Using APA to Build a Screaming Fast Network Server Connection for HP-UX 11i v2 and v3

Using APA to Build a Screaming Fast Network Server Connections
SNSL -- Page 13 --
This illustration has several key points of failure, which are protected with a combination
of APA and the MC/SG product. The first point of failure is the switch on the left-hand
side of the illustration. The server on the left-hand side of the diagram has a logical Fail-
Over group, which consists of a Link Aggregate (connected to the switch on the left) and
a single (orange) back-up link connected to the switch on the right. If the switch on the
left-hand side fails, APA will automatically fail-over to the 1 Gb (orange) back-up link
and continue to run.
The second point of failure is the 1 Gb (orange) back-up link, which basically brings the
entire Link Aggregate down on the server on the left. At this point MC/SG will
automatically move the IP address(es) from the server on the left to the server on the
right. The server on the right has the same setup as the server on the right.
The last major point of failure is one of the servers. In this case, MC/SG will migrate all
of the applications from the failed server to the other server and continue to run.
This combination of APA and MC/SG provides mission critical customers with the
highest levels of availability in all major areas of the system and network.
1Gb Ethernet
Fail-Over
Hot Standby
1Gb Ethernet
Fail-Over
Hot Standby
MAC_
A
MAC_
B
HP 9000 w/APA &
MC/Service
Guard
HP
XP256
RAID-5
ARRAY
HP 9000 w/APA &
MC/Service
Guard
Fail-Over
Aggregat
e
8G
b
Fail-Over
Aggregat
e
8G
b
SWITC
H
SWITC
H
Gigabit
Backbone