INTERNET DRAFT S. Bandyopadhyay draft-shyam-mshn-ipv6-21.txt April 26, 2015 Intended status: Informational Expires: October 26, 2015 Mesh Structured Hierarchical Networking and IPv6 draft-shyam-mshn-ipv6-21.txt Abstract This document tries to address an approach for the reorganization of the entire network in a large address space. It describes how a three-tier mesh structured hierarchy can be established based on fragmenting the entire space into some regions and some sub regions inside each of them. It addresses issues which could be relevant to this architecture in the context of IPv6. This document also tries to come out with an approach how IP switch based network can perform as good as ATM network for the processing of real time traffic. Status of this Memo This Internet-Draft is submitted in full conformance with the provisions of BCP 78 and BCP 79. Internet-Drafts are working documents of the Internet Engineering Task Force (IETF). Note that other groups may also distribute working documents as Internet-Drafts. The list of current Internet- Drafts is at http://datatracker.ietf.org/drafts/current/. Internet-Drafts are draft documents valid for a maximum of six months and may be updated, replaced, or obsoleted by other documents at any time. It is inappropriate to use Internet-Drafts as reference material or to cite them other than as "work in progress." This Internet-Draft will expire on October 26, 2015. Copyright Notice Copyright (c) 2015 IETF Trust and the persons identified as the document authors. All rights reserved. This document is subject to BCP 78 and the IETF Trust's Legal Provisions Relating to IETF Documents (http://trustee.ietf.org/license-info) in effect on the date of publication of this document. Please review these documents carefully, as they describe your rights and restrictions with respect to this document. Bandyopadhyay Expires October 26, 2015 [Page 1] Internet Draft MSHN and IPv6 April 26, 2015 Table of Contents 1. Introduction.....................................................2 2. A Three tier mesh structured hierarchical network................3 2.1. Route propagation...........................................5 2.2. Determination of prefix lengths.............................8 2.2.1. A pseudo optimal distribution of prefixes in a 64bit architecture.........................................9 2.2.2. Whether to go for a two tier or three tier hierarchy ....................................................10 2.3. Issues related to Satellite communications.................11 2.4. Issues related to PI addressing and IP mobility............12 2.4.1 IP address aliasing..................................14 2.4.2. Changes expected with the specifications related to IP mobility......................................15 2.5. Refinements over existing IPv6 specification...............16 2.6. Distributed processing and Multicasting....................18 3. Solution for site multihoming...................................19 3.1. Multihoming and IP Mobility................................22 3.1.1. IP Address Stacking.................................23 3.2. Implementation aspects.....................................26 3.2.1. Processing of system call getsrcaddr................27 3.2.2. Changes required in ip_output and ip_forwarding modules.............................................28 3.2.3. Processing of protocol input routines and socket IO system calls........................................29 3.3. Multihoming, VPN and load sharing..........................29 4. Processing of real time packets (QoS issue).....................31 4.1. Dual mode operation........................................33 4.2. Expected changes at the application layer..................34 5. IANA Consideration..............................................34 6. Security Consideration..........................................34 7. Acknowledgments.................................................35 8. Normative References............................................35 9. Informative References..........................................36 10. Author's Address...............................................36 1. Introduction Transition from IPv4 to IPv6 is in the process. Work has been done to upgrade individual nodes (workstations) from IPv4 to IPv6. Also, there are established documents to make routers/switches to work to support IPv4 as well as IPv6 packets simultaneously in order to make the transition possible [1]. CIDR[2] based hierarchical architecture in the existing 32-bit system is supposed to be continued in IPv6 too with a large address space. There are documents/concerns over BGP table entries to become too large in the existing system [3]. There are proposals to upgrade Autonomous System number to 32-bit from Bandyopadhyay Expires October 26, 2015 [Page 2] Internet Draft MSHN and IPv6 April 26, 2015 16-bit to support the demand at the same time [4]. The challenge relies on how to make the transition smooth from IPv4 to a real IP world with least changes possible. ATM network performs faster than the network with IP switches. The difference becomes more prominent for real time applications. Whereas they have disadvantages as far as bandwidth usages is concerned compared to the IP-switch based network. This document tries to address approaches for IP-switch based network to process real-time applications as fast as ATM network also a mesh structured hierarchical network with flat address space for routing convenience. It provides a solution for site multihoming of stub networks. The term "real IP environment" is referred to an environment where hosts in a customer network will possess globally unique IP addresses and communicate with the rest of the world without the help of NAT[5]. A customer network may get IP services from more that one service providers also can maintain private IP addresses to communicate within users belonging to the same organization. 2. A Three-tier mesh structured hierarchical network Existing system is in work with Autonomous System (AS) and inter-AS layer with the approach of CIDR. In order to meet the need within the 32-bit address space, Autonomous Systems of various sizes maintain CIDR based hierarchical architecture. With the help of NAT, a stub network can maintain an user ID space as large as a class A network and can meet its useful need to communicate with the rest of the world with very few real IP addresses. With the combination of CIDR and NAT applied in the entire space, most of the part of 32-bit address space gets effectively used as network ID. This is how, 16-bit 'Autonomous System Number' is realized as insufficient in order to meet the need of growing customers. If the same gets continued with a larger network ID, load in the switches will become too high. With traditional CIDR based hierarchy, a node of higher prefix can be divided into number of nodes with lower prefixes. Each divided node can further be subdivided with nodes of further lower prefixes. This process can be continued till no further division is possible. The point worth noting is at each point the designer of the network has to preconceive the future expansion of the network with the concept in the mind that the resource can not be exhausted at any point of time. This phenomenon leads the designer to allocate resources much higher than whatever is needed which leads to a space of unused address space and the concept of H-D (host-density) ratio comes into play. The problem gets aggravated once resource gets exhausted by any chance. e.g. a node of prefix /16 can be divided with a number of nodes of prefixes /24. If any one of the nodes /24 gets exhausted, Bandyopadhyay Expires October 26, 2015 [Page 3] Internet Draft MSHN and IPv6 April 26, 2015 resources of other nodes of prefixes /24 can not be used even if they are available. Transition from private IP to real IP may not appear to be a simple task. This has happened due to the desperate attempt of the service providers to provide internet services with the help of NAT. e.g. a large educational institute meets its current requirement with 4 real IP addresses; one for its mail server, one for its web server, one for its ftp server and another one for its proxy server to provide web based services to all of its users. These four types of services are used by any organization of any size(it may be 400 or even 40000). In the current provider network these organizations are supported their need with 4 IP addresses and the CIDR based tree has been built using these components together. When private IP will be replaced with real IP, each customer network will require IP addresses based on its size and requirement. So, even if CIDR based architecture is maintained with real IP space, existing provider based network needs to be reorganized. The desired approach will be to assign address block that will be proportional to the sizes (bandwidth) of the ports of the switches of the provider network. As Autonomous Systems of various sizes are supported, Autonomous Systems and the nodes inside the Autonomous Systems can be viewed as graphically lying on the same plane within the address apace. If network can be viewed as lying on different planes, routing issues can be made simpler. If network is designed with a fixed length of prefix for the Autonomous System everywhere, routing information for the rest will get confined with the other part of the network prefix. Which means the maximum size of AS gets assigned to all irrespective of their actual sizes. This can be made possible with the advantage of using a large address space and dividing it into number of regions of fixed sizes inside it. Thus entire network can be viewed as a network of inter-AS layer nodes. Each node in the inter-AS layer can act either only as a router in the inter-AS layer or as a router in the inter-AS layer with an Autonomous System attached to it with a single point of attachment or as an Autonomous System with multiple Autonomous System border routers (ASBR) appearing like a mesh. Thus two tier mesh structured hierarchy gets established between AS layer and inter-AS layer with each AS having a fixed length of prefix. Based on the definition of Autonomous System, it is a small area within the entire network that maintains its own independent identity that communicates with the rest of the world through some specific border routers. In the similar manner, if a larger area (say region or state) can be considered as network of Autonomous Systems, that can maintain its own identity by communicating with the rest of the world through some border routers (say, state border router), mesh structured hierarchy can be established within the inter-AS