IPv4 Address Exhaustion

Transcription

1 UNIVERSITY OF HERTFORDSHIRE IPv4 Address Exhaustion Essay Mohammad Alqahtani 5/11/2009

2 The Internet Protocol (IP) is designed for use in interconnected systems of packet-switched computer communication networks. In other words, The Internet Protocol (IP) specifies how communications take place between one device and another through an addressing system (Marina del Rey, 1981).The first version published of the IP was Internet Protocol Version 4 (IPv4) which provides more than 4 billion different IP addresses identifying any connected system to the internet. This was considered be enough in the early stages of the Internet, but later when the internet being spread rapidly over worldwide the consumption of address space has been increased significantly. Consequently, address space are allocated and network is classified to manage Internet addresses through open and transparent policy frameworks. In the early 1990s, even after the network classification, this would not be enough to prevent IPv4 address exhaustion. As a result of this, in 1992, "IP Next Generation" (IPng) was published and by 1996, series of RFCs were released defining Internet Protocol Version 6 (IPv6). Despite that it was widely expected that IPv4 will be supported beside IPv6 for future but IPv4 nodes are still not able to communicate directly with IPv6 nodes. This is critical for the future of the the internet growth because all new users connecting to the Internet, and all companies that require IP addresses for their expansion, will be affected by the transferring from the current status of ready availability of unallocated IPv4 addresses(dionisiatabureguci). This essay will discuss the IPv4 Address exhaustion in term of the differences between IPv4 and IPv6, evaluation of the situation of IPv4 address space remaining unallocated, allocation of IPv4 addresses policy on the internet, transition methods from IPv4 to IPv6, current state of IPv6 deployment and possible problems might happen when IPv4 addresses are exhausted. 2

3 The IPv4 can be compared to IPv6 in term of IP address format, packet structure, and address types. The IPv4 address format is 32 bits long with a notation nnn.nnn.nnn.nnn (n<=255) such as and is classified as: A, B, C, D, or E depending on prefix bits and mask address which specifies range of IP addresses that belong to one network. The address space of IPv4 is and. Furthermore, there are three address types in IPv4: unicast, multicast, and broadcast each of them has different purpose (IBM). Generally address lifetime concept is not applicable in IPv4 apart from addresses assigned using DHCP. Additionally, the maximum Packet size is octets with a header that has variable size between 20 and 40 bytes. The packet header consists of different fields which are Ver4, IHL, Type of service, Total Length, Identification, Flags, Fragmentation Offset, Time To Live, Protocol, Header Checksum, Source Address, Destination Address and Options as shown in figure 1. This packet can be fragmented many times during routing by routers (IBM). Figure1 IPv4, unlike IPv6, supports Dynamic Host Configuration Protocol (DHCP), File Transfer Protocol (FTP), host table, Layer 2 Tunnel Protocol (L2TP), Network Address Translation (NAT), packet filtering, packet forwarding, Quality of service (QoS) and Point-to-Point Protocol (PPP). Despite the above, The IPv6 address format is 128 bits long with a notation xxxx:xxxx:xxxx:xxxx:xxxx:xxxx:xxxx:xxxx (x is a hexadecimal digit, representing 4 bits,each block is separated by semicolon) such as (3ffe:0501:0008:0000:0260:97ff:fe40:efab) and the address space of IPv6 is which is 1028 times larger than the address space of IPv4. IPv6, unlike IPv4, does not use address mask also IPv6 addresses have two 3

4 lifetimes which are preferred is used for source IP address and valid is used for destination IP address but some of IPv6 addresses do not have lifetimes. Moreover, the address types in IPv6 are unicast, multicast, and anycast identifying multiple interfaces. The maximum IPv6 Packet size is octets with a fixed header size of 40 bytes. This packet header is simpler than IPv4 packet header which consists of Ver6, Prio, Flow Label, Payload Length, Next Header, Hop Limit, Source Address and Destination Address as shown in figure2. Furthermore, packet can be fragmented only once at the source node and reassembled at the destination node (Microsoft TechNet, 2004). figure2 On other hand, both IPv4 and IPv6 use the same concepts of ports, trace route, transport layers, services table, Internet Control Message Protocol, Domain Name System, interfaces, protocol table, network table, LAN connection and loopback address (IBM). The allocation of IPv4 is very important to the integrity of the public IP and, so the IPv4 address space is allocated and managed by The Internet Assigned Numbers Authority (IANA) and The Regional Internet Registries. The RIR can be defined as a regional organization which are responsible to allocate and register the Internet addresses to the internet service provider and end user organizations for a certain region in world ( Wikipedia, Regional Internet registry ). The allocation of the Internet address space is mainly managed by the IANA under a contract with ICANN (Internet Corporation for Assigned Names and Numbers). IANA allocates units of /8 IPv4 4

