InfiLINK 2x2 and InfiMAN 2x2: Switching

08

Traffic redundancy schemes

Common link redundancy scheme. Application area

The InfiNet equipment supports various traffic reservation schemes in the "1+1" mode. Hot standby can be built according to a scheme in which the main and backup links are organized with the InfiNet devices by using transmitting and receiving elements duplication.

Figure 4.1 - Common redundancy scheme

Generally, the redundancy scheme consists of two communication channels: an arbitrary L2 link and a radio link on the InfiNet equipment (see figure above). The link selection for direct traffic transmission is based on the one of service protocols running on the InfiNet devices or on the other manufacturers network equipment. Such protocols are:

  • MINT "Failover";
  • switching loop prevention protocols (STP for example);
  • dynamic routing protocols (OSPF for example).

MINT "Failover" application

Infinet devices InfiLINK 2x2 anb InfiMAN 2x2 families allows to build a redundant links. MINT "Failover" option allows to configure hot standby for any link, wired or wireless. In this case, the main link equipment type is not important. The only condition is connectivity at the L2 level between Infinet devices through both communication links.

The Failover option is a part of the MINT proprietary protocol, so the task implementation will require joining all devices into a single MINT area using RF and PRF interfaces. After joining devices into a single MINT area, there is a risk of a switching loop appearance. In this article, we will review two ways to prevent the switching loop appearance:

  • Using different switch groups on the main and backup links.
  • Using the switch border option.

The operation principle is very simple. The Infinet device, on which the MINT "Failover" is activated, checks for the specific MAC address presence in frames transmitted through the main link. If these frames with specified MAC address are present, then the backup link operation is blocked. In this case the master device of the backup radio link turns off the radio module and stops broadcasting, and the slave device only "listens" to the air. If as the main link is also used radio, the backup and main links can operate on the same frequency without mutual influence. When the monitored MAC address disappears from frames headers received by the device through the main link, the backup link starts to transmit the traffic. The process is fully automatic. However, the transition to a backup link cause a short connectivity interruption. The traffic transmission through the main link will be resumed only after the failure of the backup link.

If devices of InfiLINK 2x2 and InfiMAN 2x2 families are used for main link, it provides additional benefits. In this case both radio links can be configured with using MINT "Failover" option. For both links, the established connection quality will be automatically evaluated. The connection with the best performance will be automatically used as the main. Link selection parameters can be set manually.

It is also possible to connect two Infinet devices to one external antenna or use one radio frequency.

MINT "Failover" over the L2-link

In case the radio link between the Infinet devices is used as the backup link, and an arbitrary link is used as the main link, the configuration is performed in two steps:

  1. PRF interfaces are created on the Master 1 and Slave 1 devices.
  2. Enable MINT "Failover" option on the Master 1 device with Slave 1 MAC address tracking.

In this case, the Master 1 device will receive frames with the Slave 1 MAC address in two ways: by the RF interface via the backup radio link and by the PRF interface via the main channel. By default, the PRF interface has less cost than RF, so the radio interface of the Master 1 device will be turned off. If the main link fails, the Master 1 device will stop receiving frames with the monitored Slave 1 MAC address and turn on the radio interface. The connection between PC1 and PC2 will be restored via the backup radio link. At the time when the main link is restored, the Master 1 device will start to receive frames with the monitored MAC address via two links and turn off the radio module again.

Figure 4.2 - MINT "Failover" over L2-link scheme

The described redundancy configuration consist of the following steps:

    1. PRF interface configuration: go to the "Network Settings" section and click the "Create PRF" button. After that go to the "Link Settings" section, activate created PRF interface and join with RF interface:

Figure 4.3 - PRF interface configuration on Master 1 and Slave 1

    2. Switch groups configuration: It is necessary to create switch groups on Master 1 and Slave 1, and add eth* and rf5.0 interfaces:

Figure 4.4 - Switch group configuration

    3. MINT "Failover" configuration: MINT "Failover" can be configured by using command line, go to the "Command Line" section and enter the "mint rf5.0 failover MAC_ADDRESS" command, there "MAC_ADDRESS" is the Slave 1 MAC address:

Figure 4.5 - MINT "Failover" configuration

MINT "Failover" over the radio between Infinet devices

If both links main and backup are built on Infinet devices, there are possible schemes:

  • MINT "Failover" with MAC address tracking specified manually and different switch group IDs on the main and backup links;
  • MINT "Failover" with MAC address tracking selected automatically and different switch group IDs on the main and backup links;
  • MINT "Failover" with MAC address tracking specified manually and using the switch border option;
  • MINT "Failover" with MAC address tracking selected automatically and using the switch border option.
MINT "Failover" with MAC address tracking specified manually and different switch group IDs on the main and backup links

