Why does a fixed-length IPv6 base header make routing faster, and how can it still be extended?

A fixed 40-byte header lets routers parse fields at known offsets with no options to scan; extra functionality goes into optional extension headers chained via the Next Header field.

Why fixed length helps:

• IPv4 had a variable header with an Internet Header Length field and inline options — routers had to compute where the data started every time
• IPv6's header is always 40 bytes, so a router reads fields at constant positions → faster, simpler, even feasible in cheap hardware

How it stays flexible:

• The Next Header field points to either an upper-layer protocol (TCP/UDP) or the first extension header
• Extension headers form a chain, each with its own Next Header pointing to the next one
• Routers can mostly skip extension headers they don't need to process

Tip: IPv6 moved options out of the mandatory header and into an optional chain — you pay the parsing cost only when a feature is actually used.

Go deeper:

• IPv6 packet — Extension headers (Wikipedia) — the fixed 40-byte header plus the Next-Header chaining that keeps it extensible.
• RFC 8200 §4 — IPv6 Extension Headers — how optionality moved into the chained extension headers.