layer. Bandyopadhyay Expires October 26, 2015 [Page 4] Internet Draft MSHN and IPv6 April 26, 2015 The inter-AS layer will be split into inter-AS-top and inter-AS- bottom. To maintain this hierarchy, each node of inter-AS-top needs to have multiple regional or state border routers (say, SBR) through which each one will communicate with the rest of the world in the similar manner an Autonomous System maintains ASBR. Thus, entire network will appear as a network of nodes of inter-AS-top layer. To maintain hierarchy, each node of the inter-AS-top needs to have a fixed length of prefix. i.e. each node of the inter-AS top will be assigned a maximum (fixed) number of nodes of Autonomous Systems. Thus, with three-tier mesh structured hierarchy in the network layer, network ID can be viewed as A.B.C. If pA, pB and pC be the prefix lengths of inter-AS-top, inter-AS-bottom and AS layers respectively, there will be 2^pA nodes at the topmost layer, 2^pB at the inter-AS- bottom layer and 2^pC nodes at the AS layer. Thus the entire space gets divided into a fixed number of regions and each region gets divided into fixed number of sub regions. This division is supposed to be made based on geography, population density and their demands and related factors. Let nMaxInterASTopNodes be the possible maximum number of nodes assigned at the top most layer and nMaxInterASBottomNodes be that at the inter-AS-bottom layer and nMaxASNodes at the AS layer. Where nMaxInterASTopNodes <= 2^pA and nMaxInterASBottomNodes <= 2^pB and nMaxASNodes <= 2^pC. 2.1. Route propagation With hierarchy established, routing information that gets established inside a node of inter-AS-top, does not need to be propagated to another node of inter-AS-top. Entire routing information of inter-AS- top layer needs to be propagated to inter-AS-bottom layer. So, each router of inter-AS layer will have two tables of information, one for the inter-AS-top and another for the inter-AS-bottom of the inter-AS- top node that it belongs to. BGP (with little modification) will work very well with a trick applied at the SBRs. Each SBR will not propagate the routing information of inter-AS-bottom layer of its domain to another SBR of neighboring domain. i.e. SBR of one top layer node will propagate routing information only of inter-AS-top layer to SBR of another top layer node. Inside a node of inter-AS- top, routing information of inter-AS-top and inter-AS-bottom need to be propagated from one ASBR to another neighboring ASBR. Inside a top layer node A, routing information of another top layer node B will have two parts; one for the list of SBRs through which a packet will traverse from top layer node A to B and another for the list of ASBRs through which the packet will traverse from one AS to another inside A. In terms of BGP, AS_PATH attribute will be split into two parts; one for the information of the top layer and another for the bottom Bandyopadhyay Expires October 26, 2015 [Page 5] Internet Draft MSHN and IPv6 April 26, 2015 layer. Within the same node A routing information of one AS to another AS will not have any top layer information. i.e. the top layer information will be set to as NULL. Similarly, each node of the AS layer will have three tables of routing entries. One for the inter-AS-top, one for the inter-AS- bottom and another for the routing information inside the Autonomous System itself. Introduction of hierarchy at the inter-AS layer reduces the size of the routing table substantially. With the availability of hardware resources if flat address space is maintained at each layer, problems related to CIDR can be avoided. With flat address space, no hierarchical relationship needs to be established between any two nodes in the same layer. So, all the nodes inside each layer can be used till they get exhausted. With flat address space (i.e. without prefix reduction), BGP tables will have maximum nMaxInterASTopNodes + nMaxInterASBottomNodes entries. IGP like OSPF has got provision to divide AS into smaller areas. OSPF hides the topology of an area from the rest of the Autonomous System. This information hiding enables a significant reduction in routing traffic. With the support of subnetting, OSPF attaches an IP address mask to indicate a range of IP addresses being described by that particular route. With this approach it reduces the size of the routing traffic instead of describing all the nodes inside it, but introduces another level of hierarchy. If subnetting concept can be avoided from the AS layer(with the additional overhead of computation inside the SPF tree), each area can be configured from a free pool of addresses based on its requirement dynamically. So, an AS can be divided into number of areas of heterogeneous sizes with the nodes from a free pool of address space. Similarly, the concept of area can be introduced in the inter-AS- bottom layer the way it works in OSPF. The area border routers in the inter-AS-bottom layer have to behave exactly in the similar manner the way an ABR behaves in OSPF. i.e. an area border router will hide the topology inside an area to the rest of the world and will distribute the collected information inside the area to the rest. It will distribute the collected routing information from outside to the nodes inside as well. In order to implement this, protocol running in the inter-AS layer (say BGP) will have to introduce a 'cost' factor. This cost factor can be interpreted as the cost of propagation of a packet from one AS to another. The protocols running inside AS layer (RIP/OSPF, etc) will have to the supply the cost information for a packet to travel from one ASBR to another. All the protocols must behave in unison for supplying this information. The cost factor is needed for a remote node while sending a packet to a node inside an Bandyopadhyay Expires October 26, 2015 [Page 6] Internet Draft MSHN and IPv6 April 26, 2015 area while more than one area border routers are equidistant from that remote node. Thus inter-AS-bottom layer (i.e. one inter-AS-top level node) can be divided into number of areas of heterogeneous sizes with nodes of AS from a free pool of address space. BGP adopts a technique called route aggregation. Along with route aggregation it reduces routing information within a message. In the similar manner, introduction of area inside inter-AS-bottom layer will not only reduce the complexity of the protocol, but will reduce the size of a BGP packet substantially. With this architecture, each node(router) inside an AS is represented as A.B.C. Each node may or may not be attached with a network which acts as a leaf node (i.e. a network will not act as a transit). In order to make use of user-id space properly and to support customer networks of heterogeneous sizes, the user-ID space needs to be divided as subnet-ID and user-ID. Profoundly, a VLSM (variable length subnet mask) type of approach has to be adopted at each node of an AS. So, each node of the AS layer will act as the root of a tree whose leaves are independent small customer networks which will act as stub. As the routing information of inter-AS layer as well as AS layer need not be passed inside any node of the VLSM tree, each router inside the tree should maintain default route for any address outside of its network. With this approach, load on each router of the service providers will become negligible. Protocols that supports VLSM with MPLS/VPN has to be implemented inside the tree (inside the VLSM tree, all the physical ports of a switch have to be configured with the subnet mask. So, mere MPLS on top of static routing table should do the rest). The fundamental assumptions based on which this architecture lies can be summarized as follows: i) Entire network can be viewed as a network of regions or states where each region or state can have its own identity by communicating with the rest of the world through some state border routers. Each region or state is a network of Autonomous Systems. Each region as well as each Autonomous System inside them will have a fixed (maximum) length of prefix. ii) Availability of hardware resources is such that flat address space can be maintained at the inter-AS layer. Introduction of mesh-structured hierarchy will have several advantages: o Load at each router will get reduced substantially. o Concept of CIDR style approach and complexity related to prefix reduction can be easily avoided. Bandyopadhyay Expires October 26, 2015 [Page 7] Internet Draft MSHN and IPv6 April 26, 2015 o Mesh structured hierarchy will make traffic evenly distributed. o Physical cable connection can be optimized. o Administrative issues will become easier. 2.2. Determination of prefix lengths With this architecture, IP address can be described as A.B.C.D where the D part represents the user id. Each router in the inter-AS layer will have two tables of information, one for the inter-AS-top and another for the inter-AS-bottom of the inter-AS-top node that it belongs to. Whereas, each node of the AS layer will have three tables of routing entries; one for the inter-AS-top, one for the inter-AS- bottom and another for the routing information inside the Autonomous System itself. In the worst case. a node inside an AS needs to maintain nMaxInterASTopNodes + nMaxInterASBottomNodes + nMaxASNodes entries in its routing table. The dynamic nature of allocating an area from a free pool of address space is more frequent at the AS layer than at the inter-AS-bottom layer. As OSPF supports all the features needed, it can be considered as default choice in the AS layer. Existing implementation of OSPF (Version 2) supports subnetting, by which an entire area can be represented as a combination of network address and subnet mask. With this approach, entire routing table gets reduced substantially. With the removal of subnetting, all the nodes inside an area will have an entry inside the routing table (OSPF Version 1). So the deterministic factor is what is the maximum number of nodes inside an AS OSPF can support once subnetting support gets removed. So the prefix length of AS layer will be determined by this factor of OSPF. With the introduction of hierarchy in the inter-AS layer, number of entries in the BGP routing table will get reduced substantially. Even if pA and pB both are selected as 16, number of routing entries come within the admissible range of existing BGP protocol. But, it is the responsibility of IANA to come out with a scheme how nMaxInterASTopNodes and nMaxInterASBottomNodes are to be selected. Each top level node will have nMaxInterASBottomNodes nodes. It will be a waste of address space if each country gets assigned a top level nodes (e.g. china has got a population of 1,306,313,800 