps movement begins consolidating practices that reduce the organizational and technical separation between software development and system operations. Automated testing, configuration, deployment, monitoring and feedback increasingly become part of application architecture rather than external manual procedures. LINK *** ## NODE 51 NAME DISTRIBUTED TRACING AT SCALE DATE 2010 PLACE USA - California WHO GOOGLE BRIEF DESCRIPTION Google describes Dapper, a distributed tracing infrastructure for large production systems. Requests can be followed across many services and machines using correlated trace identifiers and spans. Distributed tracing becomes essential as application architecture moves from monoliths toward large graphs of network services. LINK *** ## NODE 52 NAME APACHE KAFKA DATE 2011 PLACE USA - California WHO Jay Kreps, Neha Narkhede, Jun Rao and LINKEDIN BRIEF DESCRIPTION Kafka introduces a distributed append-only log architecture for durable high-throughput event streams. Producers and consumers become decoupled through persistent ordered partitions. Event streaming develops into a core architectural mechanism for integration, asynchronous processing, analytics and event-driven systems. LINK *** ## NODE 53 NAME TWELVE-FACTOR APPLICATION DATE 2011 PLACE USA - California WHO Adam Wiggins and HEROKU BRIEF DESCRIPTION The Twelve-Factor methodology formalizes practices for applications intended to operate on modern service platforms. It emphasizes explicit dependencies, external configuration, backing services, stateless processes, disposability, environment parity and logs as event streams. These principles strongly influence cloud-native application architecture. LINK *** ## NODE 54 NAME FOG COMPUTING DATE 2012 PLACE USA - California WHO Flavio Bonomi, Rodolfo Milito, Jiang Zhu, Sateesh Addepalli and CISCO BRIEF DESCRIPTION Fog computing formalizes an architecture that extends computation, storage and networking away from centralized cloud data centers toward resources located closer to physical devices and users. It becomes an important precursor to modern edge computing and distributed cyberphysical architectures. LINK *** ## NODE 55 NAME DOCKER DATE 2013-03-15 PLACE USA - California WHO Solomon Hykes and DOTCLOUD BRIEF DESCRIPTION Docker is publicly demonstrated as a developer-oriented container system. It packages an application and its dependencies into a standardized portable unit while using Linux isolation mechanisms underneath. Containers dramatically reduce friction between development, testing and production and create a new standard deployment abstraction. LINK *** ## NODE 56 NAME ETCD DATE 2013-06-02 PLACE USA WHO COREOS BRIEF DESCRIPTION The etcd project begins as a distributed strongly consistent key-value store intended for configuration and coordination. It later becomes the persistent state store behind Kubernetes control planes and illustrates the growing importance of consensus-backed coordination services in cloud-native architecture. LINK *** ## NODE 57 NAME MICROSERVICES ARCHITECTURE DATE 2014-03 PLACE GLOBAL WHO James Lewis, Martin Fowler and distributed systems community BRIEF DESCRIPTION Microservices are clearly articulated as an architectural style in which an application is decomposed into small independently deployable services organized around business capabilities. Services communicate through lightweight mechanisms and can evolve independently, but the approach makes automation, monitoring, fault isolation and distributed-system engineering essential. LINK *** ## NODE 58 NAME KUBERNETES DATE 2014-06-10 PLACE USA WHO GOOGLE AND OPEN SOURCE COMMUNITY BRIEF DESCRIPTION Google publicly announces Kubernetes as an open-source system for orchestrating containerized applications. The architecture introduces declarative desired state, pods, controllers and reconciliation mechanisms derived from experience operating large internal clusters. Infrastructure management increasingly becomes an automated control-system problem. LINK *** ## NODE 59 NAME INFRASTRUCTURE