Presentation is loading. Please wait.

Presentation is loading. Please wait.

LANMAR for IPv6 Jason Chen Teresa Breyer Yeng-Zhong Lee.

Similar presentations


Presentation on theme: "LANMAR for IPv6 Jason Chen Teresa Breyer Yeng-Zhong Lee."— Presentation transcript:

1 LANMAR for IPv6 Jason Chen Teresa Breyer Yeng-Zhong Lee

2 Agenda Review of LANMAR IPv4 and IPv6 Addressing LANMAR implementation Problems faced Demonstration

3 LANMAR Combines features of Fisheye State Routing (FSR) and Landmark routing. Each set of nodes that moves as group forms a logical subnet. Each subnet elects a landmark. Logical Subnet Landmark

4 LANMAR Each node has unique logical identifier. Each identifier has subnet and host field. Host field is unique in subnet. Each node stores the following information: For each destination within fisheye scope:  Link State  Time Stamp For each destination within fisheye scope and for each Landmark node:  Next Hop  Distance

5 LANMAR Fisheye Scope Nodes exchange Link State only with their neighbors. Update rate decreases as hop distance increases. Routes within Fisheye scope are accurate. As hop distance increases routes accuracy decreases.

6 LANMAR Fisheye Scope Cont. Difference to FSR: LANMAR routing table includes only nodes within scope and landmark nodes. ⇒ Improves Scalability by reducing Routing Table size. Typically all members in logical subnet are within fisheye scope.

7 LANMAR Routing A packet to destination within neighbor scope is routed directly using local tables. A packet to destination out of scope is routed to corresponding Landmark using the subnet field of destination. Once the packet is in scope of destination, the direct route is found in local tables. Packet does not need to pass through Landmark.

8 LANMAR Routing Cont. Node address = {subnet ID, Host ID} Lookup local routing table to locate destination. If successful: Send packet directly. If failure: Look up landmark table to find destination subnet  LM1. Send a packet toward LM1. LM2 source destination Logical Subnet Landmark LM1 LM3 Long haul routing local routing

9 IPv4 32 bits Originally 4 classes of addresses: Shortage of Group A addresses 0 network host 10 network host 110 networkhost 1110 multicast address A B C D 1.0.0.0 to 127.255.255.255 128.0.0.0 to 191.255.255.255 192.0.0.0 to 223.255.255.255 224.0.0.0 to 239.255.255.255

10 IPv4 Header: ver Total length 32 bits data (variable length, typically a TCP or UDP segment) 16-bit identifier IP header checksum time to live source IP address head. len type of service flgs fragment offset protocol destination IP address Options (if any) 20 bytes

11 IPv6 Motivation: 32-bit address space exhaustion 128 bit  Fragmentation fields moved out of base header.  IP options moved out of base header. Header Length field eliminated  Header Checksum eliminated.  Type of Service field eliminated.  Time to Live Hop Limit, Protocol Next Header  Precedence Priority, added Flow Label field  Length field excludes IPv6 header.

12 IPv6 Cont. Header: fixed 40 bytes length. Destination Address (16 bytes) Version Priority Flow Label Payload LengthNext HeaderHop Limit Source Address (16 bytes, 128 bits)

13 IPv6 Cont. How IPv6 addresses look like: 2 128-1 : 0xffffffffffffffffffffffffffffffff Separator after 16 bits, omit leading 0x. 2 128-1 : ffff:ffff:ffff:ffff:ffff:ffff:ffff:ffff fe80:0000:0000:0000:0202:2dff:fe2e:3ad6 One sequence of adjacent 16 bit blocks of 0’s can be omitted. fe80::0202:2dff:fe2e:3ad6 Leading zeros of each block can be omitted. fe80::202:2dff:fe2e:3ad6

14 IPv6 Cont. 3 types of addresses. Unicast: identifies single interface. Anycast: identifies a set of interfaces. Multicast: identifies a group of interfaces replaces broadcast addresses in IPv4 prefix: 1111 1111 or ff for example link scope: ff02: …

15 LANMAR Implementation 2 complementary and cooperating routing schemes, 3 threads: A local “myopic” proactive routing scheme (we used RIP) that operates within limited scope centered at each node and exchange route information about nodes up to only a few hops (Fisheye Scope) ⇒ one thread for sending RIP messages. A long haul distance vector routing scheme that propagates the elected landmark of each subnet and the path to it into entire network ⇒ another thread for sending Landmark messages. Common thread for receiving messages.

16 LANMAR Implementation Advertisement messages  Proactive routing protocol Periodic routing update message is sent.  LANMAR routing control packet Landmark routes advertised to all nodes.  Local routing control packet Routes only advertised to nodes within its scope.

