The Basics of Internet Routing and Addressing ccna course institute in delhi ncr

Network Bulls
www.networkbulls.com
Best Institute for CCNA CCNP CCSP CCIP CCIE Training in India
M-44, Old Dlf, Sector-14 Gurgaon, Haryana, India
Call: +91-9654672192

The original design for the Internet called for the assignment of globally unique IPv4 addresses
for all hosts connected to the Internet. The idea is much like the global telephone
network, with a unique phone number, worldwide, for all phone lines, cell phones, and
the like.
To achieve this goal, the design called for all organizations to register and be assigned one
or more public IP networks (Class A, B, or C). Then, inside that organization, each address
would be assigned to a single host. By using only the addresses in their assigned network
number, each company’s IP addresses would not overlap with other companies. As a result,
all hosts in the world would have globally unique IP addresses.
The assignment of a single classful network to each organization actually helped keep Internet
routers’ routing tables small. The Internet routers could ignore all subnets used inside
each company, and instead just have a route for each classful network. For instance, if
a company registered and was assigned Class B network 128.107.0.0/16, and had 500 subnets,
the Internet routers just needed one route for that whole Class B network.
Over time, the Internet grew tremendously. It became clear by the early 1990s that something
had to be done, or the growth of the Internet would grind to a halt. At the then-current
rate of assigning new networks, all public IP networks would soon be assigned, and
growth would be stifled. Additionally, even with routers ignoring the specific subnets, the
routing tables in Internet routers were becoming too large for the router technology of
that day. (For perspective, more than 2 million public Class C networks exist, and two million
IP routes in a single IP routing table would be considered quite large–maybe even too
large–for core routers in the Internet even today.)
To deal with these issues, the Internet community worked together to come up with both
some short-term and long-term solutions to two problems: the shortage of public addresses
and the size of the routing tables. The short-term solutions to these problems included
■ Reduce the number of wasted public IP addresses by using classless IP addressing
when assigning prefixes—assigning prefixes/lengths instead of being restricted to assigning
only Class A, B, and C network numbers.
■ Reduce the need for public IP addresses by using Port Address Translation (PAT, also
called NAT overload) to multiplex more than 65,000 concurrent flows using a single
public IPv4 address.
■ Reduce the size of IP routing tables by making good choices for how address blocks
are allocated to ISPs and end users, allowing for route summarization on a global scale.
This section examines some of the details related to these three points, but this information
is not an end to itself for the purposes of this book. The true goal is to understand
outbound routing (from the Enterprise to the Internet), and the reasons why you may or
may not need to use a dynamic routing protocol such as Border Gateway Protocol (BGP)
between the Enterprise and the Internet.
The process starts with ICANN and IANA. These organizations maintain a set of currently
unallocated public IPv4 addresses. (See http://www.iana.org/numbers/, and look
for the IPv4 addresses link, to see the current list.) When ARIN, the RIR for North America,
notices that it is running out of IPv4 address space, ARIN requests a new public address
block. IANA examines the request, finds a currently unallocated public address
block (Step 1 in the figure), and allocates the block to ARIN (Step 2 in the figure).
Next, an ISP named NA-ISP1 (shorthand for North American ISP number 1) asks ARIN for
a prefix assignment for a /16 sized address block. After ARIN ensures that NA-ISP1 meets
some requirements, ARIN assigns a prefix of 198.133.0.0/16 (Step 3 in the figure). Then,
when Company1 becomes a customer of NA-ISP1, NA-ISP1 can assign a prefix to Company
1 (198.133.219.16/29 in this example, Step 4).
Although the figure shows the process, the big savings for public addresses occurs because
the user of the IP addresses can be assigned a group much smaller than a single
Class C network. In some cases, companies only need one public IP address; in other
cases, they may need only a few, as with Company1 in Figure 12-1. This practice allows
IRs to assign the right-sized address block to each customer, reducing waste.
Internet Route Aggregation
Although the capability to assign small blocks of addresses helped extend the IPv4 public
address space, this practice also introduced many more public subnets into the Internet,
driving up the number of routes in Internet routing tables. At the same time, the number
of hosts connected to the Internet, back in the 1990s, was increasing at a double-digit
rate–per month. Internet core routers could not have kept up with the rate of increase in
the size of the IP routing tables.
The solution was, and still is today, to allocate numerically consecutive addresses–addresses
that can be combined into a single route prefix/length—by geography and by ISP.
