Detailed Overview of Profiles in BIG-IP LTM

---

1. "FastHTTP" Profile

• Purpose: A combination of TCP and Limited HTTP with OneConnect to improve performance for certain HTTP traffic.

Key Features:

• OneConnect: Used to allow connection persistence on the server-side.

  • The server must support connection persistence with Keep-Alive Headers (e.g., IIS servers).

  • Faster than the standard HTTP profile due to limited features.

Limitations:

• No SSL/TLS Profile.

• No Compression support.

• No RAM Cache.

• No HTTP Pipelining.

• No IPv6 support.

• Limited iRule support (Client-Accepted, Server-Connected, HTTP-Request, HTTP-Response).

• Explicit SNAT required for client source address translation.

Usage Scenarios:

• When persistence and source IP addresses are not essential.

• When aiming for high performance and low latency without complex features like SSL or compression.

---

2. "HTTP Class" Profile

• Purpose: Allows classification of HTTP traffic based on content such as headers or URI for routing decisions.

Key Features:

• Can classify traffic based on:

  • Host Names

  • URIs

  • Headers

  • Cookies

• Supports actions like:

  • Load-balancing Pool

  • Redirection URL (Location Header)

  • Rewrite URI

Replacement:

• Replaced by the Local Traffic Policy (LTP) object starting in TMOS v11.4.

---

3. "HTTP Request Logging" Profile

• Purpose: Allows detailed logging of HTTP request and response data for monitoring and debugging.

Key Features:

• Customizable logging for Request and Response.

• Supports detailed data formats for logs, including:

  • BIG-IP Host Name

  • Client IP

  • SNAT IP

  • Virtual Server

  • User-Agent

• Available ONLY through TMM-Interfaces, not via MGMT Port.

Sample Log Entries:

Request:

BIG-IP Host Name = $BIGIP_HOSTNAME --- Date & Time = $DATE_NCSA --- Virtual Server = $VIRTUAL_NAME --- Client-IP = $CLIENT_IP --- SNAT-IP = $SNAT_IP --- Pool Member-IP:PORT = $SERVER_IP:$SERVER_PORT --- Domain Name = ${host} --- Referer = ${Referer} --- Origin = ${origin} --- User-Agent = ${User-agent} --- Request = $HTTP_REQUEST

Response:

HTTP Response Code = $HTTP_STATCODE --- Elapsed Time (ms) Between Receiving the REQ. and Sending the RES. = $RESPONSE_MSECS

4. "FastL4" Profile

• Purpose: Optimized for Layer 4 (L4) traffic where there is no need to process higher-layer protocols like HTTP.

Key Features:

• No TCP/HTTP Optimizations.

• Limited persistence options: Source address, Destination address, Universal, and Hash.

• Ideal for L4 traffic like basic TCP load balancing.

Usage Scenarios:

• Performance (L4) Virtual Servers.

• Forwarding (L2/IP) Virtual Servers (L2 forwarding not applicable).

Modes of Hardware Acceleration:

1. Full: All L4 traffic is processed by hardware, with no use of software-based features.

2. Assisted: LTM performs SYN load balancing, while hardware assists with the remaining traffic.

3. None: L4 traffic is managed entirely by BIG-IP application without hardware acceleration.

---

5. "Stream" Profile

• Purpose: Performs string matching and replacement on HTTP requests and responses.

Key Features:

• No Data Buffering.

• Searches and replaces all occurrences of a string in both requests and responses.

• Case-sensitive for pattern strings.

Usage Scenarios:

• Search and Replace on content (headers and bodies) within HTTP traffic.

• When no need for complex conditions, but a simple match/replace is required.

Dependency:

• Can be used without an HTTP profile, where it searches and replaces content in both the header and body.

• With an HTTP profile, it only searches and replaces content in the body.

---

6. "HTTP Compression" Profile

• Purpose: Compresses HTTP response content to reduce bandwidth and improve performance.

Key Features:

• Compresses HTTP responses based on Accept-Encoding header from the client.

  • Supports gzip and deflate methods.

• CPU-intensive and should be used cautiously in high-traffic environments.

Compression Levels:

• You can choose from levels 1-9 for gzip compression (level 9 is the highest compression but most CPU-intensive).

• Compression Strategy defines the compression provider:

  1. Speed (default)

  2. Size

  3. Ratio

  4. Adaptive

---

7. "Web Acceleration" Profile (RAM Cache)

• Purpose: Caches HTTP objects in memory for faster subsequent access and reduced load on backend servers.

Key Features:

• Stores static content (CSS, JS, images) for faster access.

• Content can be cached based on Vary headers or Cache-Control settings.

• Does not cache private data (e.g., responses with Cache-Control: private).

Cacheable Items:

• HTTP 200, 203, 206, 300, 301, 410 responses.

• GET methods by default (non-cacheable methods can be customized via iRules).

• Memory-intensive, so should be used for high-demand or static content.

---

Summary of Profile Features

Profile	Purpose	Key Features	Limitations
FastHTTP	High performance HTTP handling	OneConnect, TCP + limited HTTP, fast but no SSL/TLS or compression	No SSL/TLS, No Compression, No IPv6
HTTP Class	Classifies HTTP traffic	Classify traffic based on hostnames, URIs, headers, cookies	Replaced by LTP (TMOS v11.4)
HTTP Request Logging	Logs HTTP request/response data	Detailed log formatting options	Available only through TMM-Interfaces
FastL4	Optimized for Layer 4 traffic	No TCP/HTTP optimizations, limited persistence options	No Layer 7 processing
Stream	String match/replace	Search and replace on all occurrences of a string	Case-sensitive, no complex conditions
HTTP Compression	Compresses HTTP content	gzip/deflate compression, customizable levels	CPU-intensive, should be used cautiously
Web Acceleration (RAM Cache)	Caches static content	High-demand object caching, reduces backend load	Memory-intensive, only caches cacheable content