Classless Inter-Domain Routing

(Redirected from /32)

Classless Inter-Domain Routing (CIDR /ˈsdər, ˈsɪ-/) is a method for allocating IP addresses for IP routing. The Internet Engineering Task Force introduced CIDR in 1993 to replace the previous classful network addressing architecture on the Internet. Its goal was to slow the growth of routing tables on routers across the Internet, and to help slow the rapid exhaustion of IPv4 addresses.[1][2]

IP addresses are described as consisting of two groups of bits in the address: the most significant bits are the network prefix, which identifies a whole network or subnet, and the least significant set forms the host identifier, which specifies a particular interface of a host on that network. This division is used as the basis of traffic routing between IP networks and for address allocation policies.

Whereas classful network design for IPv4 sized the network prefix as one or more 8-bit groups, resulting in the blocks of Class A, B, or C addresses, under CIDR address space is allocated to Internet service providers and end users on any address-bit boundary. In IPv6, however, the interface identifier has a fixed size of 64 bits by convention, and smaller subnets are never allocated to end users.

CIDR is based on variable-length subnet masking (VLSM), in which network prefixes have variable length (as opposed to the fixed-length prefixing of the previous classful network design). The main benefit of this is that it grants finer control of the sizes of subnets allocated to organizations, hence slowing the exhaustion of IPv4 addresses from allocating larger subnets than needed. CIDR gave rise to a new way of writing IP addresses known as CIDR notation, in which an IP address is followed by a suffix indicating the number of bits of the prefix. Some examples of CIDR notation are the addresses 192.0.2.0/24 for IPv4 and 2001:db8::/32 for IPv6. Blocks of addresses having contiguous prefixes may be aggregated as supernets, reducing the number of entries in the global routing table.

Background

edit

Each IP address consists of a network prefix followed by a host identifier. In the classful network architecture of IPv4, the three most significant bits of the 32-bit IP address defined the size of the network prefix for unicast networking, and determined the network class A, B, or C.[3]

Class Most-significant bits Network prefix size (bits) Host identifier size (bits) Address range
A 0 8 24 0.0.0.0–127.255.255.255
B 10 16 16 128.0.0.0–191.255.255.255
C 110 24 8 192.0.0.0–223.255.255.255
D(multicast)
E(reserved)
1110
1111
224.0.0.0–255.255.255.255

The advantage of this system is that the network prefix could be determined for any IP address without any further information. The disadvantage is that networks were usually too big or too small for most organizations to use, because only three sizes were available. The smallest allocation and routing block contained 28 = 256 addresses, larger than necessary for personal or department networks, but too small for most enterprises. The next larger block contained 216 = 65536 addresses, too large to be used efficiently even by large organizations. But for network users who needed more than 65536 addresses, the only other size (224) provided far too many, more than 16 million. This led to inefficiencies in address use as well as inefficiencies in routing, because it required a large number of allocated class-C networks with individual route announcements, being geographically dispersed with little opportunity for route aggregation.

Within a decade after the invention of the Domain Name System (DNS), the classful network method was found not scalable.[4] This led to the development of subnetting and CIDR. The formerly meaningful class distinctions based on the most-significant address bits were abandoned and the new system was described as classless, in contrast to the old system, which became known as classful. Routing protocols were revised to carry not just IP addresses, but also their subnet masks. Implementing CIDR required every host and router on the Internet to be reprogrammed in small ways—no small feat at a time when the Internet was entering a period of rapid growth. In 1993, the Internet Engineering Task Force published a new set of standards, RFC 1518 and RFC 1519, to define this new principle for allocating IP address blocks and routing IPv4 packets. An updated version, RFC 4632, was published in 2006.[5]

After a period of experimentation with various alternatives, Classless Inter-Domain Routing was based on variable-length subnet masking (VLSM), which allows each network to be divided into subnetworks of various power-of-two sizes, so that each subnetwork can be sized appropriately for local needs. Variable-length subnet masks were mentioned as one alternative in RFC 950.[6] Techniques for grouping addresses for common operations were based on the concept of cluster addressing, first proposed by Carl-Herbert Rokitansky.[7][8]

CIDR notation

edit

CIDR notation is a compact representation of an IP address and its associated network mask. The notation was invented by Phil Karn in the 1980s.[9][10] CIDR notation specifies an IP address, a slash ('/') character, and a decimal number. The decimal number is the count of consecutive leading 1-bits (from left to right) in the network mask. Each 1-bit denotes a bit of the address range which must remain identical to the given IP address. The IP address in CIDR notation is always represented according to the standards for IPv4 or IPv6.