Scheme with different switch group IDs is shown below:


Figure 4.6 - MINT "Failover" with different switch group IDs

Device configuration is performed in the following sequence:

  • Step 1: create PRF interfaces on all devices (figure 4.3). The "switch border" option must be disabled;
  • Step 2: join the RF and PRF interfaces on all devices (figure 4.3);
  • Step 3: create switch groups on all devices (figure 4.4);
  • Step 3a: note that switch groups with different IDs should be used on the main and backup links. Otherwise, there is a chance of a loop appearance;
  • Step 4: activate the MINT "Failover" option on the Master 1 (figure 4.5). The Slave 1 MAC address must be specified as the tracking address for the Master 1.
  • Step 5: activate the MINT "Failover" option on the Master 2 (figure 4.5). The Slave 2 MAC address must be specified as the tracking address for the Master 2.
MINT "Failover" with MAC address tracking selected automatically and different switch group IDs on the main and backup links

This scheme coincides with the scheme where MAC address specified manually and is shown in figure 4.6. The devices configuration in accordance with the scheme is performed in the following sequence:

  • Stage 1 - 3a: similar to the scheme with MAC address specified manually.
  • Step 4: activate the MINT "Failover" option on Master 1 and Master 2 (figure 4.5). The command line will be as follows: "mint rf5. 0 failover auto".
  • Step 5: Master 1 and Master 2 will automatically select the MAC addresses of the devices to implement the MINT "Failover" option.
MINT "Failover" with MAC address tracking specified manually and using the switch border option

The scheme for configuring MINT "Failover" using the "switch border" option is shown in the figure below:


Figure 4.7 - MINT "Failover" with switch border option usage

Device configuration is performed in the following sequence:

  • Step 1: create PRF interfaces on all devices (figure 4.3). The "switch border" option must be enabled;
  • Step 2: join the RF and PRF interfaces on all devices (figure 4.3);
  • Step 3: create switch groups on all devices (figure 4.4);
  • Step 3a: note that switch groups with the same IDs can be used on the main and backup links;
  • Step 4: activate the MINT "Failover" option on the Master 1 (figure 4.5). The Slave 1 MAC address must be specified as the tracking address for the Master 1.
  • Step 5: activate the MINT "Failover" option on the Master 2 (figure 4.5). The Slave 2 MAC address must be specified as the tracking address for the Master 2.
MINT "Failover" with MAC address tracking selected automatically and using the switch border option

The scheme coincides with the scheme where MAC address specified manually and is shown in figure 4.7. The devices configuration is performed in the following sequence:

  • Stage 1 - 3a: similar to the scheme with MAC address specified manually.
  • Step 4: activate the MINT "Failover" option on Master 1 and Master 2 (figure 4.5). The command line will be as follows: "mint rf5. 0 failover auto".
  • Step 5: Master 1 and Master 2 will automatically select the MAC addresses of the devices to implement the MINT "Failover" option.

STP

In case an arbitrary L2 channel is main, and the backup link is a radio channel between InfiNet devices operating in the switching mode (Figure 4.1), there is a risk of a switching loop. Loops occurrence mechanism is described in the previous lesson. STP (spanning-tree protocol) is one of the solutions for loop prevention and fault tolerance.

STP using on the InfiNet units

The STP protocol can be started on InfiNet devices during switch groups configuration. In this case, the equipment will look for loops and prevent them by disabling the traffic transfer through some switch groups interfaces. An example of the STP operation is shown in the figure below. At the same time, in case of main link failure, the protocol will track this and change the data transfer scheme using a backup radiolink.

Figure 4.8 - Redundancy scheme with STP using

Create PRF interfaces and switch groups on Master 1 and Slave 1 devices (Figures 4.3-4.4). In switch group settings activate the STP operation and set the VLAN number associated with this switch group:

Figure 4.9 - STP configuration in switch group

In case the main link (Figure 4.8) is a radio link between InfiNet devices, apply the configuration above to all wireless devices.

STP on network equipment of other manufacturers

It is possible to implement a hot standby scheme in which the STP is started on the network equipment of other manufacturers, and the InfiNet wireless devices are used for the main and / or backup radio link. Such scheme is shown on the figure below. STP is enabled on Switches 1 and 2. By exchanging service BPDU frames, switches build a tree structure and, in the case of loop detection, block one of the ports. On the figure below, STP on switch 1 blocked the port to an arbitrary L2 link, choosing a radio link as the main for data transmission.

To implement this scheme, no additional settings on the InfiNet devices are required - it is enough to create switch groups (Figure 4.4).

Figure 4.10 - Redundancy scheme with STP using on the equipment of other manufacturers

OSPF protocol

One more method of building redundancy schemes is using dynamic routing protocols. In this the OSPF protocol will be reviewed, but using of any other routing protocol will not be significantly different.

