So You Actually Read Doyle and Ackers
Picking up Routing Tcp Ip Volume 2 after finishing the first volume is where most people get overwhelmed. The book jumps from basic distance-vector and link-state concepts into OSPFv3, IS-IS extensions, BGP path attributes, multicast routing, MPLS TE, and policy-based routing — all packed into about 600 pages with dense technical diagrams that look fine in print but mean nothing when you are trying to troubleshoot a misconfigured AS_PATH filter at 2 AM. I want to walk through what the book actually covers, what it leaves out, and how to use it without burning three weekends on theory that does not apply to your environment.
Routing Tcp Ip Volume 2: What It Actually Covers
The book is split into clear sections. Part one handles OSPF extensions and IPv6 OSPF. Part two covers IS-IS in more depth than most production networks need. Part three dives into BGP path selection logic and best practices for eBGP and iBGP design. Part four moves into multicast routing with PIM and MSDP, and part five wraps up with policy-based routing, route-maps, and MPLS traffic engineering basics. What most people miss is that Volume 2 is deliberately written for engineers who already understand what happens during a neighbor adjacency. If you do not know how OSPF LSA types work or why AS_PATH prepending matters, you will bounce off pages 40 through 120 before the book gets into anything useful. Do not read Volume 2 straight through. Treat it as a reference manual organized by protocol, not a narrative.
Where the Book Falls Short
The biggest limitation of Routing Tcp Ip Volume 2 is that it is old. The original editions came out before modern SD-WAN architectures, before VXLAN overlay routing became standard, and before declarative configuration platforms like Junos or Ansible changed how people deploy routing policies. The BGP chapters are solid, but they assume you are running pure iBGP with full-mesh peering or route reflectors in a way that most cloud-native deployments no longer follow. The multicast section is essentially historical reference material at this point. Very few networks today operate standalone PIM-SM infrastructure outside of specific video distribution use cases. If your team is building greenfield designs, skip straight past chapter ten and focus on the BGP and policy routing sections. There is also a serious gap around automation. The book walks you through route-maps and access-lists the old way. It does not discuss structured routing policies, BGP community-based path filtering, or integrated control-plane tools that most senior engineers rely on now.
Get the Full Details
What to Focus On First
The OSPFv3 and IPv6 chapters are worth reading if your organization is migrating away from IPv4. The IS-IS sections are only useful if you work in carrier or large enterprise core environments. For everyone else, spend your time on the BGP path selection chapter and the policy-based routing sections. BGP path selection is the part of the book that actually explains why routes behave the way they do under load. The sequence of match criteria — local preference, AS_PATH length, origin type, MED, eBGP over iBGP, IGP metric to next hop, and so on — is laid out in a way that matches real vendor implementations closely enough to be actionable. When you move into policy-based routing, the book explains route-maps well but does not make clear that most production networks use route-maps primarily for redistribution control, not for actual per-flow traffic steering. If you are trying to implement per-application routing through route-maps alone, you will hit platform limits quickly.
A Problem I Had With BGP Route Leaking
About five years ago I was redesigning a multi-homed enterprise edge and needed to leak specific prefixes from one ISP circuit into another using BGP communities. The Routing Tcp Ip Volume 2 BGP chapters describe community usage, but they do not cover the exact scenario where the receiving router strips unrecognized communities before the route enters the IGP redistribution process. My route-maps looked correct in isolation, but the prefixes were never advertised outbound because the community values that triggered the match statements were being dropped upstream. The workaround was straightforward once I figured it out. I switched from community-based route-maps to prefix-list matching for the initial filter, then used a separate community-set statement applied only after the prefix match succeeded. That kept the route in the RIB regardless of whether the downstream peer preserved the community value. It added about twenty minutes of configuration review, but it stopped the silent prefix losses immediately.
Advanced Nuance Most People Miss
One counter-intuitive point the book touches on but does not emphasize enough is that BGP best-path selection is not globally deterministic across all vendors. Cisco, Juniper, and Arista all follow the same general algorithm, but tie-breaking behavior can differ when LOCAL_PREF, AS_PATH length, and origin type are identical across multiple paths. If you are running dual-homed BGP sessions and expect symmetric return traffic without explicitly controlling LOCAL_PREF, you should test failover behavior rather than assuming the path selection will match your lab expectations. Another thing beginners consistently get wrong is the relationship between route-maps and prefix-lists in redistribution. A route-map without an explicit permit statement will deny everything that does not match a prior sequence number. This means if you configure a route-map with only deny entries for certain prefixes, all other prefixes are silently dropped during redistribution unless you add a final permit entry. The book mentions this in passing, but it causes more outages than almost any other configuration mistake I have seen.
How to Use This Book Practically
Download or borrow the book. Do not try to read it cover to cover. Open the BGP chapters first and work through the path selection examples with your own lab topology. Build a small dual-homed test environment with two routers and validate that the best-path decisions match what the book describes. Then move into policy-based routing and configure route-maps that actually filter traffic in your test lab before touching production equipment. The multicast chapters can wait until you have a documented need. The IS-IS advanced sections are only useful if you work in service provider environments. The OSPFv3 material is worth skimming even if you are not currently deploying IPv6, because the conceptual differences between OSPFv2 and OSPFv3 explain a lot about why IPv6 routing behaves differently in transition scenarios. The book is still one of the better technical references available for advanced IP routing. It just requires selective reading and hands-on validation to make it useful instead of overwhelming.
