Understanding Persistence in F5 BIG-IP LTM

Why Sessions Stick to the Same Server (Support Engineer’s View)

When learning F5 BIG-IP LTM, one of the most confusing questions is:

“Why is the same user always hitting the same backend server?”

The answer is Persistence.

In this Document , I’ll explain what persistence is, why it exists, how it works internally, and how to troubleshoot common persistence-related issues, from a Technical Support Engineer’s perspective.

 

What Is Persistence in F5 LTM?

Persistence ensures that multiple requests from the same client are sent to the same backend server for a specific duration.

Without persistence:        

            - Each request can go to a different server 

            - Many applications will break

With persistence:

            - Session data remains consistent

            - Applications behave correctly

 

Why Persistence Is Required (Real-World Examples)

Persistence is critical for applications that:

            - Store session data locally on the server

            - Don’t share session state across servers

Examples:

            - Login-based applications

            - Shopping carts

            - Banking portals

            - Legacy web apps without centralized session storage

Without persistence, users may see:

            - Random logouts

            - Session resets

            - “Invalid session” errors

 

Where Persistence Fits in Traffic Flow

Traffic flow with persistence looks like this:

Client → VIP → Persistence Lookup → Pool Member → Server

At a high level:

            1. Client sends request to VIP

            2. F5 checks if a persistence record already exists

            3. If yes → send traffic to the same server

            4. If no → select a server and create a persistence entry

 

Common Types of Persistence in F5 BIG-IP

1️. Source Address Persistence

            - Based on client IP address

            - Common for non-HTTP traffic

            - Simple but not ideal for NATed users

Example:

Client IP → Server A

Issue:

• Multiple users behind the same NAT may stick to one server

 

2️. Cookie Persistence (Most Common)

            - Used for HTTP/HTTPS applications

            - F5 inserts a cookie to identify the server

Types:

            - Cookie Insert (F5-managed)

            - Cookie Passive (Application-managed)

This is the most widely used persistence method in LTM.

 

 

3️. SSL Persistence

            - Uses SSL session ID

            - Useful when cookies are not available

            - Less common in modern deployments

 

4️. Universal Persistence (Advanced)

            - Custom persistence using iRules

            - Used for complex or non-standard applications

Example:

            - Persist based on header value, token, or custom logic

 

Persistence Profile – Where It Is Configured

Persistence is configured using a Persistence Profile, which is then attached to:

            - A Virtual Server

            - Or sometimes a Traffic Policy

Key things inside a persistence profile:

            - Timeout value

            - Cookie settings

            - Match criteria

            - Fallback options

 

Important Concept: Persistence vs Load Balancing

This is often misunderstood.

            - Load balancing decides which server gets the first request

            - Persistence decides where subsequent requests go

Once persistence is in effect:

            - Load balancing is bypassed for that session

Persistence Records (Behind the Scenes)

F5 stores persistence entries in memory, similar to connection entries.

Each record maps:

Client identifier → Pool member

This is why persistence issues can often be fixed by:

            - Clearing persistence records

            - Adjusting timeout values

 

Final Thoughts

Persistence is not just a feature — it’s a critical application requirement.

Many F5 issues that look complex become simple once you understand:

            - Why persistence exists

            - How F5 enforces it

            - How long persistence entries live