Routing concept

Table of contents

Introduction

The main task of switching is to ensure nodes connectivity within one network (see. InfiLINK 2x2 and InfiMAN 2x2: Switching). To organize communication between networks, different class of devices (routers) must be used (see Figure 1). This article describes the applications areas and configuration of Infinet devices used as routers.

Terminology

• Switching - the process of connecting subscribers through intermediate devices. In most modern networks, frame switching is done based on the Ethernet header (destination MAC address and vlan ID). In example (Figure 1a) data exchange between PC-1 and PC-2 is performed based on MAC addresses. In this article, the terms switching and L2 data transmission technology are identical.
• Switch - the device that performs switching.
• Routing - the process of determining the data transmission path between nodes in different networks, which is the best according to some criteria. Most modern networks route packets based on the IP header (destination IP address). In example (Figure 1b) data exchange between PC-1 and PC-2 is performed based on IP addresses. n this article, the terms routing and L3 data transmission technology are identical.
• Router - the device that performs routing.
• Local network - the network part that is in the responsibility area of the organization. The organization's employees are responsible for assigning IP addresses to devices on this network, an address conflict is very unlikely.
• Global network - network having global scale. Usually, the Internet is understood as a global network. Since many local networks are connected to the global network, the IP addresses allocation is performed centrally by special organizations.

Switching

Let's look at the difference in processing service headers for data transmission by switching and routing using an example (Figure 1).

In the scenario when PC-1 sends data to PC-2 (Figure 1a), PC-1 fills in the service fields following way:

• Destination MAC address: MAC address PC-2 - MAC-2;
• Source MAC address: MAC address PC-1 - MAC-1;
• Destination IP address: IP address PC-2 - IP-2;
• Source IP address: IP address PC-1 - IP-1.

The switch receives a frame from PC-1 and redirects it to PC-2 in accordance with the switching table. Thus, data transmission is performed based on the Ethernet service header, since transmission is at the data link level. This mechanism is called switching.

In the scenario when PC-1 sends data to PC-3 (Figure 1b), PC-1 fills in the frame service fields following way:

• Destination MAC address: router MAC address - MAC-R2;
• Source MAC address: MAC address PC-1 - MAC-1;
• Destination IP address: IP address PC-3 - IP-3;
• Source IP address: IP address PC-1 - IP-1.

The switch receives such a frame and transmits it to the router in accordance with the switching table. The router receives the frame, decapsulates the IP packet and transmits it to LAN-2. In this case, service headers will be set in the following way:

• Destination MAC address: MAC address PC-3 - MAC-3;
• Source MAC address: outer MAC address - MAC-R2;
• Destination IP address: IP address PC-3 - IP-3;
• Source IP address: IP address PC-1 - IP-1.

Note that the IP packet header is left unchanged, the receiver and sender MAC addresses in the Ethernet frame header are changed. This operation was performed because MAC addresses are used to transfer data within the same local network, i.e. when transferring data between different local networks, the MAC addresses will always be replaced. This data transfer mechanism is called routing.

Routing

The main networks function is the ability to organize communication between arbitrary nodes connected to this network. Using for these tasks the packet switching technologies associated with the link layer of network interaction model has a number of disadvantages:

• There is a risk of loops appearance when using some data-link protocols such as Ethernet. The risk can be minimized using third party tools such as STP, but is not limited by standard Ethernet facilities.
• The broadcast traffic amount depends on the number of devices connected to the network. To ensure that the amount of broadcast traffic in the total traffic is not large, the devices number connected to one broadcast domain should be limited. Thus, all network devices cannot be connected to the same broadcast domain, which makes impossible using L2 layer protocols to organize global device connectivity.
• Switches to transmit data operate on Ethernet frames, the headers of which contain the source and destination devices MAC addresses. Each entry in the switching table contains the MAC address of the device interface and does not support the mechanism for grouping these addresses. Thus, ensuring global connectivity will require switching tables, which include the MAC addresses of all devices in the world at each network node.

The IP network layer protocol, which is widely used to provide connectivity in large and global networks, lacks these disadvantages. IP is not a replacement for Ethernet, these protocols work together and perform different functions: Ethernet provides data transfer within the communication channel, IP is responsible for global addressing and node communication.

Currently, two versions of the IP protocol have become widespread: IPv4 and IPv6. Since Infinet devices currently support only the IPv4 protocol, further article will contain the description of only this version.

IP protocol

IP address

The IP protocol provides for using 32 bits for addressing nodes in the network, which are usually divided into four octets and written in decimal form, separating octets with dots (Fig. 2). IP addresses examples:

• 10.94.200.7
• 192.17.0.0
• 201.15.2.255

Network mask

IP provides the grouping of addresses on a network using network masks. A netmask is applied to an IP address, dividing it in two parts: a network ID and a host ID. Devices connected to the same network will have the same network ID and different host IDs. To ensure the network ID is matching on all devices, use the same network mask values when configuring devices. Host IDs set allows inferring the number of devices that can be connected to this network and their IP addresses.

