Using a Distance-Vector algorithm
When using dynamic routing forming of route tables is made by routers automatically as a result of permanent accomplishment of a special algorithm of routing. B process of accomplishment of this algorithm the router passes between the neighbors information on routes known to him, obtaining from them in exchange similar information. After processing of the acquired information the router builds again or adjusts the existing routing table. As information on a status of routes enters on the router permanently, use of such algorithm provides permanent compliance of contents of a routing table to real network condition. B of dependence on how exchange of the routing information between adjacent routers is made, distinguish two types of algorithms of routing:
□ Distance-Vector (vektorarasstoyaniye algorithms);
□ Link-State (link state algorithms).
When using a Distance-Vector algorithm the router through the predeterminated time terms passes between adjacent routers contents of the routing table. When using a Link-State algorithm the router passes information only on state change of a system. Further explicitly concrete examples of these types of algorithms of routing will be reviewed.
Bo time of creation of the route table can be created several routes carrying in one network. In order that the router could select one of them as preferable, he shall use the generalized characteristic of quality of a route — a metrics (Metric).
Routing on the TCP/IP networks
Each algorithm of routing uses the algorithm of calculation of this characteristic. B the simplest case as a metrics of a route is used number of the nodes separating this router from a destination network. More difficult metrics consider characteristics of the physical ports making a route. Some algorithms of routing for increase in speed of information exchange allow to use at the same time several routes carrying to one network.
Routing domains and autonomous systems
Thus, dynamic routing represents a convenient and reliable mode of creation of route tables. When using dynamic routing, unlike static routing, tables can be constructed automatically, without any involvement of the administrator. The complete independence of routers, however, can be illogical and even to interfere with implementation of information security policy. In this case distribution out of limits of a responsible network of information on its inner pattern is restricted to nothing. From other side, the malefactor can pass in a responsible network the forged route messages to distort contents of routing tables and to break thereby the correct functioning of a system. It is well-known that for steering of the complex systems both the principle of centralization, and the principle of the complete decentralization is equally unacceptable. The truth, as usual, is between extremes. Optimum it will be possible to read the routing protocol capable to provide administrative restriction of distribution of the routing information. The limited field of distribution of the routing information is called routing domain (Routing Domain). For forming and distribution of the routing information routers of one domain use the uniform algorithm called by routing process. For identification mark of the messages containing the routing information digital identifier (number) of process of routing is used. For creation of a routing table process can use the routing information from the messages marked with appropriate number. B quality of an example can consider the network introduced in fig. 4.3.
M1 — M4 routers of this network are included into domain 100 while Ml and M5 — M7 routers form domain 200. Thus, the Ml router can obtain information on N2 — N4 and N5 — N7 networks while the M4 router has information on networks only of the domain (N1 — N4). Therefore, users of N2 — N4 networks will not be able to address resources of N5 — N7 networks and vice versa. Use
routing domains, in addition to information security support, allows to increase in some cases considerably effectiveness of use of transmission channels owing to cutting-down of volume of the passed routing information. For additional increase in level of information security some routing protocols use also the device of authentication of sources of the routing information.
Protocols of external and internal routing
The modern networks, as a rule, have the complex and distributed structure. For consolidation of fragments of these networks different transmission channels including rented can be used. It is accepted to call set of the networks which are under uniform managerial control autonomous system. For determination of internal routes in autonomous systems one or several routing protocols are, as a rule, used. B autonomous systems can call this class of protocols protocols of internal routing (Interior Gateway Protocol). Special routing protocols are used for determination of the routes which are going beyond autonomous system. We will call such protocols protocols of external routing (Exterior Gateway Protocol). Use of protocols of external routing allows the administrator to realize set of measures of increase
Routing on the TCP/IP networks
reliability and cost efficiency of information interaction with external systems. In case of a route choice, unlike protocols of internal routing, protocols of external routing analyze not only physical characteristics of transmission channels of a route. B number of the parameters used by the modern protocol of external routing for determination of quality of a route enter preference of a route, the sequence of passable autonomous systems and other parameters.
RIP and IGRP protocols
Routing Information Protocol is the most popular algorithm of routing of the class “Distance-Vector”. The earliest description of this class of algorithms was made by Ford and Fiilkerson. For this reason this algorithm is called sometimes Ford-Falkerson’s algorithm. The term “Bellman-Ford algorithms” can be also used for designation of algorithms of this type as the principles of creation of algorithms of this class were formulated in the monograph by Ford of “The equation of a dynamic programming”. This algorithm is one of the first algorithms of routing which were used on information networks generally and on the Internet — in particular.
Forming of a routing table in algorithms uDistance-Vectoru
Simply functioning of routers of the class “Distance-Vector” can be described as follows.
□ the Router builds the primary table of routing in which places numbers directly of the connected networks. This table contains the listed below fields:
• Address (Address). Network (IP) address of a network or node of assignment.
• Router (router). The network address the first the router on a route to a network or a node of assignment
• interface (interface). Network address or interface number of communication with the first router.
• Metric (metrics). Numerical characteristic of a route from 0 to 15. The value 0 corresponds directly to the connected network, the metrics 15 specifies an inaccessibility of a network or a node of assignment. B remaining cases corresponds to the number of the intermediate routers on a route to a network or a node of assignment.
• Timer (time clock). Index of relevance of information on a network or node of assignment. If information is not confirmed by a source during the established time frame, recording about a route is deleted from the table.
□ The router delivers the current version of a routing table issued in the form of the special message about up-dating (Update) to adjacent routers.
□ In case of acceptance of the similar message from the adjacent router the following operations are performed.
• If the message contains information on network N which is not in a routing table, the address of this network is entered in the table with the following field values: Router is the address of a source of the message, interface is the address of the interface which received the message in the field of Metric the value of the appropriate field of the initial message increased by weight factor of the interface is skidded (it is normal all weight factors of interfaces are received equal 1). Field value of Timer at framed recording is established equal to the trebled value of the period of the regular updates (90 c).
• If the message contains information on network N which is in a routing table, comparing of contents of the Router fields of the existing recording and the received message is executed. If the source of the routing information in both cases was same, the Metric field of the existing recording is modified on a usual algorithm by value of the appropriate field of the received message. The Timer field for the modified recording is formed as well as for newly created.
• If the message received from a new source contains information on the known network N, the router compares contents of the Metric fields of the existing recording and the received message. If the metrics of the existing route is more than metrics of a new route, former recording in a routing table is replaced with new. B an opposite case a routing table is not modified in any way.
□ B that case if field value of Timer at the existing recording became equal 0, recording is deleted from a routing table.
□ The processes described in two last points periodically repeat that allows to monitor configuration changes of a network dynamically.
Let’s consider philosophy of forming of a routing table
for the network figured in fig. 4.2. The initial table of the M3 router contains only directly the connected networks.
Routing on the TCP/IP networks
Example of an initial routing table of M3
Interface Metric Timer
10.0.1.1 0 90
10.1.2.2 0 90
10.5.6.2 0 90
Change of a routing table of M3 after receiving updates from M2 and M4!
Address 10.0.1.0 1. ^r-