5 address blocks to RIRs which are authorized to re-allocate these addresses within their region. Then, Internet Service Providers and other organisations obtain IP address space from the RIR that covers their area. There are five Regional Internet Registries which are AFRINIC, APNIC, ARIN, LACNIC and the RIPENCC as shown in fingure3 (IPv4 Address Report). However, some IPv4 address blocks are reserved for different purposes as described in table1 (IANA IPv4 Address Space Registry). IPv4 address block(s) Purposes of use Reference /8 self-identification RFC /8 Loopback RFC /16 Link Local RFC /12 Private Networks RFC /24 Test-Net RFC /24 6to4 Relay Anycast RFC /16 Private Networks RFC /15 Network Interconnect Device RFC3330 Benchmark Testing Table 1 Figure 3 (Wikipedia, Regional Internet registry) In the last decade, the rapidly increase in development of the Internet industry has led to the depletion of the unallocated IPv4 address space to the Regional Internet Registries (RIRs). Regarding IPv4 Address Report website, the free addresses are only 12 percent of IPv4 address space and the 5

6 predictive exhaustion date of Unallocated IPv4 Addresses in The (IANA) will be on 12-Jun-2011 while the predictive exhaustion date of (RIRs) Unallocated Address Pool will be on 04-Oct-2012 (IPv4 Address Report). This does not mean we will not be able to use the existing IPv4 Internet after the address pool runs out, but there are no more unallocated IPv4 addresses available from the Regional Internet Registries. In other words, the servers that have only IPv4 address will continue to exist even after the IPv4 address exhaustion and the Internet Service Providers (ISPs) cannot assign additional global IPv4 addresses to their end-users in order to access such servers. Additionally, beside the problem of the IPv4 address space depletion, rapid increase in number of routing tables which reaches the maximum number of routing table at 60,000. However, there are four methods that might delay the IPv4 address space consumption. Firstly, using Network Address Transition (NAT) or share connection which can delay the IPv4 address exhaustion by allowing a large number of hosts to be globally represented by a single public IP address. These hosts are connected to the internet throw NAT device (gateway) that use one public IP address (Martin Dunmore, 2005). However, this methods has some disadvantages such as complicity to configure and maintain NAT, some application would be hard to configure such as FTP and peer to peer application, problems with security protocols such as IPSec protocol which notices the changes in packet s headers made by NAT and then drops these packets. Another disadvantage is reducing the performance of the network because of the transition time of packets from private to public IP address(the TCP/IP Guide, 20,9,2005). Second method is Classless Inter-Domain Routing (CIDR) that permits for a more proficient use of IP addresses on both network devices and LANs by using a network subnet anywhere from 13 to 27 bits rather than using normal IPv4 address classes A, B or C (8, 16 or 24) only. This solves the problem of limited number of routing table (Tom McLaughlin, 26-Oct-2005). Another method is DHCP address sharing by using limited hire time to the connected hosts. This limited connected times of hosts will allow the addresses to be used many times rather than being reserved by computers not in use. Final method is RFC 1918 Address Space (VPN ) that can apply for network nodes that only need to communicate in a district area or campus environment by using private IP addresses instead of public routed IP addresses (UEN IP Address Allocation Policy, 2008). However, This techniques still cannot handle the pressure on the consumption of IPv4 space Address because the heavily populated countries in Asia such as India and China which are rapidly rising in 6

7 economy and industry, and also Internet-connected personal devices technology such as connecting via mobile phones. This view has been supported in the work Vilho Räisänen of (86, 2003). The good solution of the lack of IPv4 addresses is to transfer to IPv6 which provides huge number of IP addresses. There is a suggested transition mechanism for IPv6 hosts and routers defined in RFC 2893 and RFC2185 documents. This is called Simple Internet Transition which includes dual-stack IP implementations for both corporative hosts and routers, embedding IPv4 addresses in IPv6 addresses, IPv6-over-IPv4 tunneling mechanisms and IPv4/IPv6 header translation(wikipedia, IPv6). Dual-stack technique is a protocol stack that has both IPv4 and IPv6. Therefore, the same transport protocols, TCP, UDP, and others, can run over both IPv4 and IPv6. Also, the same applications can run over both IPv4 and IPv6 as demonstrated in figure 4( IPv6 Administration Guide, 2003). Figure 4 This technique can be applied until all IPv4 addresses upgraded to IPv6. However, this mechanism needs additional requirements which are memory and CUP power in the host to handle the two separated stacks, competent DNS resolver to resolve both IP address versions, two tables for each protocol and the firewall must be configured in order to protect both IPv4 and IPv6 networks. This view has been supported in the work of Silvia Hagen (254, 2002). The second method is (6to4 address) or embedding IPv4 addresses in IPv6 addresses to convert IPv4 address to IPv6 address by adding the 32 bit of IPv4 address after 96 bits of zeros so the IPv6 devices can deal with new form of IPv4 address (Figure 5). This technique makes the new addresses unique 7