The network mask has 32 bits and is written in the same way as the IP address with one difference: the mask consists of a one bits sequence followed by zero bits, i.e. the set of masks is preset and contains 33 values: from 0 to 32. The finite range of possible values allows to write the network mask in an abbreviated form, in which the number of single bits in the mask is indicated after a slash (see the table below).

One bits in the network mask define the network identifier: the bits of the IP address corresponding to one bit values of the mask must be fixed and cannot be changed. The remaining bits of the IP address, corresponding to the zero bit values of the mask, can take arbitrary values and determine the host ID.

When configuring devices connected to the network, IP addresses are not used without the network mask, since routing rules imply a different approach when transferring data to a device from "own" network and to other devices (see Switching). Note that the network mask is indicated in the device configuration and is not transmitted in the service header of the IP packet.

Addresses types

The IP address can be divided according to several criteria:

• by application area;
• by belonging.

By the application area, addresses can be divided in two large groups: public and private (Figure 3). Global connectivity can only be established between public addresses, i.e. private addressing is used on the enterprise local network, and public addressing is used on the Internet. The public address is unique, private addresses can be reused, i.e. devices PC-2 and PC-6 may have the same address and this is not a problem, since there is no connectivity between LAN-1 and LAN-2. However, addressing within the same local network must be unique, i.e. the addresses of PC-5 and PC-6 must be different.

In addition to public and private addresses, several service ranges are allocated, for example, to transmit multicast traffic, loopback interface traffic, etc.

By belonging in any network, the following addresses can be distinguished:

• Network address: the address assigned to this network. Often the network addresses are used in device routing tables, as it is shown below. The lowest address from the allowed range is used as the network address: in example 1 - 10.94.200.0, in example 2 - 192.17.0.0.
• Broadcast address: recipients of this address are all devices connected to the network. A packet with a network broadcast address set as the destination will be delivered to all devices connected to this network. The highest address from the allowed range is used as the broadcast address: in example 1 - 10.94.200.255, in example 2 - 192.17.0.3.
• Nodes addresses: addresses that can be assigned to network interfaces of devices connected to the network. All allowed addresses can be used as node addresses, except for the network address and the broadcast address: in example 1 - 10.94.200.1-10.94.200.254, in example 2 - 192.17.0.1-192.17.0.2.

Place of the router in the network

Figure 3 does not have the elements to connect networks to each other and to transfer data between networks using IP addressing. Such elements are called routers (Figure 4). Usually, a router connects several networks of an arbitrary type, not just public and private, as shown in the example.

The routers have following key features:

• The main function of a router is to transfer data between the connected networks.
• The router is connected to the network by connecting one of the router's interfaces to the network and assigning an IP address from the allowed range to this interface. Both physical and virtual interfaces can be used.
• When transmitting data, the router is guided by routing table.
• Data within the network are transmitted using switching technology, and between networks - routing, i.e. IP and Ethernet are complement each other, as mentioned before.
• For user data, the router is an intermediate device and does not change the source and destination addresses. The packet source sets the source and destination IP addresses.
• The router analyzes only the destination address to find a destination in the routing table.The source address in the service header is set to allow the recipient to send a response packet.
• The routing table is not only in specialized network devices, but also at end nodes. For example, on a Windows software controlled PC, the routing table can be displayed by running the "route print" command at the command line.

Routing table

Let's look at the network diagram (Figure 5), which includes the following elements:

• Local network LAN-1 to connect network devices PC-1 and PC-2:
  • 192.168.1.0/24 addressing is used in the network;
  • 192.168.1.10/24 is assigned to PC-1;
  • 192.168.1.20/24 is assigned to PC-2;
  • 192.168.1.1/24 is assigned to R1.
• Local network LAN-3 to connect network devices PC-3 and PC-4:
  • 172.16.3.0/28 addressing is used in the network;
  • 172.16.3.2/28 is assigned to PC-3;
  • 172.16.3.4/28 is assigned to PC-4;
  • 172.16.3.1/28 is assigned to R3.
• Local network LAN-2 to connect routers R1, R2 and R3 with each other:
  • 10.10.2.0/29 addressing is used in the network;
  • 10.10.2.1/29 is assigned to R1;
  • 10.10.2.2/29 is assigned to R2;
  • 10.10.2.3 is assigned to R3.
• R2 router connection to the WAN global network:
  • 45.94.77.7/25 is assigned to eth0 interface connected to WAN.

