13. FIG certain network topology, router R1 E1 E2 respectively connected through the interface, the LAN 432 shown in FIG. 1, the same domain
network 2, the router is connected via an interface L0 R2, and R2 is connected via a router to the Internet domain name server, the R1 U0
interface 202.118.2.1:R2 P is the IP address of the interface L0 is 202.118.2.2, P over the interface L1 is
130.11.120.1. P EO interface address is 202.118.3.1: P address of the DNS server is 202.118.3.2.
1) The P address space allocated to 202.118.1.0/24 LAN 1, LAN 2, P need to be allocated to each LAN
address number less than 10, give the allocation result, and write each LAN 1, LAN 2 subnet address, wide
address, subnet mask, P total number of addresses that can be assigned P P number of addresses and address ranges can be assigned.
2) Give R1 routing table, so that it expressly includes 1-to-LAN routing, LAN routing 2, domain name service
hosts the Internet's routing and routing.
3) Request from the polymerization technique employed, R2 is given to the LAN 1, LAN 2 reasons.
Solutions: 1) Core classless IP address is to use variable length network number and host number, and to indicate (i.e., the network number part 1, part of the host number 0) by the corresponding network address of a subnet mask according to the present problem. median 24, since the IP address is 32 bits, which is part of the host number eight. Thus subnet mask is 1,111,111,111,111,111 1,111,111,100,000,000, i.e., 255.255.255.0. Rule classless IP addresses according. Each segment has two addresses are not assigned: the whole number 0 indicates the host network address, host number 1 indicates the whole broadcast address. Thus eight host number is the number of hosts can be represented by 28-2, i.e., 254. The network to be divided into two subnets, each host 120 to the host of the number of bits X should satisfy the following three conditions: X <8, a host number as the bit length is divided into 8-bit network, certain X less than 8. 2X> 120, because of the need to accommodate 120 hosts meaning of the questions. X is an integer, the above solution equation, x = 7, the subnet mask is 1,111,111,111,111,111 11111111 1111111110000000, i.e. all 255.255.255.128 is divided into two segments:. 202.118.1.0/25 and 202.118.1.128/25.
14. Interpreting the Network Address Translation (NAT) works, how to implement NAT Traversal?
General NAT works: for out of the internal network, access to the public Internet IP datagrams, which will replace the source IP address of a legitimate public IP address of the NAT server, replacing the source port number, and replace the relational record to NAT conversion table; for returning from the public Internet IP datagrams, according to destination IP address and the destination port number to retrieve the NAT table, and using the retrieved internal private IP address and the corresponding port number replaces the destination IP address and destination port number and then forwards the IP datagram to the internal network.
15 . What motivates Ipv6 make is? Compared to its IPv6 IPv4 datagram format What are the characteristics? P160
motivation:
IPv4 address depletion hao
Features:
IP6 basic first minister of a fixed 40 bytes, and IPv4 contrast, some of the fields in the IPv4 header of IP6
has been deleted. First, in the header associated with the slice IPV4 field has disappeared, if an IPv6 packet can not pass
when through a link having a smaller MTU, directly to the router discarded, and the transmission source side transmits the datagram
number "packet too big" ICMP error messages. When the host receives an ICMP packet transmission will be a smaller data length
resend data packets IP16. IPV6 advantage of this improvement is to avoid fragmentation of the datagram, thereby speeding up the
P forwarding speed. Secondly, the header checksum field does not appear in the IPv6 datagram because the IP4 header
checksum for each hop need to be recalculated, it will increase end delay datagram sent with to the benefit
of the very small. Finally, the option field is no longer a part of the basic IPv6 datagram header, but it does not mean that the number of IPV6
reportedly no longer provide additional options, but by "Next Header" header field points to a special option, if not selected
item header, the "next header" field points to the upper layer protocol header, for example, TCP or UDP header
16. Description of link-state algorithm process and reasons for selection algorithm distance vector routing algorithm? P163 ~ 165
Link-state routing algorithm is a global routing selection algorithm, each router during route calculation need to construct the entire network topology. To construct the entire network topology, each router periodically checks for collecting charges directly connected link, and the router ID information and the like directly connected, configuration of link state packet broadcast to the entire network diffusion. Thus, each router in the network will periodically received link state packets to other routers wide interpolation, and link state information stored in the link-state database in each router. When the database to collect enough link state information, the router may be based on the link state information in the database, to build a network topology. Next, link-state routing algorithm is converted to the shortest path problem on the network topology. Typical shortest path algorithm in the figure is the Dijkstra algorithm, link-state routing algorithm is the Dijkstra shortest path algorithm.
