# ANDYL — Autonomous Networking > Self-healing application networking for software running anywhere. Mobile apps, satellite constellations, autonomous robotics, IoT fleets, and every device on the move. ## Principles ### Real-time resilience Every client should have its own independent network guidance system. Network disruptions should be autonomously healed by intelligent clients, not centralized failovers. Routing changes complete in seconds, not minutes. ### Semantic discovery DNS is the Google Maps of the internet, before live traffic, search, or rerouting. Competitive web and AI search needs routing that stores and aggregates site metadata, with real-time updates following source content. ### Verifiable history In a world where everything can be generated to project authenticity, human history needs a way to permanently stamp content in time. Every route, every resolution, every decision should be provable. ### Built on what works We use the maturity of DNS to bootstrap every request. We're not reinventing the wheel. We're adding a motor. The simplicity of the foundation is a feature, not a limitation to work around. ## Products ### Vector — Network SDK Vector is a cross-platform application networking SDK for clients that need reliable HTTP and gRPC behavior across changing network conditions. URL: https://andyl.com/products/vector **Specs** - **Product**: Vector - **Type**: Networking SDK - **Platforms**: iOS / Android / Rust - **Protocols**: HTTP / gRPC - **Runtime**: Client / On-device - **Availability**: Contact ANDYL **Capabilities** #### Modern application transport Use a consistent request and streaming model across HTTP/1.1, HTTP/2, HTTP/3, and gRPC, with response decompression and transport behavior managed below application code. - HTTP: 1.1 / 2 / 3 - APIs: Requests / Streams / gRPC - Content: Gzip / Brotli #### Network-change resilience Coordinate reachability signals, DNS refresh and caching, connection draining, retries, and per-attempt timeouts so applications can respond deliberately when connectivity changes. - Network: Reachability-aware - DNS: Cache / Refresh - Recovery: Retries / Timeouts #### Policy and built-in diagnostics Apply request and response policies close to the client, with built-in observability, dashboards, and telemetry for understanding application networking in real time. - Security: TLS / Input controls - Diagnostics: Requests / Metrics / Logs - Timings: Per-stage waterfall **Use cases** - **Cross-platform mobile applications**: Give iOS and Android clients a shared transport model while retaining native integration points for each platform. - **Clients on changing networks**: Build explicit recovery behavior for applications that move between interfaces or experience intermittent connectivity. - **Observable network-intensive software**: Diagnose request latency, connection reuse, retries, protocol selection, and transfer volume without rebuilding those signals in every application. ### Orbital — Route Control Global network control plane for millions of device fanout. Orbital coordinates, observes, and directs the ANDYL network at scale. URL: https://andyl.com/products/orbital **Specs** - **Product**: Orbital - **Type**: Control Plane - **Runtime**: Cloud / Distributed - **Scale**: Millions of devices **Capabilities** #### Global fanout Push configuration, policy, and intelligence updates to millions of endpoints simultaneously. Sub-second propagation worldwide. - Propagation: <200ms global - Endpoints: Millions concurrent - Protocol: Persistent, bidirectional #### Real-time control plane Centralized visibility and control over every device running Vector. Monitor, adjust, and respond to network conditions as they happen. - Latency: Real-time telemetry - Policy: Hot-reload, zero downtime - Uptime: 99.999% SLA #### Observability Full telemetry pipeline from edge to cloud. Network health, route decisions, and performance metrics in one place. - Metrics: Route, health, perf - Retention: Configurable - Export: OpenTelemetry compatible **Use cases** - **Fleet management**: Coordinate network behavior across global device deployments with centralized policy control. - **Network operations**: Real-time visibility into every route decision, failover event, and performance metric across your fleet. - **Incident response**: Instantly push routing overrides and mitigation policies to affected devices during outages. ### GSS — Global Signalling Service Open-standard signalling protocol for internet-scale applications. GSS evolves DNS with push updates, structured records, and lightweight functions for configuration, targeting, and discovery. URL: https://andyl.com/products/gss **Specs** - **Product**: GSS - **Type**: Network Protocol - **License**: MIT - **Scope**: Internet / DNS **Capabilities** #### Structured records GSS records carry strongly-typed and structured data for applications to consume directly, like routing rules, service descriptors, and configuration. - Records: DNS-native + structured - Encoding: Binary, schema-aware - Compatibility: Full DNS bridge #### Push updates and client-side functions Publishers push changes in real time. Built-in functions handle cohort targeting, rollouts, load balancing, and routing redundancy on the client. - Updates: Server push, sub-second - Functions: Targeting, rollouts, LB - Evaluation: Client-side, deterministic #### One protocol, many uses Name resolution, service discovery, feature flags, and proxy configs share one wire format and one delivery path. Applications integrate against the GSS client library to consume any of them. - Coverage: Names, configs, flags, routes - Transport: QUIC, TCP fallback - Integrity: End-to-end signed **Use cases** - **Application configuration**: Distribute feature flags, proxy rules, and runtime policies over the same protocol that resolves your names. - **Targeted rollouts**: Run canaries, cohort experiments, and regional rollouts with deterministic client-side assignment. - **Modern service discovery**: Publish service endpoints with weights, locality, protocol metadata, and routing redundancy as structured data. ### AOS — Immutable Linux AOS is an early-preview, immutable Linux distribution for headless servers and edge systems. Its bootstrap toolchain, userspace, packages, images, and tests are built from source without nixpkgs package outputs. URL: https://andyl.com/products/andyl-os **Specs** - **Product**: AOS - **Type**: Linux distribution - **Status**: Early preview - **Architecture**: x86_64 + aarch64 - **Build**: Hermetic / Nix - **Target**: Server / Edge **Capabilities** #### Built from source AOS builds its bootstrap toolchain, userspace, packages, system images, and test tools through its own package set instead of consuming nixpkgs package outputs. - Toolchain: Bootstrapped - Packages: Hermetic builds - Tests: Source-built tools #### Immutable image, runtime policy The release image owns boot-critical capabilities while authenticated host.nix input can activate hostname, networking, users, services, access policy, and desired packages as an atomic configuration generation. - Root: Read-only EROFS - Configuration: Runtime host.nix - Activation: Atomic generations #### Guarded upgrades and recovery AOS implements independent configuration and A/B image generations, boot counting, rollback, dm-verity, and Secure Boot and measured-boot test variants. Production deployments must supply and qualify their own trust material. - Images: A/B generations - Integrity: dm-verity - Trust keys: Deployment-owned **Use cases** - **Headless servers**: Qualify immutable server images whose machine-specific services and policy are activated from authenticated deployment input. - **Edge systems**: Use the edge runtime role and declarative networking on systems that benefit from a small, policy-neutral base image. - **Reproducible OS development**: Inspect, build, boot, and test the operating system and its package set from a single public source tree.