AS CODE DATE 2014-07-28 PLACE USA WHO Mitchell Hashimoto and HASHICORP BRIEF DESCRIPTION Terraform is announced as a tool for expressing infrastructure through declarative configuration. Servers, networks, DNS, storage and external services can be represented as version-controlled software definitions and reconciled with actual infrastructure. Infrastructure becomes an explicit programmable architectural artifact. LINK *** ## NODE 60 NAME AWS LAMBDA DATE 2014-11-13 PLACE USA WHO AMAZON WEB SERVICES BRIEF DESCRIPTION AWS Lambda introduces an event-driven compute model in which developers deploy functions while the platform manages infrastructure, scaling and execution. Serverless architecture moves the abstraction boundary beyond virtual machines and containers toward individual units of application behavior triggered by events. LINK *** ## NODE 61 NAME EXTENDED BPF DATE 2014-12 PLACE GLOBAL WHO LINUX KERNEL COMMUNITY BRIEF DESCRIPTION Extended BPF evolves the Linux kernel's packet-filtering technology into a general mechanism for safely executing verified programs inside the kernel. eBPF later enables programmable networking, security, tracing and observability without requiring conventional kernel modules and becomes a major infrastructure primitive in cloud-native systems. LINK *** ## NODE 62 NAME GRPC DATE 2015-03 PLACE USA - California WHO GOOGLE AND OPEN SOURCE COMMUNITY BRIEF DESCRIPTION Google open-sources gRPC as a modern RPC framework derived from its internal service infrastructure. Protocol Buffers define strongly typed service contracts while HTTP/2 provides multiplexed and bidirectional communication. gRPC becomes a major communication mechanism for microservices, infrastructure control planes and distributed systems. LINK *** ## NODE 63 NAME GRAPHQL DATE 2015 PLACE USA - California WHO FACEBOOK BRIEF DESCRIPTION Facebook open-sources GraphQL after using it internally since 2012. Clients query a strongly typed schema and request precisely the data structures they need. GraphQL introduces an alternative API architecture particularly suited to complex frontend applications and rapidly evolving data graphs. LINK *** ## NODE 64 NAME OPEN CONTAINER INITIATIVE DATE 2015-06 PLACE GLOBAL WHO LINUX FOUNDATION, DOCKER AND INDUSTRY COMMUNITY BRIEF DESCRIPTION Industry participants establish an open standards initiative for container image formats and runtimes. The container becomes increasingly independent of a particular vendor implementation, allowing build tools, registries, runtimes and orchestrators to interoperate around standardized artifacts. LINK *** ## NODE 65 NAME KUBERNETES 1.0 AND CNCF DATE 2015-07-21 PLACE GLOBAL WHO GOOGLE, LINUX FOUNDATION AND CLOUD NATIVE COMMUNITY BRIEF DESCRIPTION Kubernetes reaches version 1.0 and Google contributes the project to the newly established Cloud Native Computing Foundation. Cloud-native computing begins consolidating around an open ecosystem of interoperable projects for containers, orchestration, networking, observability, storage and distributed application infrastructure. LINK *** ## NODE 66 NAME ENVOY PROXY DATE 2016-09 PLACE USA - California WHO Matt Klein and LYFT BRIEF DESCRIPTION Envoy is open-sourced as a high-performance application proxy designed for distributed service architectures. Networking concerns including discovery, retries, load balancing, telemetry and protocol management can increasingly be moved from application code into specialized infrastructure components. LINK *** ## NODE 67 NAME SERVICE MESH DATE 2017-05-24 PLACE GLOBAL WHO GOOGLE, IBM, LYFT AND ISTIO COMMUNITY BRIEF DESCRIPTION Istio is publicly released as a service-mesh platform for connecting, securing, managing and observing microservices. A dedicated infrastructure layer can control service-to-service traffic, policy and telemetry without requiring every application to independently implement those capabilities. LINK *** ## NODE 68 NAME MICROVM ARCHITECTURE DATE 2018-11-26 PLACE USA WHO AMAZON WEB SERVICES BRIEF DESCRIPTION AWS releases Firecracker