The address may denote a specific interface address (including a host identifier, such as 10.0.0.1/8), or it may be the beginning address of an entire network (using a host identifier of 0, as in 10.0.0.0/8 or its equivalent 10/8). CIDR notation can even be used with no IP address at all, e.g. when referring to a /24 as a generic description of an IPv4 network that has a 24-bit prefix and 8-bit host numbers.

For example:

  • 198.51.100.14/24 represents the IPv4 address 198.51.100.14 and its associated network prefix 198.51.100.0, or equivalently, its subnet mask 255.255.255.0, which has 24 leading 1-bits.
  • the IPv4 block 198.51.100.0/22 represents the 1024 IPv4 addresses from 198.51.100.0 to 198.51.103.255.
  • the IPv6 block 2001:db8::/48 represents the block of IPv6 addresses from 2001:db8:0:0:0:0:0:0 to 2001:db8:0:ffff:ffff:ffff:ffff:ffff.
  • ::1/128 represents the IPv6 loopback address. Its prefix length is 128 which is the number of bits in the address.

In IPv4, CIDR notation came into wide use only after the implementation of the method, which was documented using dotted-decimal subnet mask specification after the slash, for example, 192.24.12.0/255.255.252.0.[2] Describing the network prefix width as a single number (192.24.12.0/22) was easier for network administrators to conceptualize and to calculate. It became gradually incorporated into later standards documents[11][12] and into network configuration interfaces.

The number of addresses of a network may be calculated as 2address length − prefix length, where address length is 128 for IPv6 and 32 for IPv4. For example, in IPv4, the prefix length /29 gives: 232−29 = 23 = 8 addresses.

Subnet masks

edit

A subnet mask is a bitmask that encodes the prefix length associated with an IPv4 address or network in quad-dotted notation: 32 bits, starting with a number of 1-bits equal to the prefix length, ending with 0-bits, and encoded in four-part dotted-decimal format: 255.255.255.0. A subnet mask encodes the same information as a prefix length but predates the advent of CIDR. In CIDR notation, the prefix bits are always contiguous. Subnet masks were allowed by RFC 950[6] to specify non-contiguous bits until RFC 4632[5]: Section 5.1  stated that the mask must be left contiguous. Given this constraint, a subnet mask and CIDR notation serve exactly the same function.

CIDR blocks

edit

CIDR is principally a bitwise, prefix-based standard for the representation of IP addresses and their routing properties. It facilitates routing by allowing blocks of addresses to be grouped into single routing table entries. These groups, commonly called CIDR blocks, share an initial sequence of bits in the binary representation of their IP addresses. IPv4 CIDR blocks are identified using a syntax similar to that of IPv4 addresses: a dotted-decimal address, followed by a slash, then a number from 0 to 32, i.e., a.b.c.d/n. The dotted decimal portion is the IPv4 address. The number following the slash is the prefix length, the number of shared initial bits, counting from the most-significant bit of the address. When emphasizing only the size of a network, the address portion of the notation is usually omitted. Thus, a /20 block is a CIDR block with an unspecified 20-bit prefix.

An IP address is part of a CIDR block and is said to match the CIDR prefix if the initial n bits of the address and the CIDR prefix are the same. An IPv4 address is 32 bits so an n-bit CIDR prefix leaves 32 − n bits unmatched, meaning that 232−n IPv4 addresses match a given n-bit CIDR prefix. Shorter CIDR prefixes match more addresses, while longer prefixes match fewer. In the case of overlaid CIDR blocks, an address can match multiple CIDR prefixes of different lengths.