17 LANMAR Implementation Message process  A subnet mask is used to extract the group ID from a node’s IP address.  Each group is a logical subnet.  Landmark election is performed only among nodes in the same LANMAR group.  Local scope: RIP

18 LANMAR Subnet Masks IPv4: IPv6: Subnet Mask x x x x LANMAR Group IDNode ID Subnet Mask x x x x 1 1 1 1 0 0 0 0 64 bits Group ID Node IDNetwork ID 16 bits48 bits 111…11 000000……………………00000000000000………0000000

19 LANMAR Routing Table Routing table management Unicast Routing Table  Stores routing information for all unicast routing protocols running on router. Routing Table has 2 sub tables  Local routing table is flat without the concept of group or subnet.  Landmark routing table keeps only one entry from each group or subnet.

20 LANMAR Routing Table Cont. Dest.PrefixNext HopMet fe80:0:0:1333::3ad6128::0 fe80:0:0:1444::4352128fe80:0:0:1333::cf492 Landmark AddressPrefixNext HopMetric 0:0:0:1333::64fe80:0:0:1333::cf491 0:0:0:1444::64fe80:0:0:1444::43521 Local routing table Landmark routing table

21 Challenges in porting IPV4 LANMAR to IPV6

22 Data Structure Selected examples: struct sockaddr_in6 { sa_family_t sin6_family; /*AF_INET6*/ in_port_t sin6_port; /*Port number*/ uint32_t sin6_flowinfo; /*IPv6 traffic class and flow information*/ struct in6_addr sin6_addr; /*IPv6 address */ uint32_t sin6_scope_id; /*Set of interfaces for a scope*/ } instead of struct sockaddr { sa_family_t sin_family; /*AF_INET*/ in_port_t sin_port; /*Port number*/ struct in_addr sin_addr; /*IP address*/ }

23 Data Structures Cont. struct in6_addr { uint8_t s6_addr[16]; } where typedef unsigned char uint8_t; instead of struct in_addr { in_addr_t s_addr; } where typedef uint32_t in_addr_t; typedef unsigned int uint32_t; ⇒ Write library to compare, assign, logical &, and logical | uint8_t arrays.

24 Uintfunc Library void uintcmp(uint8_t array1[], const uint8_t array2[], int n); int uintcmp(const uint8_t array1[], const uint8_t array2[], int n); void uintand(uint8_t result[], const uint8_t array1[], const uint8_t array2[], int n); void uintor(uint8_t result[], const uint8_t array1[], const uint8_t array2[], int n); int uintzero(const uint8_t array[], int n);

25 Syntax for IPv6 Address and protocol family constants: AF_INET and PF_INET Replace with: AF_INET6 IPv6 address family. PF_INET6 IPv6 protocol family. Numerous other syntax changes.

26 Substituting IPv4 functions with newer IPv6 versions gethostbyname(): retrieves the network host entry referenced by a host name with an IPv4 address Replace with: gethostbyname2(): retrieves the network host entry referenced by a host name and its address family, which will be AF_INET6 inet_addr(): interprets a character string representing an address in Internet ``dot'' notation (IPv4) and returns a value suitable for use as an Internet address Replace with: inet_pton(): performs the same operation for both IPv4 and IPv6 address notations inet_ntoa(): interprets an IPv4 internet address and converts it to a character string. Replace with: inet_ntop(): performs the same operation for both IPv4 and IPv6 addresses.

27 Masks IPv4 SUBNET_MASK = 255.255.255.0 LANDMARK_MASK = 255.255.255.0 NORMAL_MASK = 255.255.255.255 IPv6 SUBNET_MASK6 = 0000:0000:0000:ffff:: LANDMARK_MASK6 = 0000:0000:0000:ffff:: NORMAL_MASK6 = ffff:ffff:ffff:ffff:ffff:ffff:ffff:ffff REVERSE_MASK6 = ffff:ffff:ffff::ffff:ffff:ffff:ffff

28 Routing Table Kernel IPv4 Routing Table (route)  Destination  Gateway  Genmask  Flags  Metric  Ref  Use  Iface Kernel IPv6 Routing Table (route –A inet6)  Destination/Prefix  Source/Prefix  Next Hop  Flags  Metric  Ref  Use  Iface ⇒ change part of code updating routing table (adding, deleting, and searching routes)

29 Routing Table Cont. Change part of code updating routing table (adding, deleting, searching for routes). Problem: new field Prefix. Calculate Prefix for group entry via LANDMARK_MASK6. For unicast routes prefix = 128.

30 Demonstration


Download ppt "LANMAR for IPv6 Jason Chen Teresa Breyer Yeng-Zhong Lee."

Similar presentations


Ads by Google