8 for every IPv4 address and also provides a automatic technique to find an IPv4-to-IPv6 gateway for those IPv6 addresses (The TCP/IP Guide -IPv6/IPv4 Address Embedding, 2005). Another mechanism is tunneling mechanisms which capsulate IPv6 packets inside IPv4 packets then these packets are routed through the IPv4 routers until their destinations as can be seen in figure 5 (The TCP/IP Guide -IPv6/IPv4 Address Embedding, 2005). Figure 5 There are many type of tunnelling methods such as ISATAP and 6over4 which are available forthe switching of local non router sites or for use over campus networks. The ISATAP tunneling technique is quite similar to 6to4 tunneling, with the IPv4 address embedded in the last 32 bits rather than the first 48 bits of the IPv6 address. 8

9 The main benefit of Tunneling mechanism is to increase deployment of IPv6 gradually apart from the ISP support, since IPv6 can communicate to other nodes over an IPv4 network. Since 1998, the deployment of IPv6 has been started by different countries, organisations, ISPs, internet services vendors and hardware and software vendors. For example, the USA has issued permission to all internet vendors to transfer to an IPv6 platform by the end of 2008 and the General Services Administration in 2007 has spent over $150 billion on contracts for the infrastructure of this project. The benefit of this is to increase the migration to the next generation of internet. In Canada, IPv6 node can communicate with other nodes successfully in the USA and different countries via local IPv6 and through IPv4 tunnels (Kaushik Das,2008). In Japan, the first deployment of IPv6 address using tunnelling mechanism to more than 100 customers was in Furthermore, most of Japanese ISPs have started IPv6 services to their customers and Japan Gigabit Network Company uses IPv6 in ATMs and Native IPv6 transport. Additionally, hardware and software vendors have started functioning the requirements of IPv6 networks and Japan Telecom and KDDI have started trials in some areas such as mobile phones and train internet (Kaushik Das,2008). Moreover, the China has published the China's Next Generation Internet project (CNGI) for five year long. The aim of this project is to have a significant proportion of the IP address space by switching from IPv4 network to IPv6 early. Additionally, there is project to link 25 universities in different cities across China by private IPv6 network (Ingrid Marson, 2004). In Korea, KOREAv6 project has been started in 2004 with objective of creating backbone of IPv6 internet in order to be ready for use by This project has three phases. Phase I started in 2004 to build the nation IPv6 network which provides IPv6 Internet gateway service and tests /39 IPv6 gears such as switches,routers and VPN. In 2005, Phase II applied IPv6 technologies to the eight services of IT839 such as VoIP, WiBro, and Home network that expands IPv6 network to the public network. Finally, Phase III has been started in 2006 to allow end users connecting to the internet via IPv6. In France, the tunnelling and dual stack mechanisms are successfully used to connect IPv6 nodes to the public Internet (Kaushik Das,2008). 9

10 In Europe, there was 6NET project which constructed and run a European IPv6 network from 2002 to 2005 with the objective of connecting sixteen countries to achieve experience of IPv6 deployment over IPv4-based networks. This project demonstrated successfully IPv6 networks using the dual stack mechanism (Martin Dunmore, 2005). Over most of Hardware vendors such as IBM, CISCO, HP, HITACHI, they have started since 1998 to enhance the internet backbone equipments (such as switches, routers and so on ) to be compatible with the IP6 networks, achieve requirements of transition methods and support the protocol translation mechanisms that permit communication between IPv6-only and IPv4-only hosts such as NAT-PT and BIS ( Cisco). Application and operation systems vendors have started in the beginning of For example, Microsoft has started providing the IPv6 services in windows XP, and vista and UNIX as well. Currently, IPv6 address can be configured in most of the OS. However, the current state of allocation IPv6 address is less than one percent of the IPv6 address space. Regarding the Internet Number Resource Status Report issued on 31 march 2009 by Regional Internet Registries, the total number of IPv6 /32 allocated to the ISPs and LIR is less than 3000 while the total number of of IPv6 /32 allocated by IANA to RIRs is about 73,000 as shown in figure 6 (NRO, 2009). Figure 6. 10