The routing table is an address directory of networks. It contains the location of the networks used for packets transmitting. The routing table may not contain the exact location of a particular network, but there are network interface through which the destination network can be reached. This logic is used by all routers along the traffic path, i.e. if there are 8 routers on the packet path, then each of them has information only about the next router along the way, and this information is contained in the routing table.

The routing table includes the following columns (Table 2a-c):

• Network address: the packet destination address specified in the service header is checked for belonging to the network whose address is indicated in the table. If the destination belongs to this network, than the current table entry can be used for data transmission.
• Gateway address: the next router address there the packet will be forwarded.
• Output interface: the network interface for the packet transmission.
• Distance: in networks with redundant communication channels, there are several paths to the same network. These routes can be obtained from one or several sources, however, only one of these routes should be placed in the routing table. To prioritize routes from different sources, use the Administrative Distance parameter (or Distance), which means the level of trust to this source. The route from the source with the lowest Distance value will be added to the routing table, as a lower Distance value means a higher level of trust. General recommendations for Distance values are followed by most manufacturers of network equipment (Table 3).
  
• Metric: a route to the same network can be obtained not only from different sources, as mentioned above, but also from the same. These routes are prioritized using Metric value when added to the routing table. Each routes source calculates the metric using different algorithms, so the metrics from different sources cannot be directly compared.

Routing table management

Each router along the packet path has the routing table management algorithm. The algorithm is following:

• Step 1: the destination address is checked for belonging to networks, which entries are in the routing table.
• Step 2: if few records satisfy the requirement of step 1, the "narrowest route" is selected, i.e. entry with maximum netmask value. For example, mask /24 is narrower than /8.
• Step 3: if at the step 2 there are several routing table entries with the same network masks, the Distance parameter is compared. The lower the value of this parameter, the higher the route priority.
• Step 4: if at the step 3 there are several entries in the routing table with the same Distance values, the metrics are compared. The lower the metric value, the higher the route priority.
• Step 5: if there is no entry in the routing table that meets the requirements of step 1 and there is no default route, the packet is dropped.

Routing tables management examples

Let's look at the examples of the routing table management in various scenarios (Figure 6a-c).

Scenario 1 - connecting PC1 to an FTP server running on PC2 (source - 192.168.1.10, destination - 192.168.1.20)

• Step 1a: PC1 generates a packet with the destination address of PC2 and transfers it for processing to the L2 layer of the network interface.
• Step 1b: The L2 layer of the PC1 network interface verifies that the destination belongs to the source network. Since PC1 and PC2 belong to the same network, the the PC2 network interface MAC address is set in the Ethernet header. The generated frame is sent to Switch1.
• Step 1c: The switch transmits the PC2 frame according to the switching table.

Data are transmitted within the same network using switching technologies, router R1 does not participate in this process.

Scenario 2 - checking the availability of PC3 from PC1 (source - 192.168.1.10, destination - 172.16.3.2)

• Step 1a: PC1 generates a packet with the PC3 destination address and sends it for processing to the L2 layer of the network interface.
• Step 1b: The L2 layer of the PC1 network interface verifies that the destination belongs to the source network. PC1 and PC3 belong to different networks, so the R1 router MAC address is set in the Ethernet header as the destination MAC address. The generated frame is sent to Switch1.
• Step 1c: Switch1 transmits frame to R1 in accordance with the switching table.
• Step 2a: Router R1 goes through the routing table: two entries match the destination address, 172.16.3.0/28 and 172.16.3.0/30. Since mask /30 is narrower than /28, R1 will redirect the packet to the 172.16.3.0/30 network. Note that if the packet destination were PC4, a different entry in the routing table would be used, even though PC3 and PC4 belong to the same network.
• Step 2b: R1 router forwards the Ethernet frame to R3 router. Source and destination IP addresses remain unchanged, as a source MAC address is set the eth2 R1 MAC address, as destination MAC address - the eth3 R3 MAC address.
• Step 2c: The switch forwards the received Ethernet frame to router R3.
• Step 3a: Router R3 goes through the routing table: the destination address matches the 172.16.3.0/28 network.
• Step 3b: Router R3 sends an Ethernet frame to Switch3.The source and destination IP addresses remain unchanged, as a source MAC address is set the eth1 R3 interface MAC address, as the destination MAC address - the PC3 network interface MAC address.

Scenario 3 - connection with the "infinetwireless.com" from PC1 (source - 192.168.1.10, destination - 82.151.200.119)

• Step 1: PC1 generates a packet with the destination address 82.151.200.119 (the IP address of the server where the infinetwireless.com website is available). The packet is sent to router R1.
• Step 2: R1 router goes through the routing table: there are no networks in the table that match the destination address, so the default route should be used. The router sends the packet to R2.
• Step 3: R2 router goes through the routing table: there are no entries matching the destination address, so the default route is used and the packet is sent to a router outside the local network (WAN).

