What is IPv6 Address Planning?

Prerequisites: Internet Protocol version 6 (IPv6) and what is IPv6?

A new IP protocol, called Internet Protocol (IP) version 6, is intended to replace IP version 4, which is currently in use and deployed worldwide. The latest IP version, IPv4, has proven to be reliable, easy to implement, and easy to use. It is interoperable and has passed the tests that make the internet a real global utility. However, the original design of IPv4 did not consider the following situations:

• Requirements for IP-level security. 
• The rapid expansion of the internet and depletion of IPv4 address space. 
• Need for easier auto-configuration and renumbering of network devices.
• Requirements for real-time data delivery also called quality of service.

With the help of methods such as Network Address Translation and private address spaces, IPv4’s life has been extended (NAT). While these methods appear to extend the address space and accommodate traditional client/server configurations, they do not meet the demands of increasing IP addresses. IP address translation, pooling, and temporary assignment techniques cannot be used to connect to always-on environments (such as private internet via broadband, cable modem, or DSL). Additionally, plug-and-play requirements for consumer Internet devices are in addition to address requirements. The IPv6 address space offers more addresses but should be used with caution. You can successfully deploy IPv6 using your existing IPv4 infrastructure. Converting from IP version 4 to IP version 6 is still possible today with proper planning and design.

IPv6 Address Format:

IPv6 represents a 128-bit addressing format with 16-byte hexadecimal fields separated by colons (:). This reduces the complexity and error rate of address coding. Below is an example of a legal IPv6 address: 2001:ef8:130F:0000:0000:09C0:876A:145B. In addition, IPv6 uses the following rules to compress IPv6 addresses and simplify their representation. Lead zeros in the address field are optional and can be omitted. 

Example: The following hexadecimal numbers can be represented compressed as follows:

• Example 1: 0000 is 0. (a compressed format)
• Example 2: 2001:ef8:129F:0000:0000:09C0:876A:145B = 2001:ef8:129 F:0:0:9C0:876A:145B (compressed format)

A pair of colons (::) indicates a set of zero-valued fields. However, the colon pair can only be used once in a valid IPv6 address. 

• Example 1 is 2001:ef8:129F:0:0:9C0:876A:145B = 2001:ef8:130F::9C0:876A:145B (compressed format)
• Example 2: FF01::1 = FF01::0:0:0:0:0:1 (compressed format)

Network Prefix:

Prefixes are denoted in IPv6 and roughly correspond to subnets in IPv4 terminology. The left-hand bits of the IPv6 prefix act as a network ID. Similar to how IPv4 addresses are represented in Classless Inter-Domain Routing (CIDR) notations, IPv6 prefixes are represented using the IPv6 prefix or prefix length format. The number of consecutive high-order bits that make up the prefix (the network portion of the address) is represented as a decimal value by the /prefix-length variable.  For example, the IPv6 prefix 2001:ef8:8086:6502::/64 is accepted.

IPv6 Address Types:

The IP address requirements for IPv4 hosts differ significantly from those for IPv6 hosts. An IPv6 host can have multiple IP addresses, while an IPv4 host typically uses only one. There are three main categories of IPv6 addresses:

• Unicast: Single interface address. When a packet is sent to a unicast address, the interface with that address receives it. 
• Anycast: A group of interface addresses, often belonging to different nodes. The routing protocol used determines the nearest interface over which packets sent to anycast addresses are delivered. This interface can be identified by an anycast address. 
• Multicast: An address for a collection of interfaces (within a certain range). Usually sent from different nodes. When a packet is sent to a multicast address, all interfaces listed in this range will receive it.

Why IPv6 Address Plan Required?

The following justify creating an IPv6 address plan:

• Routing tables are more compact and efficient. 
• Enforcing safety regulations can be easier. 
• You can practice application guidelines. 
• Easier network management and deployment. 
• Especially visual identification makes troubleshooting easier. 
• Scaling becomes easier when new devices and locations are added.