CIDR is also used for IPv6 addresses and the syntax semantic is identical. The prefix length can range from 0 to 128, due to the larger number of bits in the address. However, by convention, a subnet on broadcast MAC layer networks always has 64-bit host identifiers.[13] Larger prefixes (/127) are only used on some point-to-point links between routers, for security and policy reasons.[14]

 

Assignment of CIDR blocks

edit

The Internet Assigned Numbers Authority (IANA) issues to regional Internet registries (RIRs) large, short-prefix CIDR blocks. However, a /8 (with over sixteen million addresses) is the largest block IANA will allocate. For example, 62.0.0.0/8 is administered by RIPE NCC, the European RIR. The RIRs, each responsible for a single, large, geographic area, such as Europe or North America, subdivide these blocks and allocate subnets to local Internet registries (LIRs). Similar subdividing may be repeated several times at lower levels of delegation. End-user networks receive subnets sized according to their projected short-term need. Networks served by a single ISP are encouraged by IETF recommendations to obtain IP address space directly from their ISP. Networks served by multiple ISPs, on the other hand, may obtain provider-independent address space directly from the appropriate RIR.

 

For example, in the late 1990s, the IP address 208.130.29.33 (since reassigned) was used by www.freesoft.org. An analysis of this address identified three CIDR prefixes. 208.128.0.0/11, a large CIDR block containing over 2 million addresses, had been assigned by ARIN (the North American RIR) to MCI. Automation Research Systems (ARS), a Virginia VAR, leased an Internet connection from MCI and was assigned the 208.130.28.0/22 block, capable of addressing just over 1000 devices. ARS used a /24 block for its publicly accessible servers, of which 208.130.29.33 was one. All of these CIDR prefixes would be used, at different locations in the network. Outside MCI's network, the 208.128.0.0/11 prefix would be used to direct to MCI traffic bound not only for 208.130.29.33, but also for any of the roughly two million IP addresses with the same initial 11 bits. Within MCI's network, 208.130.28.0/22 would become visible, directing traffic to the leased line serving ARS. Only within the ARS corporate network would the 208.130.29.0/24 prefix have been used.