Filling the routing table

Speaking about the mechanisms for filling the routing table, two terms should be added:

• RIB (routing information base) - routing information data obtained from all sources.
• FIB (forwarding information base) - data forwarding table used to handle transit traffic. FIB is generated from RIB by filtering and combining routing information (Figure 7).

The sources routing information are:

• Operating system routes: service networks used by the device operating system. For example, the loopback interface network 127.0.0.0/8.
• Directly connected networks: networks to which the device is connected directly, i.e. device interfaces are associated with IP addresses that belong to these networks. A Distance parameter of such routes is minimal and equals to 0 (Table 2a-c).
• Static routes: routes added to the table manually. Distance of such routes is equal to 1 (Table 2a).
• Dynamic routing protocols: routes obtained using dynamic routing protocols. A Distance value is assigned to each dynamic routing protocol, examples are shown in Table 3.

Routing table at Infinet Wireless devices

Depending on the family, Infinet Wireless devices support different sources of routing information:

Routing table output

Further in the article we will use the tools for outputting and analyzing routing information. These tools depend on the device family and will be shown below.

Routing tables for InfiLINK 2x2, InfiMAN 2x2 families devices

InfiLINK 2x2 and InfiMAN 2x2 families devices supports routing settings for management traffic and for customer traffic, moreover, static routes and dynamic routing protocols are supported.

Routing information output can be performed in two ways:

• Web interface: go to the "Network settings → Routing parameters" (Figure 8a). The interface allows to view only static routes.
• Command line: the "nestat -r" command displays FIB data. Also there are commands allowing to evaluate routing information by separate sources, which will be described in the following sections.

Unknown node#1> netstat -r
Routing tables
Destination        Gateway            Flags     Refs     Use  Interface
10.10.10.0/24      link#6             UC          0        0  svi1
10.10.10.101       00:0c:29:40:72:d0  UHL         0        1  svi1
10.10.10.254       link#6             UHL         0        0  svi1
10.10.20.0/24      link#2             UC          0        0  eth0
10.10.20.101       00:0c:29:40:72:d0  UHL         1     1307  eth0
127.0.0.1          127.0.0.1          UH          1        0  lo0
224.0.0.0/8        127.0.0.1          UGS         0        0  lo0

Routing tables for InfiLINK XG, InfiLINK XG 1000 families devices

InfiLINK XG, InfiLINK XG 1000 families devices supports routing configuration of management traffic only. Default gateway and static routes can be set. Routing information output can be performed in two ways:

• Web interface: go to the "Network access" section (Figure 8b).
• Command line: run the "nestat -r" command.

#1> netstat -r
Routing tables
Destination        Gateway            Flags     Refs     Use  Interface
10.10.10.0/24      link#2             UC          0        0  mgmt
10.10.10.101       00:0c:29:40:72:d0  UHL         1      512  mgmt
10.10.10.254       link#2             UHL         1        0  mgmt
10.10.20.0/24      10.10.10.254       UGS         0        0  mgmt
127.0.0.1          127.0.0.1          UH          0        0  lo0
224.0.0.0/8        127.0.0.1          UGS         0        0  lo0

Routing tables for Quanta 5, Quanta 70 families devices

Quanta 5, Quanta 70 families devices support only routing configuration for management traffic, allowing to set a default gateway. Routing information output can be performed in two ways:

• Web interface: go to the "Network" section (Figure 8c).
• Command line: run the "nestat -r" command.

#1> netstat -r
Routing tables
Destination        Gateway            Flags     Refs     Use  Interface
10.10.10.0/24      link#2             UC          0        0  eth0
10.10.10.101       00:0c:29:40:72:d0  UHL         5     3222  eth0
127.0.0.1          127.0.0.1          UH          0        0  lo0
224.0.0.0/8        127.0.0.1          UGS         0        0  lo0

Additional materials

Online courses 

1. InfiLINK 2x2 / InfiMAN 2x2: Initial Link Configuration and Installation
2. InfiLINK 2x2 and InfiMAN 2x2: Switching
3. InfiLINK XG Family Product
4. Quanta 5: Installation and Configuration

Webinars

1. Typical scenario of routing setting using Infinet Wireless devices. Part I.
2. Typical scenario of routing setting using Infinet Wireless devices. Part II
   

Other

1. InfiNet Wireless R5000 - Web GUI - Technical User Manual
2. InfiLINK XG / InfiLINK XG 1000 - Technical User Manual
3. Quanta 5 family - Technical User Manual
4. netstat command