IPv6
Is the time right for IPv6?
The Internet as we know it is slowly running out of addresses. When the modern version of the Internet Protocol (IPv4) was first published in 1981, the world assumed that the address space -- the number of possible addresses that you could create with its 32 bits -- would serve Internet users for a long time to come.
Unfortunately, the huge success of the Internet, led predominantly by the introduction of HTTP and the Web, has rapidly exhausted
the number of available IP addresses held by the Internet Assigned Numbers Authority (IANA), the organization that allocates
them in blocks to countries.
Current estimates suggest that IANA’s pool of remaining addresses will be empty by
2012 or sooner. Foreseeing the problem, the Internet Engineering Task Force (IETF), began work on the next major version of
the protocol, called IPv6, which was ratified as a draft standard in 1998.
The biggest difference between IPv6 and IPv4 is the increased address space. IPv6 uses 128 bits for its addresses, which creates
an address range that would be almost impossible to exhaust.
The number of combinations allows for 655, 570, 793, 348,
866, 943, 898, 599 addresses for every square meter on the earth’s surface.
One way in which this could be exploited is to create an ‘Internet of things’. For several years now, ubiquitous computing experts have advocated the idea of connecting everyday devices to the Internet. If your refrigerator could talk to the dairy counter at the grocery store, it might tell it to remind you pick up milk and yoghurt, for example. Vehicles with IP addresses could talk to everything from traffic services to IP-enabled garage doors. IP-enabled cargo crates might communicate their position and internal temperature to a central supply chain management system.
However, even with the onset of ubiquitous computing, there will be vast surplus of addresses that can be used to make Internet routing easier. Unlike IPv4, IPv6 addresses are designed to support the hierarchical nature of the Internet. There are consequently different types of IPv6 address, to be used depending on whether an organization wants an address to be globally accessible across the entire Internet, or just locally accessible from a single site or link.
This ability to define the scope of an IPv6 address using various prefixes, along with the large number of available addresses, removes the need for Network Address Translation (NAT). NAT is used in the IPv4 world as a means of hiding a range of private Internet addresses from the rest of the Internet behind a single address. The problem with NAT is that it made it difficult to connect specific devices inside an organization with external services. VoIP, unified communications and video services, for example, could be troublesome to configure in a NAT environment.
In an IPv6 world, any local address can be made to talk to another address on the public Internet directly, if properly configured.
That will make the ubiquitous computing concept more achievable, and will also make it easier for network administrators to
support the delivery of richer, application-layer services seamlessly between different organizational address spaces.
However,
NAT isn’t simply a way of limiting the consumption of public IP addresses. It is also seen as a security measure by many organizations,
preventing external addresses from reaching internal clients. But IPv6-based addressing can be configured to restrict external
devices from reaching site or link-specific addresses. And unlike IPv4, IPv6 must support traffic encryption; it is not simply
an optional add-on.
The larger an organization is, the more it will appreciate another feature of IPv6: address autoconfiguration. The system is designed to be used in situations where the Dynamic Host Configuration Protocol (DHCP) is not available. DHCP is used to automatically assign IP addresses, and IPv4 required these to be assigned manually if a DHCP-enabled device is not present. However, IPv6 devices can still auto-assign their own addresses in the absence of a DHCP service, making it much easier to install devices on the network.
The other advantage to IPv6 is built-in quality of service provisioning. IPv4 requires additional layer 3 quality of service
mechanisms to be used for quality of service during routing. DiffServ (ratified in 1998) serves as a means of classifying
traffic into broad classes, prioritizing some over others should a network connection have to begin dropping packets due to
congestion.
Multi-protocol label switching (MPLS) is used in core backbone architectures for mapping layer 3 quality
of service information onto layer 2 network transports. Both of these systems require significant investments either by the
network provider or the sending and receiving parties. DiffServ in particular requires traffic priority information to be
included in the network header by the sending party.
IPv6 cuts through the whole mess by including support for traffic flow control directly in the header as part of the
standard. This means that services supporting IPv6 can automatically take advantage of traffic prioritization without layering
extra complexity into the network.
Therein lies the main challenge to IPv6. Few of those IPv6-based services exist. The transition to this standard has been
relatively slow. One of the earliest widespread adopters was the US military, which has specified IPV6 as a standard within
its own operations as a basis for supporting network-centric warfare.
Non-governmental organizations have been slow to
adopt it, however. IPv6 headers are incompatible with IPv4. While the stack is now supported in modern client and server operating
systems, and while many network equipment vendors are including support for both protocols in their equipment, most customers
have traditionally stuck with the earlier standard.
Much of the inertia is down to a lack of services. For IPv6 to be useful, there must be a critical mass of services available online, and these have failed to emerge. Tech-focused companies such as Google have created IPv6 interfaces to their services, but they are the exception, rather than the rule. Many network providers are now rolling out IPv6 support in the core in preparation for a mass migration, but it will take time for customers to catch up -- especially in a financial climate where new capital expenditure and hardware refresh will be a low priority.
Even after the addresses begin to run out at the top level, countries will still have blocks of IP addresses to allocate.
Eventually, however, there will be no choice but to implement this exciting and beneficial new standard, as the world begins
its slow but steady transition.
print this page
send to a colleagueall you need to know about enterprise network technology
Over the last 30 years computer networks have established themselves as a cornerstone of modern business. With the commercialization of the Internet in the 1990s, networks have become fundamental…
Move beyond perimeter security to protect your organization
What does network security mean in 2009? Thanks to rapid developments in both technology and business culture, it means something vastly different than it did even at the start of the decade.
The emerging applications that will shape tomorrow’s enterprise network
n the modern enterprise, the network is central to countless critical business processes: think of how a single centralized SAP platform may be controlling manufacturing processes and schedules on six …