IPv6 Address Plan Considerations:

Due to IPv6`s substantially bigger address space than IPv4, logical and practical addressing strategies can be defined with a great deal of flexibility. You can assign subnet prefixes using a variety of logical schemes that take into account both the IP Addressing Guide’s listed factors and extra IPv6-specific considerations, such as: 

• Using already-in-use IP addressing schemes. Converting VLAN IDs into IPv6 subnet IDs and existing subnet numbers into IPv6 subnet IDs.
• Updating your IP address strategy to assign IPv6 addresses based on your requirements. You can allocate when revamping IP addressing systems in accordance with your needs. 

Such a logical addressing scheme can simplify network management, troubleshooting, operations, and service provision. Your addressing strategy should take into account the following elements:

• Prefix Aggregation: If network designers do not thoroughly pursue prefix aggregation, large IPv6 addresses may bloat routing tables.
• Network expansion: It’s crucial to consider network expansion while designing the address infrastructure. 
• Use of special local addresses (RFC4193): Like IPv4, IPv6 has a private address space. The main difference is that IPv4 allows each organization access to the same private address space. With IPv6, there is only one network and only one globally unique address space. This private address space can be used to address hardware and software not connected to the Internet.

We will look at the factors you should take into account when developing an IPv6 address plan for a service provider network. 

Address Planning for Infrastructure:

Imagine that we are organizing the addresses for the network of an ISP. We may need to consider our requirements for:

• Loopback identifiers
• Links between the two points
• Server internal networks (for NMS and other NOC servers)
• networks of external servers (for mail and DNS services, for example)

Let’s designate a /40 for backbone infrastructure for the purposes of this demonstration. All infrastructure assignments made from this block will be routed using our preferred IGP.

• Loopbacks: One /48 can be set aside for the assignment of loopback (/128) addresses. You may determine that you just need a /60 or /64 for this on smaller networks.
• Point-to-point: According to RFC 6164 and RFC 6547, we can allocate a /64 for each point-to-point link but should use a /127 (provided the hardware supports it).
• Network inside the server: If you require multiple subnets for your management server network, you can allocate from /60. /64 should work in a less complex NOC LAN.
• Network of external servers: /64 allows 264 hosts for public services.

Address Planning for Customers:

• Enterprise Customers: Consider before distributing IPv6 address blocks to clients and decide whether to aggregate them by range. To easily manage traffic per region, it might make sense to assign a /40 to each region. Corporate Banking and Enterprise customers can often get a /48 to access over 65,000 subnets. For very small consumers, some carriers allocate small blocks like /52s, /56s, or /60s. However, to give end users the freedom to assign /64s as they grow, we do not recommend assigning prefixes below /60.
• WAN link: Reserve a /48 block for customer WAN circuits for monitoring purposes. Don’t be confused with infrastructure point-to-point links. Again, it’s good practice to put /64 in your links but really use /127. As with any infrastructure connection, these should be routed through an IGP and aggregated at a gateway or POP router.
• Broadband Customers: Depending on your setup, IPv6 address blocks can be assigned to broadband subscribers. For example, SLAAC can be used to assign her WAN address from her BRAS /64 on the CPE WAN side (Stateless Address AutoConfiguration).
  We recommend that you assign at least a /40 to your broadband network and a /48 to your BRAS so that you can advertise these prefixes via BGP.
• Data Center Services: Continue to assign loopback and point-to-point connection addresses to data centers through infrastructure blocks. For hosting services, another block must be named (for example, /40). Of course, how you divide these depend on your data center design. Different /64s and VLANs can be configured for each service or customer. Datacenter border routers use IBGP to advertise these subnets to other networks. Traffic Regulation Assign customer prefixes (pulling traffic from both ends of the address space) so that incoming traffic can be balanced. You can easily achieve this by allocating a child aggregate block, like /33.