people where as Vatican City has got only 920 according to a census of 2006). So a moderate value of nMaxInterASBottomNodes is desirable, with which larger countries will have a number of top level nodes. e.g. each state of USA can be assigned a top level node. With the introduction of area in the inter-AS-bottom layer, each top level node can be divided into number of areas of heterogeneous sizes. So, a group of neighboring countries with less population can share the address space of a top level node. Similarly, user-id space has to be decided based on the largest area VLSM tree should be spanned through. All Bandyopadhyay Expires October 26, 2015 [Page 8] Internet Draft MSHN and IPv6 April 26, 2015 these issues are completely geo political and have to be decided by IANA. 2.2.1. A pseudo optimal distribution of prefixes in a 64bit architecture In order to have optimal use of cable connections, length of the VLSM tree is expected to be as short as possible. Also any single organization may prefer to have its user id space to be under the same network id. So, a 16bit user-id may become insufficient for places like large university campus, where as 32bit will become too large. Hence, 24bit user-id will be a moderate one which is the class A address space in ipv4 (also used as the space for private IP). As published in 1998 [7], OSPF can support an area with 1600 routers and 30K external LSAs. So, 11 bits are needed to support this space. With the assumption that OSPF can support much more address space with the advancement of hardware technology as well as to keep the space open for future expansions, 12 bits are assigned for the AS layer. 16 bits are assigned for the inter-AS-bottom layer. So, if on the average, 16bit equivalent space gets used within the user-id space (i.e. one out of 256) and 8bit equivalent nodes gets used inside an AS (16% of 1600), for a top level node (with 16bit equivalent AS nodes), it will generate 2^40 IP addresses, which will give 8629 IP addresses per person in Japan (with a population of 127417200; Japan is at the 10th position from the top in the population list of the world). So, even if all the countries with population less than or equal to Japan are assigned a top level node and all the provinces/states of countries with larger population are assigned a top level node each, total number of nodes will come well under 1024. If a number of neighboring countries with lesser population shares a top level node, total number of top level nodes will come down further. This suggests that 62 bit equivalent (10(pA)+16(pB)+12(pC)+24(user-id)) space will be good enough for unicast addresses. This distribution expects OSPF to support 65K (64K+1K) external LSAs. 64bit address space may be divided into two 63bit blocks as follows: i. Global unicast addresses with the most significant bit set to 0. In order to separate out router address space from the host computers of customer networks, routers may be assigned a prefix 01 whereas the host computers will have prefix 00. With three-tier hierarchy, network ID is represented as A.B.C. Any router inside the VLSM tree including the root will have an address 01A.B.C.router-id. Where as a host interface inside a customer network will be represented as 00A.B.C.uid. As the number of nodes representing routers in the provider network will be way too less than the user-id space for the customer networks, in order to keep more space for unicast addresses of Bandyopadhyay Expires October 26, 2015 [Page 9] Internet Draft MSHN and IPv6 April 26, 2015 customer networks as well as to keep the option open for future expansion, entire 63 bit address space with the MSB set to 0 has been assigned to customer networks for unicast addresses. So, the distribution will look like 10(pA)+17(pB)+12(pC)+24(user-id). Router address space will be assigned from the address space with the MSB set to 1. One can think of a larger size for the VLSM tree. It has to be compensated with a smaller size for the inter-AS space. Say the distribution may look like 10(pA)+15(pB)+12(pC)+26(user-id). As the size of the user-id space (or the VLSM tree) is fixed, larger the size of the tree, larger will be the waste. This factor can be decided based on the data supplied (or suggested) by the service providers. ii. Address space with the MSB set to 1 will be distributed within the rest. Each of them will have a fixed prefix which will be determined with the consultation with IANA. This distribution will be based on the requirements and the work that have already been done in connection to IPv6 along with the following requirements: a) Router address space: Any node in the router address space will be designated with a prefix followed by A.B.C.router-id. b) Address space for multicasting: c) Address space for private IP: A 32 bit address space should be good enough for private IP. d) Provider independent address space: This space will be used for the customers who would like to retain their number even after changing their providers. With this architecture, addressing is based on the routing topology i.e. all unicast addresses will be based on the provider assigned address space. So, each of these provider independent addresses has to be mapped with an address from the global unicast address space. Section 2.4 describes issues related to PI addressing and IP mobility in detail. In order to provide support of IP mobility as well as provider independent addressing, each customer network has to be assigned some extra space along with their usual need. The actual amount of space to be reserved has to be determined by IANA. 2.2.2. Whether to go for a two-tier or three-tier hierarchy Establishment of hierarchy in the inter-AS layer reduces the size of BGP entries to a great extent, but leads to an improper use of address space due to geo-political reason. If hierarchy in the inter- Bandyopadhyay Expires October 26, 2015 [Page 10] Internet Draft MSHN and IPv6 April 26, 2015 AS space gets removed, entire 26bit (10+16) space will be available for a single layer and use of inter-AS space will be true to its sense, but will increase external LSA (and/or number of entries in the BGP table) dramatically. So, it depends on to what extent OSPF can support external LSAs. BGP expects the packet length to be limited to 4096 bytes. BGP manages to make it work with this limitation with the concept of prefix reduction in the CIDR based environment. As the number of inter-AS nodes increases, BGP has to change this limit in order to make it work in flat address space. The alternate will be to divide the inter-AS space into number of areas as defined in section 2.1. The area border routers will advertise the aggregated information to the rest of the world. BGP may have to incorporate both the options at the same time. As the number of nodes in the inter-AS layer increases, in order to reduce the number of entries in the routing table, inter-AS space has to be split into two separate planes. So, two-tier hierarchy can be considered as an interim state to go for three-tier hierarchy. If it so happen that current available data is good enough to support the present need, it will be worth to look for to what extent it can support in the future. Assignment of inter-AS nodes in two-tier hierarchy should be based on the geographical distribution as if it is part of three-tier hierarchy. Otherwise, introduction of three-tier hierarchy in the future will become another difficult task to go through. Based on the report of year 2011, BGP supports ~400,000 entries in the routing table. With this growing trend, BGP may have to change the limit of packet length even in a CIDR based environment. With the introduction of two-tier hierarchy, number of entries in the routing table will come down drastically and with the three-tier approach, it will come down further. 2.3. Issues related to Satellite communications Establishment of hierarchy in the inter-AS layer expects the only way any two autonomous systems in two different top level nodes communicate is through their SBRs. If two autonomous systems inside the same top level node communicate through satellite, it will be considered as a direct link between them. Whenever autonomous system 'ASa' of top level node 'A' communicates with autonomous system 'ASb' of top level node 'B' through satellite, they have to go through their state border routers. i.e. satellite port inside 'A' that communicates with a satellite port inside 'B' will be considered as state border router. If multiple such ports exists inside node 'A', all of them will be equidistant from any port inside 'B'. Which expects any satellite port inside 'B' to have prior knowledge of list of autonomous systems that will be under the purview of any port inside 'A'. So, all the satellite ports of 'A' have to exchange such group of information with all the satellite ports of 'B' and vice versa. These group of autonomous systems can be considered as a Bandyopadhyay Expires October 26, 2015 [Page 11] Internet Draft MSHN and IPv6 April 26, 2015 cluster of autonomous systems inside an area of a top level node. If number of such ports is small, some heuristics can be applied while assigning AS numbers in order to reduce the processing time during the circuit establishment phase. It will become difficult to maintain such heuristics once the number of such ports becomes large. So, in case of satellite communication, the advantage of establishing hierarchy inside inter-AS layer diminishes as the number of satellite ports increases. If any private corporate maintains its own satellite channel to communicate between its offices at distant locations, all of these offices are going to be considered as under the user-id space of its network. Service providers that provide satellite services to the end-site customers, can operate in the usual manner as they will provide connection to customer networks which will act as stub. 2.4. Issues related to PI addressing and IP mobility As far as implementation is concerned, provider independent addressing will be a costly affair. First of all in order to resolve the currently mapped location, there has to be a mechanism which is to some extent similar to the DNS entry resolution. Inside a customer network which is based on the provider assigned address space, routing of IP packets will be based on the provider assigned addresses. So, for every IP packet that is destined to a PI address will have a stack of addresses; the mapped address (or the care-of address) and the PI address. While initiating communication with a PI address, the mapped address has to be resolved first and then both the PI address as well as the mapped address has to be passed down