perconnection

Perconnection, often written as per-connection, is a design approach in which system behavior, policy enforcement, and resource allocation are determined and applied on an individual network connection basis rather than uniformly across an entire system. The concept is used in networking, security, and application infrastructure to tailor handling to the characteristics and state of each connection.

Common uses include per-connection rate limiting to curb abuse while preserving throughput for legitimate users, per-connection firewall or access control decisions, per-connection encryption context management, and per-connection logging or auditing. The approach is relevant for protocols with persistent connections such as TCP streams, TLS sessions, WebSockets, and VPN tunnels, as well as for stateless protocols implemented with connection-oriented abstractions.

Implementing perconnection policies requires maintaining per-connection state, which can challenge scalability and memory usage. Efficient data structures, connection lifecycle management, and cleanup of stale state are important. Privacy and security concerns include preventing leakage of connection-specific data and ensuring isolation between connections. Performance overhead can arise from per-connection lookups and policy evaluations.

Perconnection enables finer-grained control, improved security posture, and more accurate accounting, but it increases system complexity, resource usage, and potential latency. It trades simplicity and predictability for adaptability and specificity.

Examples include web servers applying rate limits per TCP connection, intrusion prevention systems enforcing per-connection rules, and cloud load balancers maintaining per-connection routing state to steer subsequent packets or sessions. Some implementations combine per-connection policies with higher-level per-user or per-tenant controls to balance scalability and protection.

See also per-flow, per-session, and connection-oriented networking concepts.