TCP/IPv6 Configuration and Management Manual (G06.29+, H06.03+, J06.03+)

Manage the NonStop TCP/IPv6 Subsystem
HP NonStop TCP/IPv6 Configuration and Management Manual—524523-012
6-3
Fault-Tolerant Operations, Ethernet Failover
Guidelines
The processes sharing the UDP port should not maintain a context for previous
messages because a sequence of messages might not be delivered to the same
socket if the port is shared. In fact, with round-robin enabled, a sequence of messages
is distributed to each of the port-sharing sockets, in turn.
If, for example, an application assumes that all packets from a given source will be
directed to it (the application process), the application must assume that it is the only
instance of itself on that UDP port. If another instance of itself is sharing the UDP port,
packets from the same source could go to two different instances of the application
process resulting in one of the application processes missing some of the packets
destined for it.
To run multiple instances in parallel for applications that must in the course of normal
operation, maintain a context across multiple received messages, you can circumvent
the problem introduced by round-robin filtering by changing the application to bind to
an IP address. Changing the application to bind to an IP address allows one instance
of the application for each SUBNET to be supported by NonStop TCP/IPv6, while still
sharing the same port number. NonStop TCP/IPv6 distributes incoming packets that
come in from one SUBNET only to the application bound to that SUBNET. Therefore,
by binding the application to the IP address, the packets received by the application
retain their contexts. This solution circumvents the problem introduced by round-robin
distribution of incoming packets among sockets sharing the same port.
Fault-Tolerant Operations, Ethernet Failover Guidelines
NonStop TCP/IPv6 provides failover at the adapter level with a feature called Ethernet
failover.
Unlike other failover implementations, NonStop TCP/IPv6 Ethernet failover does not
require one of the LIFs to act as a “hot standby” in anticipation of a failure. Both LIFs
are active and allow inbound and outbound network traffic to be distributed between
them. Therefore, you gain scalability when all your adapters are performing correctly.
In NonStop TCP/IPv6, a total of 64 SUBNETs can be configured. To achieve a failover
configuration, two SUBNETs are associated as a failover pair, so a maximum of 32
failover pairs can be configured.
With Ethernet failover, a single IP address can be shared between both LIFs (referred
to as shared IP), or each LIF can be configured to have its own IP address (referred to
as nonshared IP). HP recommends the shared IP configuration for most cases
because shared IP provides a bandwidth advantage for outbound traffic. With shared
IP, outbound traffic can flow over either SUBNET, and this effectively doubles the
throughput capacity for outbound traffic. New connections are distributed across both
adapters, but won't necessarily be distributed in a balanced manner. The benefits of
shared IP are increased bandwidth for outbound traffic and flexibility for inbound traffic.
Nonshared IP can provide the extra bandwidth for outbound traffic that shared IP
provides if you add every ROUTE twice, once per SUBNET, and if the application has
selected INADDR_ANY as a source IP address (allowing NonStop TCP/IPv6 to choose
the interface to assign to it). If the application binds to a source IP address, NonStop