as a lightweight virtual-machine manager designed for serverless and container workloads. MicroVMs attempt to combine the stronger workload isolation of virtual machines with startup speed and resource efficiency closer to containers, demonstrating continuing convergence between virtualization and cloud-native execution. LINK *** ## NODE 69 NAME OPENTELEMETRY DATE 2019-05 PLACE GLOBAL WHO OPENTRACING, OPENCENSUS AND CNCF COMMUNITY BRIEF DESCRIPTION OpenTracing and OpenCensus merge to create OpenTelemetry. A common vendor-neutral architecture begins standardizing how applications generate, collect and export traces, metrics and later logs. Observability becomes an explicit cross-cutting interface in modern distributed systems rather than a collection of application-specific logging mechanisms. LINK *** ## NODE 70 NAME WEBASSEMBLY 1.0 DATE 2019-12-05 PLACE GLOBAL WHO W3C WEBASSEMBLY WORKING GROUP BRIEF DESCRIPTION WebAssembly becomes a W3C Recommendation as a safe portable low-level execution format. Code compiled from multiple languages can execute efficiently within a sandboxed runtime. The architecture creates a new portable execution boundary distinct from both native binaries and traditional language-specific virtual machines. LINK *** ## NODE 71 NAME RETRIEVAL-AUGMENTED GENERATION DATE 2020-05-22 PLACE GLOBAL WHO Patrick Lewis, Ethan Perez, Douwe Kiela and collaborators BRIEF DESCRIPTION Retrieval-Augmented Generation combines a generative neural model with external non-parametric information retrieval. The architecture separates model computation from an external knowledge source and establishes a major pattern for modern AI systems: model, embedding or retrieval subsystem, indexed data, context assembly and generation operating as cooperating components. LINK *** ## NODE 72 NAME WASI 0.2 AND WEBASSEMBLY COMPONENT MODEL DATE 2024-01-25 PLACE GLOBAL WHO BYTECODE ALLIANCE AND WEBASSEMBLY COMMUNITY BRIEF DESCRIPTION WASI 0.2 establishes a stable interface set built around the WebAssembly Component Model. Components expose typed imports and exports through explicit interface definitions and can be composed even when implemented in different programming languages. The architecture advances portable, sandboxed and interface-driven software composition beyond browser execution. LINK *** ## NODE 73 NAME MODEL CONTEXT PROTOCOL DATE 2024-11-25 PLACE USA - California - San Francisco WHO ANTHROPIC AND OPEN SOURCE COMMUNITY BRIEF DESCRIPTION The Model Context Protocol is introduced as an open protocol for connecting AI assistants with external data sources and software tools. AI applications increasingly acquire an explicit integration architecture involving models, clients, servers, resources, prompts and callable tools rather than embedding every external-system integration directly into application logic. LINK *** ## NODE 74 NAME OPENTELEMETRY GRADUATION DATE 2026-05-11 PLACE GLOBAL WHO CLOUD NATIVE COMPUTING FOUNDATION AND OPENTELEMETRY COMMUNITY BRIEF DESCRIPTION OpenTelemetry reaches graduated-project status within the Cloud Native Computing Foundation. Standardized traces, metrics and logs have matured into common infrastructure for observing cloud-native and increasingly AI-intensive distributed systems, demonstrating that observability itself has become a first-class architectural subsystem. LINK *** ## NODE 75 NAME CLOUD EDGE AI SYSTEMS OF SYSTEMS DATE 2026 PLACE GLOBAL WHO GLOBAL COMPUTER ENGINEERING COMMUNITY BRIEF DESCRIPTION Modern computer architecture increasingly consists of interacting layers spanning local functions, packages, applications, services, containers, virtual machines, clusters, clouds, edge devices, sensors, accelerators and artificial intelligence systems. Components communicate through explicit contracts, events and protocols while deployment, security, observability, data management and fault recovery operate as independent architectural subsystems. Contemporary architecture is therefore increasingly a system of systems rather than a single executable program. LINK