Backup Router Protocol VRRP.

In the previous article, we touched on the problem of redundant network infrastructure and examined how the technology works MC-LAG. Today we tell about another technology, more complex and more reliable and enables you to reserve at the level of the data center. We are talking about increasing the availability of routers using the protocol VRRP (Virtual Router Redundancy Protocol).
The article will be of interest primarily to those who would like to ensure continuous availability of business-relevant Internet resources, but it does not have sufficient technical capacity: has neither its own stand-alone system or block IP-addresses or connections to providers via BGP.

Choosing a backup scheme
Imagine that we have a business critical Internet resource that must always be available to a large number of users.

Network configuration resource (address, mask and default gateway) are as follows:
ifconfig eth0 address 12.34.56.78 mask 255.255.255.248 gw 12.34.56.73

If .78 - is the address of the host, the .73 - the default gateway address. This address - area of ​​responsibility of the operator, and if the host is placed in the data center - the area of ​​responsibility of the data center. Graphically, this scheme can be represented as follows:

In the final host prescribed address 12.34.56.78, the router - .72, and between organized unified L2-domain (usually a separate VLAN):

To increase the availability of the final host, the reservation is required a network infrastructure.
For redundancy at L2 in the simplest case, a Virtual Chassis / Fabric / MC-LAG. Then of course the host network connects data centers with LAG (Etherchannel):

Possible points of failure are very destination host and a router.
Reservation final host - it is the responsibility of the customer. It is very desirable to reserve the host end and are located in different data centers. This will avoid many problems (with the network structure, the availability of a particular physical server, with power and cooling in individual markets).

To organize the transfer of IP-addresses between the primary and secondary hosts can be different within one segment of the L2-it can be done using protocols CARP / HSRP / VRRP and their analogues:

The end customer and standby hosts are located in different data centers. Routers that belong to the operator, are also located in different data centers. Data centers can be connected to multiple channels of communication.

In the event of a fault in one of the data centers host the final still remains affordable. The described approach can be used for both redundancy L2-, and L3-on circuits.

Backup Routers
An example of a reservation at L3 can serve anycast-routing and the use of the BGP speaker with the higher operator. Each host on the router advertises network operator 12.34.56.72/29 with a different priority. In addition, each host is connected to the router operator separate subnet, a separate VLAN'om.