Security Associations (SA) and Security Policy Database
What you'll learn: How IPsec maintains connection state and enforces policy decisions through Security Associations and the Security Policy Database.
The Core Concept
When IPsec establishes a secure connection (like a VPN tunnel), it doesn't just send encrypted packets randomly. Instead, it creates a Security Association (SA) — a one-way agreement that defines exactly how to protect traffic between two endpoints. Think of an SA as a contract that specifies: "For traffic going from A to B, use AES-256 encryption, SHA-256 authentication, and here's the shared key."
Because communication is bidirectional, you actually need two SAs per connection: one for outbound traffic and one for inbound.
How the Security Policy Database (SPD) Works
Before an SA is created, IPsec consults the Security Policy Database (SPD) — a ruleset that determines what traffic needs protection and how to protect it. The SPD contains policy entries like:
- "All traffic to 192.168.10.0/24 must be encrypted with ESP"
- "Traffic on port 80 should bypass the VPN"
- "Block traffic from untrusted subnet 10.5.5.0/24"
When a packet arrives or departs, IPsec checks the SPD first. Based on source/destination IP, port, and protocol, it decides: encrypt this, bypass that, or drop something else.
The Relationship
SPD → SA workflow:
- Packet arrives matching an SPD policy requiring encryption
- IPsec checks if an existing SA matches the policy
- If no SA exists, initiate key exchange (IKE) to create one
- Apply the SA's encryption/authentication parameters to the packet
Key Takeaway: Security Associations are the actual connection states containing cryptographic parameters, while the Security Policy Database acts as the decision-maker that determines which traffic needs protection and triggers SA creation for policy-based routing.