These allocations significantly aid route summarization.
For example, continuing the same example shown in Figure 12-1, Figure 12-2 shows some
of the routes that can be used in ISPs around the globe based on the address assignment
shown in Figure 12-1.
First, focus on the routers shown in Europe and South America. Routers outside North
America can use a route for prefix 198.0.0.0/8, knowing that IANA assigned this prefix to
be used only by ARIN, which manages IP addresses in North America. The underlying
logic is that if the routers outside North America can forward the packet into North
America, then the North American routers will have more specific routes. The single route
for 198.0.0.0/24 shown in Europe and South America can be used instead of literally millions
of subnets deployed to companies in North America such as Company1.
Next, consider routers in North America, specifically those outside the NA-ISP1 network.
Figure 12-2 shows one such ISP, named NA-ISP2 (North American ISP number 2), on the
left. This router can learn one route for 198.133.0.0/16, the portion of the 198.0.0.0/8 block
assigned to NA-ISP1 by IANA. Routers in NA-ISP2 can forward all packets for destinations
inside this prefix to NA-ISP1, rather than needing a route for all small address blocks
Note: For the purposes of this book, the terms NAT, PAT, and NAT overload are used
synonymously. There is no need to distinguish between static NAT, dynamic NAT without
overload, and dynamic NAT with overload (also called PAT).
NAT reduces the need for public IPv4 addresses to only a few addresses because of how
NAT can multiplex flows using different TCP or UDP port numbers. Figure 12-3 shows a
sample that focuses on a router performing NAT. The figure shows an Enterprise network
on the left, with the Enterprise using private Class A network 10.0.0.0/8. The Internet sits
on the right, with the NAT router using public IP address 200.1.1.2.
The figure shows how the Enterprise, on the left, can support three flows with a single
public IP address (200.1.1.2). The NAT feature dynamically builds its translation table,
which tells the router what address/port number pairs to translate. The router reacts when
a new flow occurs between two hosts, noting the source IP address and port number of
the Enterprise host on the left, and translating those values to use the public IP address
(200.1.1.2) and an unused port number in the Internet. Note that if you collected the traffic
using a network analyzer on the right side of the NAT router, the IP addresses would
include 200.1.1.2 but not any of the network 10.0.0.0/8 addresses. Because the combination
of the IP address (200.1.1.2 in this case) and port number must be unique, this one IP
address can support 216 different concurrent flows.
Private IPv4 Addresses and Other Special Addresses
When allocating the public IPv4 address space, IANA/ICANN restricts themselves in several
ways. Of course, the private IP address ranges cannot be assigned to any group for
www.CareerCert.info
Chapter 12: Internet Connectivity and BGP 395
Table 12-2 Private IP Address Reference
Number of Classful
Networks
Range of Classful Networks Prefix for Entire
Range
(1) Class A: 10.0.0.0 10.0.0.0/8
(16) Class B: 172.16.0.0 through 172.31.0.0 172.16.0.0/12
(256) Class C: 192.168.0.0 through
192.168.255.0
192.168.0.0/16
Table 12-3 Reserved Values in IPv4 Address Range (RFC 3330)
Value or
Range
Reason
0.0.0.0/8 Used for self-identification on a local subnet.
127.0.0.0/
8
Loopback testing.
169.254.
0.0/16
This “link local” block is used for default IPv4 address assignment when
DHCP process fails.
192.0.2.0
/24
Reserved for use in documentation and example code.
192.88.9
9.0/24
Used for IPv6 to IPv4 relay (6to4 relay) (RFC 3068).
198.18.0.
0/15
Benchmark testing for Internet devices (RFC 2544).
use in the public Internet. Additionally, several other number ranges inside the IPv4 address
space, as summarized in RFC 3330, are reserved for various reasons. Tables 12-2 and
12-3 list the private addresses and other reserved values, respectively, for your reference.
Table 12-3 lists other reserved ranges of IPv4 addresses that IANA will not allocate in the
public Internet.
In summary, every Enterprise that connects to the Internet must use at least one public
IP address, and often several public IP addresses. Although some companies do have a
large public IPv4 address block–often obtained before the shortage of public IPv4 addresses
in the early to mid-1990’s–most companies have a small address block, which
then requires the use of NAT/PAT. These details have some impact on whether BGP is
useful in a given case.