DECnet Phase Routing: Evolution, Architecture, and Protocols

DECnet, developed by Digital Equipment Corporation (DEC), represents one of the most significant milestones in the history of peer-to-peer networking. While modern networking is dominated by the TCP/IP suite, the innovations introduced in DECnet Phase Routing—particularly in Phase IV and Phase V—laid the conceptual groundwork for hierarchical routing, link-state protocols, and the integration of the OSI model.

This guide provides an in-depth exploration of DECnet Phase Routing, focusing on how Digital Network Architecture (DNA) evolved to manage complex, multi-area networks.

Evolution of DECnet Routing

The history of DECnet is a journey from simple point-to-point connections to a sophisticated global networking architecture.

Early Phases (Phase I - Phase III)

• Phase I & II: Routing was non-existent in the modern sense. Phase II allowed for up to 32 nodes but only supported point-to-point communication. If Node A wanted to talk to Node C, it had to be directly connected.
• Phase III (1980): Introduced Adaptive Routing. This allowed for networks of up to 255 nodes. For the first time, packets could travel through intermediate nodes (routers) to reach their destination.

Phase IV: The Golden Age of DECnet (1982)

Phase IV was the breakthrough that enabled massive scale. It expanded the address space to 16 bits, allowing for 64,449 nodes. To manage this scale, DEC introduced Hierarchical Routing, dividing the network into 63 areas, each containing up to 1023 nodes.

Phase V: DECnet Plus (1987)

Phase V, also known as DECnet/OSI, marked a shift toward international standards. It replaced the proprietary routing protocols with the OSI IS-IS (Intermediate System to Intermediate System) protocol, allowing for architecturally unlimited network sizes and multi-vendor interoperability.

Talk to a Network Expert

DECnet Phase IV Routing Architecture

To understand DECnet routing, one must look at the specific node types and the hierarchical structure defined in Phase IV.

The Hierarchical Structure: Areas and Nodes

In Phase IV, a DECnet address consists of two parts:

1. Area Number (6 bits): Ranges from 1 to 63.
2. Node Number (10 bits): Ranges from 1 to 1023.

This hierarchy is essential for reducing the size of routing tables. Instead of every router knowing the path to every node in the entire network, they only need detailed information about their local area.

DECnet Node Types

DECnet distinguishes between two primary functional roles for devices:

• End Nodes (Level 0): These are leaf devices (like VAX terminals or workstations). They can send and receive data but cannot forward traffic for other nodes.
• Level 1 Routers (L1): These routers manage traffic within a single area. They know the shortest path to every node in their specific area. If they receive a packet destined for another area, they simply pass it to the nearest Level 2 router.
• Level 2 Routers (L2): These are the backbone routers. They manage traffic between areas. An L2 router maintains a routing table for all 63 possible areas but does not need to know the specific node addresses inside those other areas.

Also Read: How ARP Connects Your IP Address to a Local Network Identity

DECnet Routing Protocol (DRP) Mechanism

Phase IV utilized a distance-vector routing algorithm, often compared to a precursor of RIP (Routing Information Protocol) but optimized for the DNA environment.

Shortest Path Selection: Cost vs. Hops

DECnet routing is based on a Least Cost Path algorithm.

• Cost: An arbitrary value (1 to 63) assigned to a circuit (link). Administrators typically assign lower costs to high-speed links (like Ethernet) and higher costs to slower serial links.
• Hops: The number of intermediate routers a packet must pass through.

The routing algorithm calculates the path with the lowest cumulative cost. If two paths have the same cost, the one with the fewer hops is selected.

Hello Messages and Adjacency

DECnet nodes use "Hello" messages to maintain awareness of their neighbors:

• End Node Hellos: Sent by end nodes to let the local router know they are active.
• Router Hellos: Sent by Level 1 and Level 2 routers to advertise their presence and status to other routers and end nodes.

Phase V: Transition to Link-State and IS-IS

DECnet Phase V (DECnet-Plus) was a radical redesign. It moved away from the distance-vector model of Phase IV in favor of Link-State Routing via the ISO IS-IS protocol.

Why the Shift?

Distance-vector protocols suffer from "slow convergence" and are prone to routing loops (the "count to infinity" problem). Phase V addressed this by:

1. Map-Based Routing: Every router maintains a complete "map" (Link State Database) of the network topology.
2. Dijkstra’s Algorithm: Each node independently calculates the shortest path using the Shortest Path First (SPF) algorithm, resulting in much faster updates when a link goes down.
3. OSI Addressing: Phase V adopted the NSAP (Network Service Access Point) addressing format, which is much more flexible than the fixed 16-bit Phase IV addresses.

Also Read: Subnetting Techniques: Guide to IP Network Efficiency

Technical Specifications: The DECnet Data Packet

A DECnet routing header contains several critical fields that ensure the delivery of data:

• Flags: Indicate the packet format (long vs. short) and whether "Return to Sender" is requested.
• Destination/Source Address: The 16-bit (Phase IV) or NSAP (Phase V) addresses.
• Visit Count: Similar to the TTL (Time to Live) in IP packets. Each time a packet passes through a router, the visit count increments. If it exceeds a set threshold (usually 16 or 32), the packet is discarded to prevent infinite looping.

Conclusion

DECnet Phase Routing represents a masterclass in early network engineering. By implementing hierarchical areas and sophisticated cost-based routing, Digital Equipment Corporation created a system that was decades ahead of its time. Whether you are studying the history of the OSI model or the mechanics of distance-vector algorithms, understanding DECnet provides a foundational perspective on how the global internet we use today was built.

Contact our Network Team Today

Key Takeaways

• Hierarchical Efficiency: DECnet Phase IV introduced a two-tier hierarchy (Level 1 and Level 2) that allowed for early large-scale networking without overloading router memory.
• Cost-Based Decisions: Routing is determined by cumulative link cost, allowing administrators to influence traffic flow based on bandwidth.
• Architectural Evolution: The transition from Phase IV (Distance Vector) to Phase V (Link-State/IS-IS) mirrored the broader industry shift toward more robust, faster-converging protocols like OSPF.
• Legacy Impact: Many concepts used in modern IS-IS (used extensively in ISP backbones today) originated or were refined during the development of DECnet Phase V.

Frequently Asked Questions (FAQs)

What is the maximum number of nodes in a DECnet Phase IV network?

A Phase IV network can support 64,449 nodes, structured as 63 areas with 1023 nodes per area.

What is the difference between a Level 1 and a Level 2 router?

A Level 1 router handles traffic within a single area. A Level 2 router acts as a gateway between different areas and handles the inter-area backbone traffic.

How does DECnet handle a link failure?

In Phase IV, routers exchange updated routing vectors periodically. When a link fails, the "cost" to reach those nodes becomes infinite, and routers eventually calculate an alternative path. In Phase V, the SPF algorithm triggers an almost instantaneous recalculation.

Is DECnet still used today?

While largely replaced by TCP/IP, DECnet remains in use in specific legacy industrial systems, some air traffic control environments, and by hobbyist communities (like the HECnet) using emulated VAX hardware.