IPv4 CIDR blocks

edit
Address
format
Difference
to last address
Mask Addresses Relative
to class
A, B, C
Restrictions
on a, b, c and d
(0..255 unless noted)
Typical use
Decimal 2n
a.b.c.d/32 +0.0.0.0 255.255.255.255 1 20 1256 C Host route
a.b.c.d/31 +0.0.0.1 255.255.255.254 2 21 1128 C d = 0 ... (2n) ... 254 Point-to-point links (RFC 3021)
a.b.c.d/30 +0.0.0.3 255.255.255.252 4 22 164 C d = 0 ... (4n) ... 252 Point-to-point links (glue network)
a.b.c.d/29 +0.0.0.7 255.255.255.248 8 23 132 C d = 0 ... (8n) ... 248 Smallest multi-host network
a.b.c.d/28 +0.0.0.15 255.255.255.240 16 24 116 C d = 0 ... (16n) ... 240 Small LAN
a.b.c.d/27 +0.0.0.31 255.255.255.224 32 25 18 C d = 0 ... (32n) ... 224
a.b.c.d/26 +0.0.0.63 255.255.255.192 64 26 14 C d = 0, 64, 128, 192
a.b.c.d/25 +0.0.0.127 255.255.255.128 128 27 12 C d = 0, 128 Large LAN
a.b.c.0/24 +0.0.0.255 255.255.255.0 256 28 1 C
a.b.c.0/23 +0.0.1.255 255.255.254.0 512 29 2 C c = 0 ... (2n) ... 254
a.b.c.0/22 +0.0.3.255 255.255.252.0 1,024 210 4 C c = 0 ... (4n) ... 252 Small business
a.b.c.0/21 +0.0.7.255 255.255.248.0 2,048 211 8 C c = 0 ... (8n) ... 248 Small ISP/ large business
a.b.c.0/20 +0.0.15.255 255.255.240.0 4,096 212 16 C c = 0 ... (16n) ... 240
a.b.c.0/19 +0.0.31.255 255.255.224.0 8,192 213 32 C c = 0 ... (32n) ... 224 ISP/ large business
a.b.c.0/18 +0.0.63.255 255.255.192.0 16,384 214 64 C c = 0, 64, 128, 192
a.b.c.0/17 +0.0.127.255 255.255.128.0 32,768 215 128 C c = 0, 128
a.b.0.0/16 +0.0.255.255 255.255.0.0 65,536 216 256 C = B
a.b.0.0/15 +0.1.255.255 255.254.0.0 131,072 217 2 B b = 0 ... (2n) ... 254
a.b.0.0/14 +0.3.255.255 255.252.0.0 262,144 218 4 B b = 0 ... (4n) ... 252
a.b.0.0/13 +0.7.255.255 255.248.0.0 524,288 219 8 B b = 0 ... (8n) ... 248
a.b.0.0/12 +0.15.255.255 255.240.0.0 1,048,576 220 16 B b = 0 ... (16n) ... 240
a.b.0.0/11 +0.31.255.255 255.224.0.0 2,097,152 221 32 B b = 0 ... (32n) ... 224
a.b.0.0/10 +0.63.255.255 255.192.0.0 4,194,304 222 64 B b = 0, 64, 128, 192
a.b.0.0/9 +0.127.255.255 255.128.0.0 8,388,608 223 128 B b = 0, 128
a.0.0.0/8 +0.255.255.255 255.0.0.0 16,777,216 224 256 B = A Largest IANA block allocation
a.0.0.0/7 +1.255.255.255 254.0.0.0 33,554,432 225 2 A a = 0 ... (2n) ... 254
a.0.0.0/6 +3.255.255.255 252.0.0.0 67,108,864 226 4 A a = 0 ... (4n) ... 252
a.0.0.0/5 +7.255.255.255 248.0.0.0 134,217,728 227 8 A a = 0 ... (8n) ... 248
a.0.0.0/4 +15.255.255.255 240.0.0.0 268,435,456 228 16 A a = 0 ... (16n) ... 240
a.0.0.0/3 +31.255.255.255 224.0.0.0 536,870,912 229 32 A a = 0 ... (32n) ... 224
a.0.0.0/2 +63.255.255.255 192.0.0.0 1,073,741,824 230 64 A a = 0, 64, 128, 192
a.0.0.0/1 +127.255.255.255 128.0.0.0 2,147,483,648 231 128 A a = 0, 128
0.0.0.0/0 +255.255.255.255 0.0.0.0 4,294,967,296 232 256 A Entire IPv4 Internet, default route.

In common usage, the first address in a subnet, all binary zero in the host identifier, is reserved for referring to the network itself, while the last address, all binary one in the host identifier, is used as a broadcast address for the network; this reduces the number of addresses available for hosts by 2. As a result, a /31 network, with one binary digit in the host identifier, would be unusable, as such a subnet would provide no available host addresses after this reduction. RFC 3021 creates an exception to the "host all ones" and "host all zeros" rules to make /31 networks usable for point-to-point links. /32 addresses (single-host network) must be accessed by explicit routing rules, as there is no room in such a network for a gateway.

In routed subnets larger than /31 or /32, the number of available host addresses is usually reduced by two, namely the largest address, which is reserved as the broadcast address, and the smallest address, which identifies the network itself.[15][16]

IPv6 CIDR blocks

edit
IPv6 CIDR prefixes
Prefix size Number of equivalent subnets Interface ID bits
/48 /56 /64
/24 16M 4G 1T 104
/25 8M 2G 512G 103
/26 4M 1G 256G 102
/27 2M 512M 128G 101
/28 1M 256M 64G 100
/29 512K 128M 32G 99
/30 256K 64M 16G 98
/31 128K 32M 8G 97
/32 64K 16M 4G 96
/33 32K 8M 2G 95
/34 16K 4M 1G 94
/35 8K 2M 512M 93
/36 4K 1M 256M 92
/37 2K 512K 128M 91
/38 1K 256K 64M 90
/39 512 128K 32M 89
/40 256 64K 16M 88
/41 128 32K 8M 87
/42 64 16K 4M 86
/43 32 8K 2M 85
/44 16 4K 1M 84
/45 8 2K 512K 83
/46 4 1K 256K 82
/47 2 512 128K 81
/48 1 256 64K 80
/49 128 32K 79
/50 64 16K 78
/51 32 8K 77
/52 16 4K 76
/53 8 2K 75
/54 4 1K 74
/55 2 512 73
/56 1 256 72
/57 128 71
/58 64 70
/59 32 69
/60 16 68
/61 8 67
/62 4 66
/63 2 65
/64 1 64
K = 1,024
M = 1,048,576
G = 1,073,741,824
T = 1,099,511,627,776