to the transport layer. Transport layer needs to form a stack of addresses while filling up the IP packet. The above complexities can be avoided if the entire customer network is assigned a contiguous set of PI addresses. So, for the entire system, provider independent addressing has to be supported either based on the individual customer basis or on the entire customer network basis but not both. Customers who would like to have mobility support, the mapped address can be considered as the "Home Address" of the mobile node as defined in the specification of "IP Mobility Support"[8]. Once a node with PI address moves to a co-located care-of address[8], system needs to make decision based on PI address, its mapped address as well as the co-located care-of address. So, provider independent address with mobility support will be the costliest operation. If PI addresses are assigned on individual customer basis, protocol control block structure associated with socket needs to introduce another field 'fmpiaddr' to store the mapped destination address. It needs to have another field 'fcladdr', the destination node care-of address to support IP mobility. If foreign address is stationary and provider independent, both 'fmpiaddr' and 'fcladdr' will have the Bandyopadhyay Expires October 26, 2015 [Page 12] Internet Draft MSHN and IPv6 April 26, 2015 same value. The existing field 'faddr' which is used to address a foreign address will hold the value of PI address for a node with PI address. Similarly it will hold the value of "Home Address" of the mobile node if it is not provider independent. Protocol output routines like 'tcp_output' and 'udp_output' need these information to fill the IP packet. A new system call 'regrmtcladdr' needs to be introduced to store both PI address and the mapped address with the PCB. int regrmtcladdr(int sockfd, const struct sockaddr *mpiaddr, socklen_t mpiaddrlen, const struct sockaddr *claddr, socklen_t claddrlen); A client program needs to call 'regrmtcladdr' before it calls 'connect' to establish connection with its peer. 'regrmtcladdr'(or its system level routine) can be used by a correspondent node while a remote mobile node registers its care-of address with the correspondent node[8]. There could be several approaches to resolve the mapped address for a PI address. This issue needs to be discussed in a separate document. A function call needs to be introduced to get the mapped address. struct in_addr getmappedaddr(struct in_addr *piaddr); It is worthwhile to introduce a function call 'connrmtaddr' that will connect a remote address of any type. 'connrmtaddr' will check whether the address is provider independent and connect the remote site accordingly. int connrmtaddr(int sockfd, const struct sockaddr *dst, socklen_t addrlen); Assignment of contiguous block of PI address space to an entire customer network apparently do not make much sense. This is just equivalent to assigning PA address space to a customer network. So, assignment of PI address space to an entire customer network has to be avoided unless there is a real need that can not be solved (or avoided) by using PA address space. PI address assignment always have to be burdened with the look up procedure to resolve the mapped address even if an entire customer network gets assigned PI addresses. Assignment of PI addresses has to be restricted to a limited number of users. This limit has to be decided by IANA. As the number of users with PI addresses increases, complexities within the entire system increases proportionately. Bandyopadhyay Expires October 26, 2015 [Page 13] Internet Draft MSHN and IPv6 April 26, 2015 2.4.1 IP address aliasing An interface of a customer network may have several IP addresses (e.g. for a multihomed customer site, each interface will have multiple global unicast addresses also it may have a private address). This phenomenon is commonly known as IP address aliasing. A second type of aliasing is required to support IP mobility and provider independent addressing. For a mobile node that has been moved to a customer network which get services from two service providers and maintains private IP addresses, will have at least four IP addresses; provider one assigned unicast address, provider two assigned unicast address, private address and its permanent "Home Address". The "Home Address" will be aliased with one of the provider assigned addresses (i.e. the co-located care-of address). Similarly for a node with provider independent address will have four IP addresses. The interface address holding the PI address will be aliased with one of the provider assign addresses as its mapped address. If the node with PI address moved to a foreign site, will have a care-of address. The mapped address will be treated as the "Home Address". So the interface structure needs to have two additional fields to hold the values of care-of address and mapped address. The PCB structure will have two additional fields 'lmpiaddr' and 'lcladdr' to hold these information. In case a PI node that has not been moved, both 'lcladdr' and 'lmpiaddr' will have the same value. So 'lcladdr' will have the current provider assigned address that a foreign node needs to use for communication. The field 'laddr' that is used to hold the vale of local address will hold the value of PI address for a node with PI address; it will hold the value of "Home Address" of a mobile node in case it does not have a PI address. In order to support multihoming, an outgoing IP packet needs to be forwarded based on its source address (see section 3). In order to support this, an outgoing packet from a mobile node or a node with PI address needs to be stacked with the associated care-of address. A client application program needs to call 'getsrcaddr' (see section 3). to get the source address based on the destination address. The client program needs to to bind this address before communicating with its peer. The 'bind' system call needs to go through the interface list and fetch the associated structure to check whether the source address is aliased or not and needs to fill the values of 'lcladdr' and 'lmpiaddr' of PCB accordingly. Protocol output routines like 'tcp_output' and 'udp_output' need this information while filling up the IP packet. IP address stacking can be performed with the approach introduced in section 6.4 of RFC6275[11]. RFC6275 talks about the stacking of IP Bandyopadhyay Expires October 26, 2015 [Page 14] Internet Draft MSHN and IPv6 April 26, 2015 addresses for a destination address (Let us call it as type 0 stacking). Two more types of stacking need to be introduced; type 1 stacking where only source address will appear in the stack and type 2 stacking where both source address and destination address will appear in the stack with a particular type of ordering. Protocol output routine like 'tcp_output' or 'udp_output' needs to fill the IP packet in the following manner. If the socket contains a valid 'lcladdr', use 'lcladdr' as the source address and 'laddr' will appear in the stack. If the socket contains a valid 'fcladdr' use 'fcladdr' as the destination address and 'faddr' will appear in the stack. If only 'fcladdr' contains a valid address where as 'lcladdr' is NULL, use type 0 stacking. If only 'lcladdr' contains a valid address where as 'fcladdr' is set as NULL, use type 1 stacking. If both 'lcladdr' and 'fcladdr' contains valid addresses, use type 2 stacking. Protocol input routine like 'tcp_input' or 'udp_input' needs to process the packet in the reverse order based on the type of stacking. For type 0 stacking, use the address in the stack as the destination address; for type 1 stacking, use the address in the stack as the source address; for type 2 stacking use both source address and destination address from the stack. When TCP receives a SYN for connection establishment, it allocates a pcb and assigns the values for 'laddr', and related fields. During this phase, TCP also needs to check whether the local address is aliased or not and needs to fill the values of 'lcladdr' and 'lmpiaddr' accordingly. Similarly if destination address is found to be aliased, based on the stacking type, it needs to fill up the field 'fcladdr'. 2.4.2. Changes expected with the specifications related to IP mobility RFC6275 demands correspondent node binding from mobile nodes for route optimization. This binding is required when a connection gets established as well as when the mobile node changes it address space. There are application like HTTP which opens up multiple connections on the run time which are very short lived. If mobile nodes need to send binding messages for all the connections, network will be unnecessarily congested. This congestion can be avoided with the establishment of binding at the time of connection establishment itself. So, if TCP server happens to be mobile, it will set the value of 'lcladdr' in the stack while sending SYN+ACK. TCP client which initiates communication through 'connect' needs to set 'fcladdr' field on receiving TCP+ACK. With this approach correspondent node binding messages need to be sent only when a Bandyopadhyay Expires October 26, 2015 [Page 15] Internet Draft MSHN and IPv6 April 26, 2015 mobile node changes its position from one address space to another. Route optimization is not applicable to applications which are of multicast type. In these cases packets need to be forwarded with the mechanism of reverse tunneling with the approach of "IP Encapsulation within IP" as defined in RFC 2003. In order to support packet delivery with route optimization method as well as with "Encapsulating Delivery Style" based on the application type the protocol control block needs to introduce another field 'hagentaddr' to hold the address of the home agent of the mobile node. The interface structure also needs to have same field. The 'bind' system call needs to go through the interface list to fetch 'hagentaddr' to the PCB along with 'lcladdr' and 'lmpiaddr' as described earlier. So, protocol output routines like 'tcp_output', 'udp_output' need to fill up the packets based on the application type. In "Encapsulating Delivery Style" packets need to be formed in the following manner. The inner header will contain Source Address: Home address of the mobile node (i.e. laddr) Destination address: Address of the correspondent node (i.e. faddr if fcladdr = NULL; if fcladdr != NULL use type 0 stacking) The outer IP header will contain Source Address: co-located care of address of the mobile node (i.e. lcladdr) Destination Address: Address of the home agent of the mobile node (i.e. hagentaddr) Protocol field: IP in IP 2.5. Refinements over existing IPv6 specification As IPv6 was envisioned