Persistence Profile in BIG-IP

Persistence profiles in BIG-IP manage how the system maintains connections between clients and servers to ensure consistent behavior. Below is a detailed breakdown of persistence types, behaviors, and advanced options.

---

General Features

• Connection Limit Overrides:

  • Pool member connection limits can be overridden for persisted clients.

  • Virtual server connection limits remain strict and are not overridden.

---

Persistence Types

1. Cookie Persistence:

   • Designed for HTTP connections.

   • Tracks sessions by inserting or handling cookies between clients and servers.

   • Options:

     • Insert (Default):

       • BIG-IP inserts a custom cookie in server responses.

       • Encodes the server's IP and port in the cookie (BIGipServer<pool_name>).

       • Enabling Always Send Cookie ensures all F5 responses include the cookie (performance impact).

     • Rewrite:

       • Modifies cookies sent by the server (BIGipCookie=<120-zeroes> format) into F5's encoded cookie format.

     • Passive:

       • Forwards all cookies from the server without inserting or modifying them.

       • Server is responsible for persistence management.

     • Hash:

       • Server provides a cookie, and BIG-IP consistently maps it to a specific node.

       • Requires specifying the cookie name in the persistence profile.

---

1. Source Address Persistence:

   • Tracks clients based on their IP address.

   • Supports TCP and UDP for host or network-based virtual servers.

2. SSL Persistence:

   • Uses the SSL Session-ID to maintain persistence for non-terminated SSL connections.

3. Destination Address Persistence:

   • For TCP/UDP connections to network-based virtual servers (supports /0 to /31 subnet masks).

4. Hash Persistence:

   • Uses an iRule to create persistence records based on a hash value.

     • Example: persist hash <session-key>

5. Universal Persistence:

   • Flexible; persistence records are created based on specific header values.

     • Example: persist uie <session-key>

6. Microsoft RDP Persistence:

   • Tracks sessions using RDP Session-ID.

7. SIP Persistence:

   • Tracks sessions using Caller-ID for SIP over TCP, UDP, or SCTP.

---

CARP (Cache Array Routing Protocol) Hash Algorithm

• Stateless load balancing for HTTP or compatible protocols.

• Key Features:

  • No persistence table required.

  • Automatic load redistribution between members.

  • Suitable for HTTP cache proxies, HTTP traffic, and OneConnect configurations.

  • Commands like tmsh show /ltm persistence persist-records and bigpipe persist show all cannot be used.

  • Excellent for stateless persistence scenarios.

---

Advanced Persistence Options

Criteria for Session Persistence

• For cookie profiles, "Cookie Hash" settings apply when VS → Pool → Node mappings include the same member addresses.

"Match Across Services"

• Port-based matching.

• Used when multiple VSs share the same virtual IP but have different ports.

  • Example:

    • VS1: 192.168.0.2:80

    • VS2: 192.168.0.2:443

"Match Across Virtual Servers"

• Allows persistence across virtual servers with different virtual IPs and ports.

• Useful when multiple VSs point to the same backend pool or node.

"Match Across Pools"

• Enables persistence across multiple pools.

• The system uses any pool containing a persistence record for the client.

  • Caution: Can redirect traffic to a pool not explicitly configured for the virtual server.

---

Summary of Use Cases

Persistence Type	Use Case
Cookie	HTTP traffic with client-side tracking via cookies.
Source Address	Sessions requiring IP-based tracking (e.g., stateless clients).
SSL	Non-terminated SSL traffic with session consistency.
Destination Address	Network-based load balancing scenarios.
Hash / Universal	Custom iRule-driven persistence logic.
Microsoft RDP	Microsoft Remote Desktop Protocol sessions.
SIP	Caller-ID-based SIP traffic.
CARP Hash	Stateless, HTTP cache proxies, and OneConnect configurations.