OSPF protocol using on network equipment of other manufacturers

InfiNet devices support the OSPF protocol operation, but this course is to show the switching mode so we assume that the dynamic routing protocol is running on the network equipment of other manufacturers. The redundancy scheme is shown in the figure below.

OSPF is started on routers 1 and 2, which are connected by two links: a backup radio link between the InfiNet devices and an arbitrary L2/L3 link as the main. After successfully neighbor relations establishing and LSDB messages exchanging, routers calculate routes to each of the known networks and add them to the routing table. User traffic will be transmitted through the main link, but in case of its failure, the route through the radiolink will become active.

The InfiNet devices configuration is the same as in the previous example - it is enough to create switch groups (Figure 4.4).

Figure 4.11 Redundancy scheme with OSPF using on the equipment of other manufacturers

Redundancy in scheme with mobile objects

Popular scenario in projects with mobile objects is the situation when base stations are installed along the object's movement path, and two subscriber terminals are mounted on the object. Using of two subscriber stations allows to increase the reliability of communication due to the radiation pattern expansion: subscriber terminals are mounted with different azimuths, therefore, at least one subscriber terminal will be in the coverage area of one Base Station sector.

The example of using two Base Stations sectors is shown in the figure below. A mobile object is a cargo ship plying in a limited water area, constantly located in the coverage area of at least one Base Station sector. The ship has two subscriber terminals which provide fault tolerance for a wireless connection:

  • in ship central position (see Figure) CPE1 is connected with BS1, CPE2 - with BS2;
  • in the leftmost position CPE2 is connected with BS1;
  • in the rightmost position CPE1 is connected with BS2.

Figure 4.12 - Project scenario with mobile objects

To explain the device configuration, let's look at the scheme in Figure 4.12 from different point of view. The scheme with configuration units is shown in Figure 4.13:

  • The ship has the user device - PC2, which is connected to CPE1 and CPE2 via the InfiMUX switch.
  • BS1 and BS2 sectors and connected to them the InfiMUX switch are situated ashore.
  • Also, a switch connected to InfiMUX which assigns VLAN tags to traffic from the user device - PC1, and the management device - PC3.
  • VLAN 100 is chosen for management, InfiNet devices have switch group 100 for this purpose.
  • VLAN 10 is chosen for data traffic and appropriate switch group 10 is created on InfiMUX switches.

Figure 4.13 - Scheme for connectivity organisation

Configuration

Joining devices into a single MINT area
  • Step 1: create PRF interfaces on all devices: InfiMUX1, InfiMUX2, BS1, BS2, CPE1 and CPE2 (see Figure 4.3).
  • Step 2: join PRF and RF interfaces on BS1, BS2, CPE1 and CPE2 devices (see Figure 4.3).
  • Step 3: on each InfiMUX create two PRF interfaces and join them (see Figure 4.3).
Device management organization

The solution involves devices management only from PC3 and, since all the InfiNet devices are in a single MINT area, a switch group must be created on InfiMUX, which will be the gateway between Ethernet and MINT. On other devices, switch groups for management traffic should not transfer frames between MINT and Ethernet; therefore, it is enough to include only radio interface.

  • Step 1: create VLAN interface for management traffic on InfiMUX1. Go to "Basic Settings → Network settings" and click the "Create VLAN" button set group number 100 and set the same VLAN ID, eth1 interface must be selected as a parent.
  • Step 2: create switch group 100 on InfiMUX1, add vlan100 and prf2 interfaces (see Figure 4.14), provide same settings on InfiMUX2, but add only prf2 interface to switch group. Create switch group 100 on each wireless device and add only rf5.0 interface (see Figure 4.14).
  • Step 3: for all devices in created switching group settings click the "Create L3 management" button, the svi100 interface will be created automatically (see Figure 4.14).
  • Step 4: for all devices, go to the "Basic Settings → Network settings" section and assign an IP address to the svi100 interface (see Figure 4.15).

Figure 4.14 Management switch group configuration

Figure 4.15 Assignment an IP address to the svi100 interface

Data traffic transmission

It is necessary to create switch groups with number 10 on InfiMUX1 and InfiMUX2, since these devices are gateways for user traffic between MINT and Ethernet. On wireless devices, switch groups for user data are not created, since they are transit devices within the MINT domain.

  • Step 1: on InfiMUX, go to the "Basic Settings → MAC Switch" section and click the "Create switch group" button (see Figure 4.16).
  • Step 2: add eth1 and prf2 interfaces to created switch groups on both InfiMUX devices (see Figure 4.16). 
  • Step 3: add rule "vlan 10" to created group (see Figure 4.16).

Figure 4.16 Data traffic switch group configuration

Previous Next