long before some of the newer technologies e.g. MPLS came into picture, some refinements can be made over the existing specification. These considerations are related to bandwidth usages and performance inside switches. Experimental results show that smaller packet size gives better result for the processing of RT packets. So, it is desirable to have IP packet header to be as small as possible. As described earlier, evaluation of the parameters nMaxInterASTopNodes, nMaxInterASBottomNodes and nMaxASNodes is geo- political and have to be decided by IANA. Once these parameters are determined with mutual agreements, values of pA, pB, pC and prefix length of user id can be determined. With 64bit address space, IP header will be reduced by 16 bytes. The 'flow label' field of IPv6 packet header may not be of any use with MPLS is in use. ATM used to have 4 priority classes. The first Bandyopadhyay Expires October 26, 2015 [Page 16] Internet Draft MSHN and IPv6 April 26, 2015 specification of IPv6 RFC-1883 used a 4bit type of service field along with a 24bits flow label field. These two were modified to a 8bit type of service field and a 20bit flow label field in the current spec RFC-2460. Too many priority classes may increase complexities to process inside switches. If type of service field of IPv6 header may be reduced to be of 4bit length as it was stated in RFC-1883 and 'flow label' field gets removed, another three bytes may be reduced from the IPv6 header. The field 'Hop Limit' has got a 8bit value in the existing spec. The role of this field needs to be discussed properly with a large address space. RFC4862[12] introduces the concept of "Stateless autoconfiguration" with the goal in mind that no manual configuration is required by individual machines before connecting them to the network. It generates a link local address with a link-local prefix and the link address (e.g. Ethernet/E.164 for ISDN) first. This link local address is used to configure global unicast address and any other configurable parameters based on router advertisement. Global unicast addresses are generated by the prefix supplied by the router advertisement and the link specific interface identifier. This identifier can be as large as 64 bit length. So irrespective of the size of the network (it may be 10000 or 100 or even less than that) every customer network will consume a 64bit equivalent addresses. This seems to be a huge blunder. What is expected is the length of the interface identifier is equivalent to support the number of nodes supported by that subnet. In order to achieve this the router itself or a server in that subnet needs to maintain a storage which will generate the interface identifier based on the request from individual hosts. It may be desirable that interface identifiers are generated from DHCP servers. With the option of generating interface identifier through DHCP, changes in the autoconfiguration process can be looked at as follows: From the point of view of a host, it can be considered as a two step process. Host needs to send Router Solicitations message to find out the presence of a router. Router Advertisement message should include an option field which will inform whether prefix information should be configured through Router Advertisement or through DHCP. Host needs to send a request message to get the interface identifier. If both the information needs to be obtained from a DHCP server they can be obtained through a single message. From the server's point of view, it needs to maintain a database for a mapping of the link-layer address and subnet specific interface identifier. Lifetime of an interface identifier has to be processed in the usual manner the way existing DHCP implementation treats IP Bandyopadhyay Expires October 26, 2015 [Page 17] Internet Draft MSHN and IPv6 April 26, 2015 addresses. There seem to be another possible danger to obtain prefix information through Router Advertisement. As the Router Advertisement comes in the form of ICMP messages, once it is received by the ICMP layer, it looses information from which interface the message has been received (This problem arises for hosts that are having multiple interfaces and not all of them are attached to the same subnet). So, autoconfiguration of a host has to be performed one interface at a time by making all other interfaces disabled. Once configuration of all the interfaces are done, all of them have to be enabled. If it is expected that hosts should reconfigure their addresses dynamically based on Router Advertisement message, Router Advertisement needs to generate a special message for a certain amount of time that needs to include old prefix and the corresponding new prefix in the message. In order to support multihoming (see section 3), prefix information needs to include the fields 'default router' and 'next hop address' to reach the default router for each of the prefixes. In a 64bit architecture, link-local address can be formed with a link-local prefix and link-layer address in a suitable manner; say it can be formed with a 16bit link-local prefix followed by a 48bit link-layer address. For hardware that supports more than 48bit addressing (say E.164), the least significant 48bits may be considered to generate link-local addresses. 2.6. Distributed processing and Multicasting With the inherent hierarchy involved in this architecture, distributed applications can also be structured in a suitable manner. Say, for a commonly used web based application a master level server will be there at every top level node. Any change that might happen in the application, has to be synchronized within these master level servers first. There might be servers at the middle layer (inside each inter-AS-bottom) inside each top level node. Once the changes get reflected at the master node, all the servers at the middle layer needs to update themselves with their master level node. This will reduce network traffic substantially. Inherent hierarchy in the architecture will also help establishing multicast tree in the similar manner. Work on these issues can be progressed only after this architecture gets approved. Bandyopadhyay Expires October 26, 2015 [Page 18] Internet Draft MSHN and IPv6 April 26, 2015 3. Solution for site multihoming Based on the definition of "multihoming" as stated in RFC3582[22], "A "multihomed" site is one with more than one transit provider. "Site-multihoming" is the practice of arranging a site to be multihomed." This is a general solution for site multihoming of stub networks in a real IP world irrespective of the framework supported by the service provider network. It is applicable to any version of IP, i.e. IPv4, IPv6 or any new generation of IP that may emerge by removing the drawbacks associated with IPv6(see section 2). Within a provider assigned address space, each customer network will possess as many global unicast address space as many service providers it gets connected with. So, an user interface of a host may have as many global unicast addresses as many service providers it will be connected with. Users will have an option of selecting the service provider while initiating a connection with the outside world. Users can maintain multiple connections through multiple service providers simultaneously. A customer network can maintain private IP addresses to communicate within its users and can share its load while maintaining VPN services. Customer networks can provide IP mobility support as well. There are many variants of UNIX systems (as well as real time operating systems) which make use of BSD source code for their implementation of TCP/IP stack. The solution given below highlights the changes required with the BSD release 4.4 source code with the notations used by IPv4. It addresses issues relevant to IPv6 wherever applicable. All other implementations of TCP/IP have to be updated in the similar manner. In this document the term "default router" will refer to the customer edge (CE) router that communicates with the provider network. Also the term "intermediate routers" will refer to all the routers apart from the CE routers. RFC1122[9] made an extensive study related to different aspects of multihoming. Some of the requirements suggested in that document related to UDP and the application layer were avoided for multihomed hosts in a connected network with a single gateway to reach the outside world. This was achieved by the implementation of TCP/IP by making sure that the interface address of an outgoing packet gets selected based on the route to be followed by the destination address. This criterion holds good in a connected environment with a single gateway to reach the outside world. Once more than one gateway Bandyopadhyay Expires October 26, 2015 [Page 19] Internet Draft MSHN and IPv6 April 26, 2015 comes into play to reach the outside world, either routing table of the entire world has to be brought in or needs some enhancements within the existing system to make the things work. Whenever a customer network gets service from more than one service provider, the customer network can be viewed as having multiple source-id (user-id) space. Each of these IP domain gets connected to different service providers through different routers. So each interface of customer network may have as many global unicast addresses as many service providers it is connected with. Number of routing entries in the routing table will (roughly) become a multiple of IP domains that it supports. Communication between any two hosts within the customer network will follow the traditional routing mechanism. In order to provide multihoming services it is needed that a host computer always forwards packets to the customer edge router associated to the same IP domain while communicating to someone in the outside world. i.e. if the interface of a host computer H receives an IP address 'addr1' and 'addr2' from two service providers P1 and P2 which are connected through routers R1 and R2 respectively, host H has to forward a packet to R1 while using its IP address as 'addr1' in order to send packets to the outside world. So, host computers as well as the intermediate routers have to use default routing based on the source domain of the source address in the IP header. In order to achieve this, host computers as well as intermediate routers need to have information related to its IP domain (net address/net mask) and the associated default router for all of its IP domains. They need to have a route entry per IP domain for all of its default routers. These information should be uploaded at the system start up time. As each interface is going to have multiple IP addresses, hosts need to have a provision to select its default IP domain (or default router) while initiating communications with the outside world. Users can select this option based on their need (i.e. whether a link is up/down/busy) dynamically. Users can execute multiple connections through different routers simultaneously as well. If no source address is specified by an application, source address has to be selected based on the outgoing interface and the default router as selected by the user. UDP (or RAW) based servers that need to support multiple clients simultaneously need to respond to a client's request with the same source address that the client had specified as the destination address. In order to satisfy this, system needs to introduce two system calls along with the existing system calls (i.e. read, write, send, sendto, recv, recvfrom) ssize_t recvwithdstaddr (int sockfd, char *buf, size_t nbytes, Bandyopadhyay Expires October 26, 2015 [Page 20] Internet Draft MSHN and IPv6 April 26, 2015 int flags, struct sockaddr *from, socklen_t *fromlen, struct sockaddr *fromcladdr, socklen_t *fromcladdrlen, struct sockaddr *dst, socklen_t *dstlen, struct sockaddr *dstcladdr, socklen_t *dstcladdrlen); 'recvwithdstaddr' receives data with destination address as specified by the sender. It is similar to 'recvfrom' with the additional field 'dst' related to the address of the receiving interface of the host. 