Network Working Group R. Hinden
Request for Comments: 4311 Nokia
Updates: 2461 D. Thaler
Category: Standards Track Microsoft
IPv6 Host-to-Router Load Sharing
Status of This Memo
This document specifies an Internet standards track protocol for the
Internet community, and requests discussion and suggestions for
improvements. Please refer to the current edition of the "Internet
Official Protocol Standards" (STD 1) for the standardization state
and status of this protocol. Distribution of this memo is unlimited.
Copyright (C) The Internet Society (2005).
The original IPv6 conceptual sending algorithm does not do load
sharing among equivalent IPv6 routers, and suggests schemes that can
be problematic in practice. This document updates the conceptual
sending algorithm in RFC 2461 so that traffic to different
destinations can be distributed among routers in an efficient
In the conceptual sending algorithm in [ND] and in the optional
extension in [ROUTERSEL], a next hop is chosen when no destination
cache entry exists for an off-link destination or when communication
through an existing router is failing. Normally, a router is
selected the first time traffic is sent to a specific destination IP
address. Subsequent traffic to the same destination address
continues to use the same router unless there is some reason to
change to a different router (e.g., a redirect message is received,
or the router is found to be unreachable).
In addition, as described in [ADDRSEL], the choice of next hop may
also affect the choice of source address, and hence indirectly (and
to a lesser extent) may affect the router used for inbound traffic as
Hinden & Thaler Standards Track [Page 1]RFC 4311 IPv6 Host-to-Router Load Sharing November 2005
In both the base sending algorithm and in the optional extension,
sometimes a host has a choice of multiple equivalent routers for a
destination. That is, all other factors are equal and a host must
break a tie via some implementation-specific means.
It is often desirable when there is more than one equivalent router
that hosts distribute their outgoing traffic among these routers.
This shares the load among multiple routers and provides better
performance for the host's traffic.
On the other hand, load sharing can be undesirable in situations
where sufficient capacity is available through a single router and
the traffic patterns could be more predictable by using a single
router; in particular, this helps to diagnose connectivity problems
beyond the first-hop routers.
[ND] does not require any particular behavior in this respect. It
specifies that an implementation may always choose the same router
(e.g., the first in the list) or may cycle through the routers in a
round-robin manner. Both of these suggestions are problematic.
Clearly, always choosing the same router does not provide load
sharing. Some problems with load sharing using naive tie-breaking
techniques such as round-robin and random are discussed in
[MULTIPATH]. While the destination cache provides some stability
since the determination is not per packet, cache evictions or
timeouts can still result in unstable or unpredictable paths over
time, lowering the performance and making it harder to diagnose
problems. Round-robin selection may also result in synchronization
issues among hosts, where in the worst case the load is concentrated
on one router at a time.
In the remainder of this document, the key words "MUST", "MUST NOT",
"REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT",
"RECOMMENDED", "MAY", and "OPTIONAL" are to be interpreted as
described in [RFC2119].
2. Load Sharing
When a host chooses from multiple equivalent routers, it SHOULD
support choosing using some method that distributes load for
different destinations among the equivalent routers rather than
always choosing the same router (e.g., the first in the list). This
memo takes no stance on whether the support for load sharing should
be turned on or off by default. Furthermore, a host that does
attempt to distribute load among routers SHOULD use a hash-based
scheme that takes (at least) the destination IP address into account,