Technology - IP Techniques 1 - IPv4 & Packet Classification

Synopsis

This report discusses the 32 bit addressing system of the IPv4 Internet, including private address spaces, Classless Inter-Domain Routing (CIDR), utilisation of IPv4 address space and the projected end of fresh space around 2011. We also discuss the two primary techniques used in router Forwarding Information Bases (FIBs) for classifying packets: TCAM (Ternary Content Addressable Memory) and ASIC-based memory lookup algorithms, such as Tree-Bitmap.

Table of Contents

• 1. Synopsis
• 2. Introduction
• 3. IPv4 - the current TCP/IP version
  • 3.1 Address space
    • 3.1.1 Subnets and prefixes
    • 3.1.2 Classless Inter-Domain Routing (CIDR)
  • 3.2 Merging networks with private addresses
    • 3.2.1 IPv6’s improvements
  • 3.3 IPv4 address utilisation
    • 3.3.1 Conserving IP addresses for webservers
    • 3.3.2 Private network addresses
    • 3.3.3 Estimating IP address utilisation
    • 3.3.4 Assigned addresses
• 4. Packet Classification
  • 4.1 The workload of high end routers
    • 4.1.1 Reliability, redundancy and power requirements
    • 4.1.2 Classification in minimum number of clock cycles
  • 4.2 Classification tasks for IP packets
  • 4.3 ASIC-based packet classification
    • 4.3.1 Juniper’s M120 and MX960
    • 4.3.2 Cisco’s CRS-1
  • 4.4 TCAM-based packet classification
    • 4.4.1 Memory cell functions as a comparator
    • 4.4.2 Classifying Internet packets
    • 4.4.3 More elaborate classification
• 5. Related reports