Skip to content

Aegis Project Roadmap

This document outlines the planned future features and optimization tracks for the Aegis Packet Processing Framework.

Phase 1: Core Framework (Current Work)

  • Cache-aligned lock-free SPSC queue infrastructure.
  • Zero-copy protocol headers and payload parsing.
  • Flat pre-allocated connection tables.
  • Double-buffered atomic rules hot-reloader.
  • Polling Output Writer architecture.
  • Micro and Macro benchmark suites.

Phase 2: Fuzzing & Testing Expansion

  • libFuzzer Integration: Set up continuous fuzz testing for packet parsing and TLS SNI extractor entry points.
  • Malformed Frame Harness: Fuzz with corrupt IP, TCP, and TLS lengths to verify parser robustness against memory violations.
  • Continuous Regression Benchmarking: Set up a GitHub Actions workflow to run nightlies and fail the build if processing throughput decreases by more than 5%.

Phase 3: Hardware Acceleration & SIMD

  • SIMD Vectorization (AVX2/SSE/NEON):
  • Vectorize IP/CIDR matching loops to check 4 or 8 addresses in parallel.
  • Vectorize TLS SNI string searches and suffix trie scanning using SIMD vector instructions (e.g. _mm256_cmpeq_epi8).
  • Kernel Bypass Ingestion:
  • Implement an optional DPDK (Data Plane Development Kit) or AF_XDP ingestion plugin to read packets directly from the NIC ring buffer, bypassing the Linux network stack.

Phase 4: Plugin ABI & External Modules

  • Dynamic Plugin Loading:
  • Develop a stable ABI interface (aegis_plugin.h) for protocol decoders.
  • Support loading custom dynamic libraries (.so / .dll) at runtime.
  • Implement registration, capability flags, and API version handshakes.