11 The most important factor to migrate to next generation IPv6 networks is the Internet service providers and Telecommunication companies over the world because the ISPs deal with the end users directly and the Telecommunication companies enhance the backbone of IPv6 networks. Additionally, the RIRs should issue a new allocation policy for the IPv4 and IPv6 such as the private network should use the IPv6 network. There might be anticipated problems happened when the IPv4 addresses run out. Firstly, Internet Service Providers cannot assign additional public IPv4 addresses to their end-users in order to access such servers. Secondly, End-users without any global IPv4 address space will not be able to access the IPv4 Internet after the exhaustion of IPv4 addresses. Another problem could be happen that end users with only IPv6 cannot access to IPv4 Servers. Additionally, most companies that provide their services throw internet will not be able to serve some of their clients such as e-banking, e-business, e-learning. Finally, the internet will be limited and many services will be eliminated. Currently, all of companies, organizations and ISPs should be ready to switch to IPv6 networks and upgrade their internet equipments and applications to be compatible with both IPv4 and IPv6 network. Companies such as Amazon should try to have a method that enables any users either using IPv4 or IPv6 address to access their servers. So the network administrator should upgrade the server in order to communicate with both versions of IP addresses. This server should have firewall that can deal with both versions of IP addresses, one of the transition mechanisms of IPv6. Additionally, big campus network can have one IPv6 network communicating with other IPv4 networks to gain experience and to be ready for dealing with two different networks. In conclusion, the situation of unallocated IPv4 address space is critical issue and still there is no direct communication between IPv4 address and IPv6 address. However, there are successful mechanisms to enable IPv6 network to communicate with IPv4 network. Furthermore, there are huge projects over the world to deploy IPv6 addresses and to establish infrastructure of the new internet generation. It is recommended that the deployment of IPv6 address starts with private internet network (LAN) in order to increase the awareness of IPv6 protocol and gain a good experience in this field. 11

12 References Cisco. IPv6 Deployment Strategies. [Online] Available at: [Accessed 3 May 2009]. Dionisia Tabureguci, Telecommunications: WHAT S THE FUTURE OF THE INTERNET ECONOMY?. [Online] Available at: ID=18084/overideSkinName=issueArticle-full.tpl [accessed 25 April 2009]. IANA IPv4 Address Space Registry. [Online] (Updated 28 April 2009). Available at: [Accessed 24 April 2009]. IBM, Compare IPv4 to IPv6. [Online] Available at: [Accessed 26 April 2009]. Ingrid Marson, China launches largest IPv6 network. [Online] Available at: [Accessed 1 May 2009]. IPv4 Address Report. [Online] (Updated 6 May 2009). Available at: [Accessed 26 April 2009]. IPv6 Administration Guide, Making the Transition From IPv4 to IPv6. [Online] Available at: [Accessed 1May 2009]. Kaushik Das,2008. IPv6 Deployment Around The World. [Online] Available at: [Accessed 6 May 2009]. Marina del Rey, 1981, Internet Protocol. [Online] Available at: [accessed 25 April 2009]. Martin Dunmore, 9,2005 An IPv6 Deployment Guide. [Online]. Available at: [Accessed 23 April 2009]. Microsoft TechNet, 2004, TCP/IP Fundamentals for Microsoft Windows. [Online] (Updated 18 April 2006). Available at: [Accessed 26 April 2009]. NRO, Internet Number Resource Status Report. [Online] Available at: [Accessed 7 May 2009]. Silvia Hagen, IPv6 essentials. O'Reilly Media, Inc. 12

13 The TCP/IP Guide, 9,2005. IP IPv6/IPv4 Address Embedding. [Online] Available at: [Accessed 21 April 2009]. The TCP/IP Guide, 9,2005. IP NAT Overview, Motivation, Advantages and Disadvantages. [Online] Available at: [Accessed 21 April 2009]. Tom McLaughlin, 26-Oct-2005, Classless Inter-Domain Routing (CIDR) Overview. [Online] Available at: [Accessed 1May 2009]. UEN IP Address Allocation Policy, 2008, [Online] Available at: [Accessed 29 April 2009]. Vilho Räisänen,, Implementing service quality in IP networks. Harwich: John Wiley and Sons Ltd. 13