The large address size of IPv6 permitted worldwide route summarization and guaranteed sufficient address pools at each site. The standard subnet size for IPv6 networks is a /64 block, which is required for the operation of stateless address autoconfiguration.[17] At first, the IETF recommended in RFC 3177 as a best practice that all end sites receive a /48 address allocation,[18] but criticism and reevaluation of actual needs and practices has led to more flexible allocation recommendations in RFC 6177[19] suggesting a significantly smaller allocation for some sites, such as a /56 block for residential networks.

This IPv6 subnetting reference lists the sizes for IPv6 subnetworks. Different types of network links may require different subnet sizes.[20] The subnet mask separates the bits of the network identifier prefix from the bits of the interface identifier. Selecting a smaller prefix size results in fewer number of networks covered, but with more addresses within each network.[21]

2001:0db8:0123:4567:89ab:cdef:1234:5678
|||| |||| |||| |||| |||| |||| |||| ||||
|||| |||| |||| |||| |||| |||| |||| |||128     Single end-points and loopback
|||| |||| |||| |||| |||| |||| |||| |||127   Point-to-point links (inter-router)
|||| |||| |||| |||| |||| |||| |||| ||124
|||| |||| |||| |||| |||| |||| |||| |120
|||| |||| |||| |||| |||| |||| |||| 116
|||| |||| |||| |||| |||| |||| |||112
|||| |||| |||| |||| |||| |||| ||108
|||| |||| |||| |||| |||| |||| |104
|||| |||| |||| |||| |||| |||| 100
|||| |||| |||| |||| |||| |||96
|||| |||| |||| |||| |||| ||92
|||| |||| |||| |||| |||| |88
|||| |||| |||| |||| |||| 84
|||| |||| |||| |||| |||80
|||| |||| |||| |||| ||76
|||| |||| |||| |||| |72
|||| |||| |||| |||| 68
|||| |||| |||| |||64   Single LAN; default prefix size for SLAAC
|||| |||| |||| ||60   Some (very limited) 6rd deployments (/60 = 16 /64 blocks)
|||| |||| |||| |56   Minimal end sites assignment;[19] e.g. home network (/56 = 256 /64 blocks)
|||| |||| |||| 52   /52 block = 4096 /64 blocks
|||| |||| |||48   Typical assignment for larger sites (/48 = 65536 /64 blocks)
|||| |||| ||44
|||| |||| |40
|||| |||| 36   possible future local Internet registry (LIR) extra-small allocations
|||| |||32   LIR minimum allocations
|||| ||28   LIR medium allocations
|||| |24   LIR large allocations
|||| 20   LIR extra large allocations
|||16
||12   Regional Internet registry (RIR) allocations from IANA[22]
|8
4

Numerical interpretation

edit

Topologically, the set of subnets described by CIDR represent a cover of the corresponding address space. The interval described by the notation   numerically corresponds to addresses of the form (for IPv4)  , where   has the lower   bits set to 0. (For IPv6, substitute 128.) For a fixed  , the set of all   subnets constitute a partition, that is a cover of non-overlapping sets. Increasing   yields finer and finer subpartitions. Thus two subnets   and   are either disjoint or one is a subnet of the other.

Prefix aggregation

edit

CIDR provides fine-grained routing prefix aggregation. For example, if the first 20 bits of their network prefixes match, sixteen contiguous /24 networks can be aggregated and advertised to a larger network as a single /20 routing table entry. This reduces the number of routes that have to be advertised.

