In this mode, each of multiple NIC (Network Interface Card) is connected to a different network and all these NICs are activated.
Just as in Fast switching mode, a virtual interface is generated and a virtual network is allocated to this interface. A TCP/IP application can conduct communication with the remote system, irrespective of the physical network redundant configuration, by using an IP address (called a virtual IP address in this document) set in this virtual interface as its own local system IP address.
The lines are monitored in accordance with the standard protocol on the Internet RIP (Routing Information Protocol). RIP is controlled by routing daemons (in.routed) on the Solaris system. The version of the routing daemons supported by the Solaris system is version 1.
Figure 2.6 Example of duplicated operation in RIP mode
Routers are placed between systems to enable communicate between them, with each communication route comprising a different network.
Because the Internet standard routing protocol RIP is used, communication can be carried out with a variety of devices in a global network environment regardless of the models. However, because the path switching by RIP is performed slowly, switching requires some time.
This mode is appropriate, for example, for the WEB server and communications between the application server and client machines in a three-tier client-server system.
Figure 2.7 System configuration for RIP mode shows a system configuration for RIP mode:
Figure 2.7 System configuration for RIP mode
The following explains each component and its meaning:
Indicates a physical interface (such as hme0 and hme1) of the duplicated NIC.
Indicates an IP address attached to a physical interface. This IP address is always active. IPv4 address can be used for a physical interface.
Indicates a virtual interface (such as sha0) so that duplicated NIC can be seen as one NIC.
Indicates a local IP address to be allocated to the virtual interface for communication with remote devices. IPv4 address can be used for a physical interface.
Indicates the IP address of a router to be monitored first when the router monitoring function is used.
Indicates the IP address of a router to be monitored after switching.
The shortest path to the remote system is selected based on the RIP packet received from the neighboring router and the selected path is used for communication. Then, monitoring is carried out to check whether any RIP packet is received from the router. If a RIP packet is normally received, the transmission line is considered to be normal. If no RIP packet is received within a specified period of time, the transmission line is considered to be faulty and the line to be used for communication is switched in accordance with the routing information received from another router. Monitoring is carried out for each router connected to NIC. Routing control via RIP is performed by the Solaris system.
Figure 2.8 Monitoring method in RIP mode (when the router monitoring function is not used)
If a failure occurs in a line, up to five minutes are required to switch the network paths via RIP.
The following failures can be detected:
Figure 2.9 Effective monitoring range in RIP mode
Because the failures in (1) to (4) appear to be the same failure, it is not possible to determine under which of the four failure types these failures should be classified. Each device has to be checked to make this determination.
Monitoring is started automatically when the virtual interface is activated. Monitoring is automatically stopped when the virtual interface is inactivated. In cluster operation, monitoring is started or stopped along with the start or stop of a RMS.
The line is switched for use in communication in accordance with the routing information received from a router that is different from the router from which RIP was received.
Figure 2.10 Outline of switching operation performed when a failure occurred in RIP mode
If a faulty line is recovered, the path is automatically restored to its original status in accordance with the RIP information. The failback of line cannot be performed manually.
Any system can be connected. However, the Fujitsu LINKRELAY Series is recommended as the router to be connected to the local system network.
The requirement for user applications that can be operated in this mode is as follows:
Applications must be operational on a system to which multiple NICs are connected and on which multiple IP addresses are defined (This system is called a multi-home host). For example, a socket application needs to operate with its local IP address fixed with the bind function or set to any value (Applications of the remote party do not check the IP address).
IPv4 address must be assigned to the physical interface.
Only one machine should run on one network in RIP mode. If RIP is sent from more than one server, the propagation of path information becomes complicated and more time is required for switching than expected.
No subnet can be created for a network to be used. Be sure to directly use a network of class A, B, or C without specifying a subnet mask. However, a subnet mask can be specified if the following conditions are met:
A subnet is created only for one network address.
A unique value in the entire network must be specified for the subnet mask for the network address for which a subnet is created.
A subnet mask value of the network address is defined in the /etc/netmasks file.
It is not possible to use an IPv6 address.
If you are using Solaris10 for OS, use the "routeadm(1M)" command to set up a routing daemon. See "3.2.2.3 System setup in RIP mode".
