- Information
- AI Chat
Was this document helpful?
Global Unicast Addresses
Course: Systems Programming (01:198:214)
23 Documents
Students shared 23 documents in this course
University: Rutgers University
Was this document helpful?
Global Unicast Addresses
Unicast addressing is by far the most crucial type of addressing that IPv6 must offer. This must
be done in a way that supports the Internet's quick rate of adding new hosts and permits scaling
of routing as the number of physical networks in the Internet increases. The unicast address
allocation scheme, which governs how unicast addresses will be assigned to service providers,
autonomous systems, networks, hosts, and routers, is thus at the core of IPv6. In actuality, the
planned IPv6 unicast address allocation scheme is strikingly similar to the IPv4 CIDR
deployment. It is useful to establish some new terminology in order to comprehend how it
functions and how it offers scalability. We can consider a transit AS as a provider and a
nontransit AS (i.e., a stub or multihomed AS) as a subscriber. We can further categorize suppliers
into direct and indirect ones.
The former are connected to subscribers directly. The latter are sometimes referred to as
backbone networks because they mainly connect other providers but do not connect directly to
subscribers. With these definitions, we can see that the Internet has some inherent order and is
more than just a loosely connected collection of independent systems. Making use of this
hierarchy without creating processes that break down when the hierarchy is not rigidly followed,
as happened with EGP, is tricky. For instance, when a subscriber joins to a backbone or when a
direct provider starts connecting to numerous additional providers, the line between direct and
indirect providers is blurred. Similar to CIDR, the IPv6 address allocation strategy aims to
aggregate routing data to lighten the load on intradomain routers. Again, the key concept is to
aggregate reachability information for a large number of networks and even huge numbers of
autonomous systems by using an address prefix, which is a group of contiguous bits at the most
significant end of the address. The most common approach to do this is to give an address prefix
to a direct provider, who will then give its subscribers lengthier prefixes that start with that
prefix. Exact evidence of this can be seen in Figure 3.22.
A provider can therefore promote a single prefix for all of its subscribers. Unavoidably, the
disadvantage is that if a site chooses to switch providers, it will have to get a new address prefix
and renumber all of its nodes. This might be a massive operation that would deter most
individuals from ever switching providers. Due to this, research is still being done on alternative
addressing schemes, like geographic addressing, in which a site's address is determined by its
location rather than by the provider to whom it is attached. However, at the moment, provider-
based addressing is essential for effective routing. Although the process of assigning IPv6
addresses is substantially the same as that of assigning IPv4 addresses since the introduction of
CIDR, IPv6 has the important advantage of not needing to accommodate a sizable installed base
of allocated addresses into its plans. One concern is whether hierarchical aggregation makes
sense at different levels of the structure.
Should all providers, for instance, get their address prefixes from a prefix assigned to the
backbone they connect to? Considering that the majority of providers connect to several
backbones, this is probably unnecessary. Additionally, there are far fewer advantages to
aggregating at this level because the number of suppliers is considerably lower than the number
of sites. At the national or continental level, aggregation may be justified. The topology of the
