In the previous example, the Internet external network is assigned an entire class C, /24 subnet of which only the last octet is shown. The next-hop gateway uses the highest address in the range (.254) and, because a physical device can run an availability protocol across redundant hardware, additional addresses (.253 and .252 in the example). Tenant Edge Services Gateways are directly connected to the network and have their interface addresses allocated from .201 upwards. This allows approximately 50 tenants to be allocated an Edge Services Gateway address from the remainder of higher end addresses on the subnet. With 50 tenants and the next-hop gateway allocated from the addresses in the /24 subnet, the “red” network still has all the addresses from .1 to .200 available for customers to use for services they want to be accessible from the Internet. These addresses can be assigned to customers in ranges of any given size using a method known as sub-allocation. See Section 5.4, External Network Address Sub-Allocation for further information on the sub-allocation process. Each tenant must also have the interface address of their Edge Services Gateway sub-allocated to itself, so that it too can be used for tenant services. See Section 5.3.3, Outbound Internet Access for the use of the interface address for outbound Internet access.