See also

edit

References

edit
  1. ^ Y. Rekhter; T. Li (September 1993). An Architecture for IP Address Allocation with CIDR. doi:10.17487/RFC1518. RFC 1518.
  2. ^ a b V. Fuller; T. Li; J. Yu; K. Varadhan (September 1993). Classless Inter-Domain Routing (CIDR): an Address Assignment and Aggregation Strategy. doi:10.17487/RFC1519. RFC 1519.
  3. ^ J. Reynolds; J. Postel, eds. (April 1985). Assigned Numbers. doi:10.17487/RFC0943. RFC 943.
  4. ^ R. Hinden, ed. (September 1993). Applicability Statement for the Implementation of Classless Inter-Domain Routing (CIDR). doi:10.17487/RFC1517. RFC 1517.
  5. ^ a b V. Fuller; T. Li (August 2006). Classless Inter-domain Routing (CIDR): The Internet Address Assignment and Aggregation Plan. doi:10.17487/RFC4632. RFC 4632.
  6. ^ a b J. Mogul; J. Postel, eds. (August 1985). Internet Standard Subnetting Procedure. sec. 2.1. doi:10.17487/RFC0950. RFC 950.
  7. ^ Carl-Herbert Rokitansky, "Internet Cluster Addressing Scheme and its Application to Public Data Networks", Proc. 9th International Conference on Computer Communication (ICCC' 88), pp. 482–491, Tel Aviv, Israel, October/November 1988
  8. ^ Cluster Addressing and CIDR in the mail archives of the IETF
  9. ^ Brian Kantor (December 2018). "Re: Stupid Question maybe?". North American Network Operators Group. /24 is certainly cleaner than 255.255.255.0. I seem to remember it was Phil Karn who in the early 80's suggested that expressing subnet masks as the number of bits from the top end of the address word was efficient, since subnet masks were always a series of ones followd by zeros with no interspersing, which was incorporated (or independently invented) about a decade later as CIDR a.b.c.d/n notation in RFC1519.
  10. ^ William Simpson (December 2018). "Re: Stupid Question maybe?". North American Network Operators Group. Actually, Brian is correct. Phil was w-a-y ahead of the times. But I don't remember him talking about it until the late '80s.
  11. ^ T. Pummill; B. Manning (December 1995). Variable Length Subnet Table For IPv4. doi:10.17487/RFC1878. RFC 1878.
  12. ^ S. Williamson; M. Kosters; D. Blacka; J. Singh; K. Zeilstra (June 1997). Referral Whois (RWhois) Protocol V1.5. doi:10.17487/RFC2167. RFC 2167. IP networks are also lexically hierarchical labels using the Classless Inter-Domain Routing (CIDR) notation, but their hierarchy is not easily determined with simple text manipulation; for example, 198.41.0.0/22 is a part of 198.41.0.0/16, which is a part of 198.40.0.0/15.
  13. ^ Carpenter, B.; Jiang, S. (February 2014). Significance of IPv6 Interface Identifiers. doi:10.17487/RFC7136. RFC 7136.
  14. ^ Kohno, M.; Nitzan, B.; Bush, R.; Matsuzaki, Y.; Colitti, L.; Narten, T. (April 2011). Using 127-Bit IPv6 Prefixes on Inter-Router Links. doi:10.17487/RFC6164. RFC 6164.
  15. ^ J. Mogul, ed. (October 1984). Broadcasting Internet Datagrams in the Presence of Subnets. sec. 7. doi:10.17487/RFC0922. RFC 922.
  16. ^ F. Baker, ed. (June 1995). Requirements for IP Version 4 Routers. sec. 4.2.3.1. doi:10.17487/RFC1812. RFC 1812.
  17. ^ RFC 4862
  18. ^ IAB/IESG Recommendation on IPv6 Address Allocations to Sites. IAB/IESG. September 2001. doi:10.17487/RFC3177. RFC 3177.
  19. ^ a b T. Narten; G. Huston; L. Roberts (March 2011). IPv6 Address Assignment to End Sites. doi:10.17487/RFC6177. RFC 6177.
  20. ^ "ARIN IPv6 Addressing Plans". Getipv6.info. 2016-03-25. Retrieved 2018-03-12.
  21. ^ "RIPE IP Allocation Rates". Archived from the original on 2011-02-03.
  22. ^ "IANA IPv6 unicast address assignments". Iana.org. Retrieved 2018-03-12.

Further reading

edit
edit