Distance vector routing algorithm is an asynchronous, distributed generation of routing algorithm . In the distance vector routing algorithm, no one node to master the complete information of the entire network. Each node can be measured with all the neighbor nodes of
the direct link between the cost, and reaches its shortest distance to each destination node (which may be the shortest distance estimation) to (object minimum distance) distance vector form switching node to all the neighbors. Each node based on the distance between the direct link with a neighbor node, and neighbor switch over distance vector is calculated and updated for each shortest distance to the destination node, and then re-advertise the new distance vector to which all neighbors, until the distance vector does not change. Distance vector routing algorithms, each router iterative routing is asynchronous, the timing synchronization is not required in some calculation is performed between a plurality of routers, but whenever a new distance from the received neighbor when the vector, or a change in local link cost, routers only need to perform calculations, if calculated after point between the own distance vector has changed, then the new distance vector advertised to all neighbors. otherwise, without notice to the distance measurement process. In the distance vector routing algorithm, the iteration count the number of each router is uncertain, a node may need to choose a point of convergence on behalf of a lot of times before, and the other node may soon be converged.
17. Why illustrates the distance vector routing algorithm may arise count to infinity problem, what can be done abatement this problem?
Distance vector routing disadvantages are: good news travels fast, bad news travels slowly. To get routing protocols to work properly, if a link is interrupted (the cost becomes infinite), then all other routers should be immediately informed of the situation, but at a distance vector routing, which is to spend some time . This problem is called count to infinity (count to infinity). After several updates need to make all routers put the cost of interrupted recording link is infinite.
In an example of FIG. 42.7, prior to the link cost changes, D, (x) = 4, D (x) = 5. When the cost is increased xy link 50, node y lowest cost to update estimated at node x DA (x) = 6, since the distance before the node vector y receives the advertised node z the shortest distance, the node 2 statement of the arrival node x is 5, and therefore, the calculation D based on BF equation (x) = min501 + 5 = 6, since the node y distance vector changes (node y to node x estimated by the shortest distance becomes 4 6), the node sends the new y distance vector to the neighboring node (including node z). After the new node from the vector y, z receive node, also based on BF equation D (x) = mn140.1 + 6} = 7. Further, the node distance vector z then the new neighbor advertisement including its painted node Y), so the cycle continues, until after the 34th cycle, the node will z "consciousness" direct link to select a lower cost. Obviously, if the link is a large cost in FIG 4.27 xz, xy link cost also becomes large, then the above scenario, node y, z are in use for a long time reaches a false node of x ' the best route ", a phenomenon known as distance-vector routing count count to infinity problem.
Adopt measures:
Definition of infinity : distance-vector protocol is typically defined as 16 to infinity, or 16 hops is unreachable, but it also means that the distance vector can not be used for large systems. In all directions, the size of the network can not exceed 15 hops.
Division range : If the node B in accordance with its routing table that the best route to go through A X, then it does not need to route the advertisement of X to A, since this information is coming from A (A is already known a). Get information from node A, modify and then back to A, which is the root cause of confusion. Therefore, the Node B before transmitting the routing table to the routing information to be deleted A routing table as the next hop of A. In this case, the node A to retain the distance X is infinite. After that, when the node A sends to B its routing table, the node B will correct its routing table. System after the first update becomes stable, because the node A and B both know that X is not reachable.
Split range and poison reverse : using the split range strategy has one drawback. Typically, the distance-vector protocol uses a timer, if a long time without news about a route, it is necessary to delete the route from the routing table. In the scenario described above, when the route X to node B to A is deleted in its announcement, the node A can not guess which is due to the division range policy (as a source of information A), or because B recently has not received any news about X's. Dividing the range of policies may be used in combination with the reversal of toxic (poison reverse) policy. Node B can still inform about the value of X, but if the information source is A, put the distance change is infinite (16) as a warning: "Do not use this value, I know information about the route comes from you."
Commonly used in the art to poison reverse technique.
18. Comparison of RIP, the similarities and differences OSPF, BGP's. P168 ~ 170
https://www.cnblogs.com/jeasonrun/p/4206211.html
RIP is a distributed routing protocol based on distance vector based on the distance vector,
OSPF is based on a distributed link state routing protocol based on link state of
BGP is a path vector based on Border Gateway Protocol, an autonomous system as a node, connected to the BGP protocol
RIP: only the adjacent routers exchange state; information routers are routers exchange information on the whole known, the routing table; fixed time interval for exchanging routing information
OSPF: sending to all routers in the autonomous system information; information transmitted is present adjacent router link state of all the routers, i.e., router network topology periphery; only when the link state changes, only the router All routers send information by flooding method.
In RIP, each router only knows the distances to all the routers and the next hop router, do not know the whole network topology, and due to the different location, different for each routing table; when network failure occurs, to go through the longer to transmit this information to all routers, so-called "good news travels fast, bad news travels slow", slow convergence
In OSPF, since the information exchange link state adjacent routers, so each router can establish a link-state database (i.e., the whole network topology), it is consistent with the database in the whole network (this is called "synchronous link state database"). In OSPF, link database can be updated quickly, so the update process converges faster OSPF
RIP transport layer protocol UDP user datagram transmitted, OSPF directly transmit IP datagrams