'fromcladdr' and 'dstcladdr' will hold the values of co-located care- of addresses (see section 2.1) of source and destination. ssize_t sendwithsrcaddr (int sockfd, char *buf, size_t nbytes, int flags, struct sockaddr *to, socklen_t tolen, struct sockaddr *dstcladdr, socklen_t dstcladdrlen, struct sockaddr *src, socklen_t srclen, struct sockaddr *srccladdr, socklen_t srccladdrlen); 'sendwithsrcaddr' sends data specifying the source address of the outgoing interface of the host. It is similar to 'sendto' with additional parameters related to source address. It behaves like 'sendto' if no address is specified for 'src'. 'srccladdr' and 'dstcladdr' will hold the values of co-located care-of addresses of source and destination. All the UDP based servers that need to support multiple clients simultaneously, need to replace 'sendto' with 'sendwithsrcaddr' and 'recvfrom' with 'recvwithdstaddr'. It has been expressed in several documents including RFC4291[10], that a single interface will possess multiple IP addresses in a real IP environment. In these cases, all the UDP servers have to be updated with the system calls 'sendwithsrcaddr' and 'recvwithdstaddr' even if a customer site gets attached to a single gateway to reach the outside world. The same logic will apply to server applications with RAW sockets. Server applications that are TCP based should work in the usual manner. Another system call needs to be introduced to get the source address based on the destination address. struct in_addr getsrcaddr(struct in_addr *dst); Client applications need to use 'getsrcaddr' and 'bind' the source address before communicating with their peer. Routing of IP packets (in the ip_output module of the hosts and in Bandyopadhyay Expires October 26, 2015 [Page 21] Internet Draft MSHN and IPv6 April 26, 2015 the ip_forwarding module of the intermediate routers) need to be modified in the following manner. If destination address of a packet falls outside of its IP domains, it has to be forwarded to the default router based on the domain that the source address belongs to. If destination address of the IP header falls within any one of its IP domains, usual routing mechanism has to be followed. If customer network maintains private IP domain, communication using private IP has to be restricted within private IP space. Applications that need to establish multiple connections dynamically should call 'getsrcaddr' once and use the same source address for all of the connections subsequently. 3.1. Multihoming and IP Mobility For a mobile node, its co-located care-of IP address[8] has to be bound to one of the IP addresses supported by the service providers (if mobile node advertises more than one address, the home agent will get confused, also there are other implications). Transport layer must ensure that the 'home address' gets tightly coupled with that particular IP address. A mobile node in a foreign site will have all the IP addresses supported by the foreign site as well as its "Home Address". As the mobile node will also communicate with the outside world with its "Home Address", user should get a provision to choose its "Home Address" while initiating communication. Selection of default router and "Home Address" will be mutually exclusive. One should not interpret it as a selection of one of the global unicast addresses. This is just because a host may have multiple interfaces. If "Home Address" is selected for communication, the transport layer of the mobile node should use its care-of address as the source address and pass its "Home Address" as an option field in the stack. This is because multihoming expects the source address as the deciding factor for packet forwarding. The IP address of a node with a provider independent address have to be mapped with one of the global unicast addresses. So for the purpose of multihoming whatever will be applicable to a mobile node will also be applicable to a node with provider independent address. As the destination address may be a PI address, a client application needs to call 'connrmtaddr' after it calls 'bind'. As all the client Bandyopadhyay Expires October 26, 2015 [Page 22] Internet Draft MSHN and IPv6 April 26, 2015 applications (either TCP/UDP/RAW) needs to call 'getsrcaddr', 'bind' and 'connrmtaddr' their pattern will apparently look alike. 3.1.1. IP Address Stacking IP address stacking in IPv6 is performed with the approach introduced in section 6.4 of RFC6275[11] with slight modification. RFC6275 describes how to pass "Home Address" as well as co-located care-of address of the destination address if it happen to be mobile. The same approach has been extended to support IP address stacking for the source address and to support IP address stacking for both source address as well as destination address. The "Reserved" space in the type 2 routing header has been split into two parts; an one octet field to address the "Stacking Type" and the rest 3 octets are left as Reserved. Stacking Type is interpreted as follows: Stacking Type=0 Source Address: Address of the sender. Destination Address: co-located care-of address of the receiver. Address 1: Home Address/PI Address of the receiver. Hdr Ext Len=2. So, type 2 routing header for stacking type 0 will be as follows: +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Next Header | Hdr Ext Len=2 | Routing Type=2|Segments Left=1| +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |Stacking Type=0| Reserved | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | | + + | | + Address 1:Home Address/PI Address of the receiver + | | + + | | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ Stacking Type=1 Source Address: co-located care-of address of the sender. Destination address: Address of the receiver. Address 1: Home Address/PI Address of the sender. Hdr Ext Len=2. So, type 2 routing header for stacking type 1 will be as follows: Bandyopadhyay Expires October 26, 2015 [Page 23] Internet Draft MSHN and IPv6 April 26, 2015 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Next Header | Hdr Ext Len=2 | Routing Type=2|Segments Left=1| +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |Stacking Type=1| Reserved | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | | + + | | + Address 1:Home Address/PI Address of the sender + | | + + | | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ Stacking Type 2 Source Address: co-located care-of address of the sender. Destination Address: co-located care-of address of the receiver. Address 1: Home Address/PI Address of the sender. Address 2: Home Address/PI Address of the receiver. Hdr Ext Len=4. So, type 2 routing header for stacking type 2 will be as follows: +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Next Header | Hdr Ext Len=4 | Routing Type=2|Segments Left=1| +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |Stacking Type=2| Reserved | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | | + + | | + Address 1:Home Address/PI Address of the sender + | | + + | | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | | + + | | + Address 2:Home Address/PI Address of the receiver + | | + + | | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ Next Header 8-bit selector. Identifies the type of header immediately following the routing header. Uses the same values as the IPv6 Bandyopadhyay Expires October 26, 2015 [Page 24] Internet Draft MSHN and IPv6 April 26, 2015 Next Header field [18]. Hdr Ext Len 4 (8-bit unsigned integer); length of the routing header in 8- octet units, not including the first 8 octets. Routing Type 2 (8-bit unsigned integer). Segments Left 1 (8-bit unsigned integer). Stacking Type 2 (8-bit unsigned integer). Reserved 24-bit reserved field. The value MUST be initialized to zero by the sender, and MUST be ignored by the receiver. Address 1 Home Address/PI Address of the sender. Address 2 Home Address/PI Address of the receiver. IP address stacking in IPv4 is performed by introducing new IP option under the option class "Datagram or Network Control", i.e. 0. The option number is 16. The CODE(144) field is followed by one octet field "Stacking Type" followed by two octet reserved space (NULL) as padding followed by the address fields based on the Stacking Type. Stacking Type is interpreted as follows: Stacking Type=0 Source Address: Address of the sender. Destination Address: co-located care-of address of the receiver. Address 1: Home Address/PI Address of the receiver. Header Length:7 Format of IP address stacking option with stacking type 0 in the IP header will be as follows: +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | CODE(144) |Stacking Type=0| Reserved | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ + Address 1:Home Address/PI Address of the receiver + +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ Stacking Type=1 Bandyopadhyay Expires October 26, 2015 [Page 25] Internet Draft MSHN and IPv6 April 26, 2015 Source Address: co-located care-of address of the sender. Destination Address: Address of the receiver. Address 1: Home Address/PI Address of the sender. Header Length:7 Format of IP address stacking option with stacking type 1 in the IP header will be as follows: +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | CODE(144) |Stacking Type=1| Reserved | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ + Address 1:Home Address/PI Address of the sender + +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ Stacking Type=2 Source Address: co-located care-of address of the sender. Destination Address: co-located care-of address of the receiver. Address 1: Home Address/PI Address of the sender. Address 2: Home Address/PI Address of the receiver. Header Length:8 Format of IP address stacking option with stacking type 2 in the IP header will be as follows: +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | CODE(144) |Stacking Type=2| Reserved | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ + Address 1:Home Address/PI Address of the sender + +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ + Address 2:Home Address/PI Address of the receiver + +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 3.2. Implementation aspects Following changes are expected with the source code of BSD. Introduce ip_domain structure and some parameters as follows: struct ip_domain { struct in_addr net_addr; struct in_addr net_mask; struct in_addr def_router; }; #define MAX_IP_DOMAINS 16 short num_ipdomains; struct ip_domain *ipdomain[MAX_IP_DOMAINS]; Bandyopadhyay Expires October 26, 2015 [Page 26] Internet Draft MSHN and IPv6 April 26, 2015 If customer network maintains private IP domain (along with the user- id space provided by the service providers) and expects its communication to be confined within its own space, def_router field has to be set as NULL. Upload IP domain information for all of its IP domains during system start up. These domain information can be uploaded through router advertisement or through DHCP. The domain information should contain the next hop address to reach the corresponding default router as well. There has to be a provision to upload these information through 'sysctl' to configure them manually. Three new 'sysctl' routines have to be introduced under the 'ip' node of the MIB tree (i.e. under CTL_NET, PF_INET, IPPROTO_IP) IPCTL_NUM_DOMAINS, IPCTL_DOMAIN and IPCTL_DEFROUTER. Both IPCTL_NUM_DOMAINS and IPCTL_DEFROUTER are of type CTLTYPE_INT and IPCTL_DOMAIN is of type CTLTYPE_NODE. Using 'sysctl' IPCTL_NUM_DOMAINS has to be configured first. Configuration of IPCTL_NUM_DOMAINS has to populate IPCTL_NUM_DOMAIN entries of nodes under IPCTL_DOMAIN and for each of these nodes three MIB attributes DOMAIN_NET_ADDR, DOMAIN_NET_MASK and DOMAIN_DEF_ROUTER (each of type CTLTYPE_NODE) has to be allocated. All the routers as well as hosts that are having interfaces connecting to more than one subnets in private IP space (see section 2.3) need to be configured through 'sysctl'. Users should get provision to change IPCTL_DEFROUTER attribute dynamically. As each interface is going to have multiple IP addresses, IPCTL_DEFROUTER has to be assigned a value that will match any one of the entries assigned for DOMAIN_DEF_ROUTER. Add a route entry for all the default routers during system start up. 3.2.1. Processing of system call 'getsrcaddr' System call 'getsrcaddr' has to be processed in the following manner: If destination address of the IP packet falls outside of its IP domains { If destination address is from private address space { get source address as the private IP address of any of its interfaces. } If user has selected its "Home Address" instead of one Bandyopadhyay Expires October 26, 2015 [Page 27] Internet Draft MSHN and IPv6 April 26, 2015 of the default routers{ /*Applicable to IP mobility/PI address*/ return its "Home Address"; } else { get default router based on the selected 'default IP domain' use 'rtalloc' to get the next hop address for the def router. select source address based on the outgoing interface 'ia', and the 'default IP domain' as selected by the user. } } else { /* i.e. destination address is inside its IP domains */ use 'rtalloc' to get the next hop address for the destination address. If destination address is from private address space { select source address based on the outgoing interface and the private address assigned to it. } else if destination address is a link local address { select source address based on the outgoing interface and the link local address assigned to it. } else { select source address based on the outgoing interface and the domain that the destination address belongs to. } } 3.2.2. Changes required in ip_output and ip_forwarding modules Execute the following steps in the 'ip_output' routine of the IP stack before it calls 'rtalloc' for route look up. If destination address of the IP packet falls outside of its IP domains { get def router address based on the IP domain the source address belongs to. use 'rtalloc' to get the next hop address for the def router. Forward the packet to the next hop. } else { /* i.e. destination address is inside its IP domains */ follow the usual procedure to forward packets } Bandyopadhyay Expires October 26, 2015 [Page 28] Internet Draft MSHN and IPv6 April 26, 2015 In BSD, the 'ip_forwarding' routine calls 'ip_output'; so it should be left as it is. 3.2.3. Processing of protocol input routines and socket IO system calls Protocol input routines need to locate the socket/process in the usual manner with the 5 unit tuple (i.e. protocol, source address, source port, destination address, destination port). When a packet is received by a mobile node (at a foreign site), it can be received in two modes. It can be received directly from the correspondent node with the 'destination address' as the co-located care-of address and its home address in the IP stack (see section 4.1 of RFC6275[11]). In the second mode the packet can be received via the home agent using IP over IP. Once the IP layer receives a packet with IP over IP, it is supposed to strip off the outer header before passing the packet to the protocol input routine. In this case packet will be received by the protocol input routine with destination address as the home address of the mobile node with no information related to its care-of address. So, protocol input routine needs to check whether the destination address of the received packet belongs to any one of its IP domains. If it does not, it needs to find out the co-located care-of address by going through the interface list if it is not already found in the packet received. This information is needed by the TCP input routing while processing a SYN message. It is also needed by the UDP/RAW modules while processing the system call 'recvwithdstaddr'. While processing the output routines like 'sendwithsrcaddr', 'sendto', UDP/RAW modules needs to check the parameters related to source address, source port, destination address, destination port, care-of address of the source, care-of address of the destination in the protocol control block. Parameters in the PCB should prevail over parameters passed by the system call while forming the IP packet. 3.3. Multihoming, VPN and load sharing For a corporate, that maintains multiple offices and communicates within themselves through private address space using VPN, can do load sharing of outgoing traffic of private IP space by segregating private IP domain of each office into number of sub domains through suitable configuration. Let us consider one of its offices gets connected to two providers P1 and P2 and gets address space as 'unicastNetAddr1'/'unicastNetMask1' and 'unicastNetAddr2'/'unicastNetMask2' respectively. It also gets assigned private address space as 'privateDomainNetAddr'/'privateDomainNetMask' from its corporate. For load sharing, it wants to maintain two sub domains with its ID space Bandyopadhyay Expires October 26, 2015 [Page 29] Internet Draft MSHN and IPv6 April 26, 2015 as 'subDomainNetAddr1'/'subDomainNetMask1' and 'subDomainNetAddr2'/'subDomainNetMask2' respectively. Domain 1 gets associated with the default router CE1 and domain 2 gets associated with CE2. Host computers and intermediate routers will be configured in the following manner: All hosts of sub domain 1 will have three entries of ip_domain: 1: 'net_addr = 'unicastNetAddr1' 'net_mask = 'unicastNetMask1' 'def_router = CE1 2: 'net_addr = 'unicastNetAddr2' 'net_mask = 'unicastNetMask2' 'def_router = CE2 3: 'net_addr' = 'privateDomainNetAddr' 'net_mask' = 'privateDomainNetMask' 'def_router' = CE1 All hosts of sub domain 2 will have three entries of ip_domain: 1: 'net_addr = 'unicastNetAddr1' 'net_mask = 'unicastNetMask1' 'def_router = CE1 2: 'net_addr = 'unicastNetAddr2' 'net_mask = 'unicastNetMask2' 'def_router = CE2 3: 'net_addr' = 'privateDomainNetAddr' 'net_mask' = 'privateDomainNetMask' 'def_router' = CE2 All intermediate routers will have four entries of ip_domain: 1: 'net_addr = 'unicastNetAddr1' 'net_mask = 'unicastNetMask1' 'def_router = CE1 2: 'net_addr = 'unicastNetAddr2' 'net_mask = 'unicastNetMask2' 'def_router = CE2 3: 'net_addr' = 'subDomainNetAddr1' 'net_mask' = 'subDomainNetMask1' 'def_router' = CE1 Bandyopadhyay Expires October 26, 2015 [Page 30] Internet Draft MSHN and IPv6 April 26, 2015 4: 'net_addr' = 'subDomainNetAddr2' 'net_mask' = 'subDomainNetMask2' 'def_router' = CE2 If any of the CE-PE link fails, that particular CE needs to forward its outgoing traffic to the other CE whose CE-PE link remains active. This can be achieved through tunneling mechanism or by providing a hot link between the CEs. Forwarding of packets should be restricted to packets with private IP space. CE routers need to communicate within themselves at regular intervals and elect a leader within themselves. The elected leader should get privilege to forward private IP broadcast packets to other sites in order to avoid multiplicity. Broadcast packets that are originated only at the local site needs to be forwarded to the other sites. For a remote site, which is connected with PE routers RPE1 and RPE2, PE router of local site can load share its outgoing traffic by segregating its outgoing traffic with a suitable manner. If any of the link between RPE1 or RPE2 fails, it needs to forward all the traffic to the active link as well. 4. Processing of real time packets (QoS issue) Here is an attempt to come out with a solution for IP switch based network to operate in the most user-friendly manner to transport data traffic (IP) as well as real time (RT) traffic (as RTP[6] packet) in the existing 32-bit system. In case of IP routing/switching entire packet gets collected at the intermediate router/switch and forwarded based on the forwarding table. Inside the switch/router the variable length IP packet gets fragmented into smaller size frames at the ingress side. The frames gets transported through the switching fabric with proper priority mechanism (to support QoS) and then reassembled at the egress side and passed through the media for the next hop. In case of ATM, packets get fragmented at the ingress edge devices into small size cells. Entire packet gets transported as a stream of cells and gets collected at the egress edge device. The success of ATM over IP routing as far as speed is concerned is due to the fact that the latency gets reduced as the entire packet does not get collected, fragmented and reassembled at the intermediate nodes. So, in case of IP switch based network, if RT packets can be passed without getting fragmented inside the switch, better performance can be expected. i.e. one RT packet needs to get to fit inside one internal frame of the switch fabric. Additionally, to make this approach successful, maximum size of MPLS label stack has to be defined. Inside the switch all the IP packets will be assumed to carry same number of MPLS labels whether they are having one or the Bandyopadhyay Expires October 26, 2015 [Page 31] Internet Draft MSHN and IPv6 April 26, 2015 maximum in real sense. In fact, to reduce overhead, this limit should be the minimum number of labels needed to satisfy all sorts of features supported by MPLS. i.e. label stacking of depth n (without limit) needs modification. If minimum frame size is selected to fit one RTP packet, overhead becomes too high due to very large (40 bytes: 20 bytes IP + 8bytes UDP + 12 bytes RTP) packet header. Again, if large frame size is used, fragmentation loss becomes too high for the small size packets (say, 40 bytes IP packets). So, a compromise is needed that will give a better result based on the IP packet size distribution. Frame size is selected based on the minimum value of the overhead due to the fragmentation loss of data packet as well as the overhead as header of the RT packets. Studies show that primarily IP data packets of three different sizes are found common in nature. Almost ~50% packets of size 40 bytes (TCP ACK), ~20% packets of size 576 bytes (path MTU set by X.25) and ~30% packets of size 1500 bytes (path MTU set by ethernet) Other packets are less compared to the above three categories and almost evenly distributed. For the sake of simplicity of calculation, traffic of the first three categories are only considered. Payload of the data traffic is the actual IP packet size where as the payload of RT traffic is the payload inside RTP packet. If totBytes are to be transported across the internet and dataPcnt be the %of data traffic, totBytes*dataPcnt/100 = data traffic and (100-dataPcnt)*totBytes/100 = RT traffic; Out of data traffic 50% of 40 bytes length; 20% of 576 bytes length;& 30% of 1500 bytes length. If totDataPkts be the total data packets, totDataPkts*(50*40/100 + 20*576/100 + 30*1500/100) = totBytes*dataPcnt/100; or, totDataPkts*58520 = totBytes*dataPcnt; Let totBytes = 58520*100, for the ease of calculation; i.e. totDataPkt = dataPcnt*100; 40 bytes packets = 50*totDataPkt/100 i.e. 50*dataPcnt 576 bytes packets = 20*totDataPkt/100 i.e. 20*dataPcnt 1500 bytes packets = 30*totDataPkt/100 i.e. 30*dataPcnt RT packets = totBytes * (100 - dataPcnt)/100 = 58520 * (100-dataPcnt); Bandyopadhyay Expires October 26, 2015 [Page 32] Internet Draft MSHN and IPv6 April 26, 2015 If n is considered to be the depth of MPLS label stack, inside the switch, actual size of 40 bytes packet = 40+4*n bytes, 576 bytes packet = 576+4*n bytes & 1500 bytes packet = 1500+4*n bytes Let frameSize be the payload of a frame (excluding the frame header) inside the switch. If a RT packet fits exactly inside frameSize, RT packet payload = (frameSize-40-4*n) bytes; Total overhead = packet header overhead (of RT packets) + fragmentation overhead (of data packets); If a plot is drawn for frameSize = 40+4*n+1 to 1500+4*n for different dataPcnt (with dataPcnt=80 to 100) minimum of overhead are found at frameSize = (84, 101, 118, 126 and 152) for n==3; frameSize = (119, 127 and 152) for n==4 and at frameSize = (118, 127 and 152) for n==5. Actual data of the IP traffic has to be collected to get the best result. As dataPcnt increases minimum values are found at a lower frameSize and it gives better result with the higher range for lower dataPcnt. With average IP packet size 585 bytes, switches will encounter a loss of 4*(n-1) bytes for packets that will need only one label. In order to make this scheme work, a standard for maximum label stack size has to be defined. RTP packet size also has to be standardized. The same scheme is applicable to all the switching systems where IP packets get transported in hop by hop basis unlike the way it works in ATM networks. 4.1. Dual mode operation Inside ingress as well as in the egress card, packets need to follow certain functional steps. In order to maximize the output, a series of processing units work in pipeline mode for these operations. Ingress service cards need to act in dual mode to process RT packets and non-RT packets. i.e. the RT packets should follow a direct path that won't need fragmentation and related complexities before they are sent to the VOQs (virtual output queues, where from packets gets picked up to be sent to the switching fabric). Whereas other packets need to follow a different path for fragmentation operations. This will prevent a RT packet to be blocked by the fragmentation procedure of not-RT packets that arrive in the service card prior to the arrival of RT packet. So, mere mapping of RT packet size with the frameSize of switch fabric will not achieve the speed of ATM switches. Bandyopadhyay Expires October 26, 2015 [Page 33] Internet Draft MSHN and IPv6 April 26, 2015 Simulation studies show that significant improvement is achieved once RT packets are directly sent to VOQs after the operation of label processing. It will be worth to study by the hardware people to figure out whether entire set of data can be placed into queues based on their priorities and segmentation operation is done in each queue in parallel mode before putting the frames into their respective VOQs. Entire operation will be lot costlier, but simulation result shows that in such case, RT packets need not be restricted to fixed size cells. Standardization of label stack depth need not be imposed as well. 4.2. Expected changes at the application layer IP packets with size 576 in most of the cases come out of those TCP layers that do not process maximum path-MTU and takes the default one that was set during X.25. The 576 factor can be corrected very easily with path-MTU set to 1500. With the consideration that label switch path do not get changed very frequently in between two arbitrary network points for any particular type of packet, most of the applications are expected to become UDP based with negative ACK. TCP in turn might go through changes. Once this comes into effect, 40 bytes packets will come down drastically. Switch fabric frame size needs to be determined keeping these two factors in mind along with changes in IP packet header. With the existing 32-bit system, frame size (excluding the frame header) of 152 and 127 are most viable solution in general for label stack depth=3,4 &5. 5. IANA Consideration This is a entry level draft for proposed standard. Hence, IANA actions should come into play at a later stage, if needed. 6. Security Consideration This document addresses three major issues. The first part talks about architectural framework of the internet in real IP space. It need not have to discuss any security related issue. The second part that provides a solution for site multihoming. It does not introduce any security related issue. All the issues related to separation of locator and identifier that were addressed in RFC4218[20] are not applicable here but for common security related issues that any site may experience, one needs to consult with the "Site Security Handbook", RFC2196[21]. For issues related to IP Mobility, section 5 of RFC5944[8] has to be consulted. The third part of the document that deals with the QoS part also does not introduce any security related issue. Bandyopadhyay Expires October 26, 2015 [Page 34] Internet Draft MSHN and IPv6 April 26, 2015 7. Acknowledgments The author would like to thank to Professor Amitava Datta of University of Western Australia for his review and constructive comments. 8. Normative References [1] Nordmark, E. and R. Gilligan, "Basic Transition Mechanisms for IPv6 Hosts and Routers", RFC 4213, October 2005. [2] Fuller V., Li. T., "Classless Inter-Domain Routing (CIDR): The Internet Address Assignment and Aggregation Plan", RFC 4632, August 2006. [3] Huston, G., "Commentary on Inter-Domain Routing in the Internet", RFC 3221, December 2001. [4] Q. Vohra, E. Chen., "BGP Support for Four-octet AS Number Space", RFC 4893, May 2007. [5] Srisuresh, P. and K. Egevang, "Traditional IP Network Address Translator (Traditional NAT)", RFC 3022, January 2001. [6] Schulzrinne, H., Casner, S., Frederick, R., and V. Jacobson. "RTP: A Transport Protocol for Real-Time Applications", RFC 3550, July 2003. [7] J. Moy., OSPF Standardization Report, RFC 2329, April 1998 [8] C. Perkins, "IP Mobility Support for IPv4, Revised", RFC5944, November 2010. [9] R. Braden, "Requirements for Internet Hosts -- Communication Layers", RFC1122, October 1989. [10] R. Hinden, S. Deering, "IP Version 6 Addressing Architecture.", RFC4291, February 2006. [11] C. Perkins, Ed., D. Johnson, J. Arkko, "Mobility Support in IPv6" RFC 6275, July 2011. [12] S. Thomson, T. Narten, T. Jinmei, "IPv6 Stateless Address Autoconfiguration", RFC 4862, September 2007. Bandyopadhyay Expires October 26, 2015 [Page 35] Internet Draft MSHN and IPv6 April 26, 2015 9. Informative References [13] Rosen, E. and Y. Rekhter, "BGP/MPLS IP Virtual Private Networks(VPNs)", RFC 4364, February 2006. [14] Postel, J., "Internet Protocol", STD 5, RFC 791, September 1981. [15] Rekhter, Y., and T., Li, "A Border Gateway Protocol 4 (BGP- 4)",RFC 1771, March 1995. [16] Deering, S. and R. Hinden, "Internet Protocol, Version 6 (IPv6) Specification, RFC 1883, December 1995. [17] Moy, J., "OSPF Version 2", STD 54, RFC 2328, April 1998. [18] Deering, S. and R. Hinden, "Internet Protocol, Version 6 (IPv6) Specification", RFC 2460, December 1998. [19] Rosen, E., Viswanathan, A. and R. Callon, "Multiprotocol Label Switching Architecture", RFC 3031, January 2001. [20] E. Nordmark, T. Li, "E. Nordmark, "Threats Relating to IPv6 Multihoming Solutions", RFC4218, October 2005. [21] B. Fraser, "Site Security Handbook", RFC2196, September 1997. [22] J. Abley, B. Black, V. Gill, "Goals for IPv6 Site-Multihoming Architectures", RFC3582, August 2003. 10. Author's Address Shyamaprasad Bandyopadhyay HL No 205/157/7, Inda Kharagpur 721305, India Phone: +91 3222 225137 e-mail: shyamb66@gmail.com Bandyopadhyay Expires October 26, 2015 [Page 36]