Further, the high availability deployments and configurations that are supported on SD-WAN Edge are described in this section.
Please take note of the following topics:
The Management of SD-WAN Edge High Availability (HA)
The high availability solution guarantees that the flow of traffic will continue even in the event of outages. The VMware data plane component that is deployed at the branch location of an end user is referred to as the SD-WAN Edge. An SD-WAN Edge that is configured in High Availability (HA) mode is a mirror image of another SD-WAN Edge, and it appears on the SD-WAN Orchestrator as a single SD-WAN Edge.
SD-WAN Edges are deployed at the branch site in pairs of Active and Standby roles in a configuration that is designed to provide high availability. Both of these Edges have configurations that are identical to one another. By utilizing a failover link that is established over a wired wide area network connection, the Active and Standby Edges are able to exchange heartbeats. Assuming the identity of the Active Edge and taking over the traffic load, the Standby Edge will assume the role of the Active Edge in the event that it loses connectivity with the Active Edge for a certain amount of time. The failover has as little of an effect as possible on the flow of traffic.
All communication between the SD-WAN Orchestrator and the Active Edge is limited to that. Any modifications that are made to the Active Edge through the use of the Orchestrator are in sync with the Standby Edge through the utilization of the failover link.
Failure Situations
In the following situations, a switch from an active edge to a standby edge is likely to happen:
- Failed WAN link: A failover move is carried out when a WAN link on the Active Edge fails. The “High Availability Going Active” event is made by the SD-WAN Orchestrator. Because the peer’s WAN interface is down, this means that another WAN link on the Standby Edge will become Active.
- LAN link failure: A failover step is taken when a LAN link on the Active Edge fails. The “High Availability Going Active” event is made by the SD-WAN Orchestrator. Because the peer’s LAN interface is down, this means that another LAN link on the Standby Edge will become Active.
- Edge functions not responding, or Edge crash / reboot / unresponsive: The Standby Edge does not receive any heartbeat messages when the Active Edge crashes, reboots, or becomes unresponsive. When the “High Availability Going Active” event is generated by the SD-WAN Orchestrator, the Standby Edge takes the role of Active.
Models for High Availability Deployment
The following deployment models are supported by the High Availability feature:
- Standard HA: In this model, the WAN links that both the Active and Standby Edges are connected to are symmetrically coupled, meaning that their configurations are identical. On the Active Edge, all ports are available for receiving and delivering traffic. On the other hand, the Standby Edge has all ports closed but for GE1. Active and Standby Edges exchange heartbeats via the GE1 interface.
- Enhanced HA: In this model, the WAN links between the Active and Standby Edges are asymmetric, meaning they are connected to separate networks but have the same configurations. Active and Standby Edges exchange heartbeats via the GE1 interface. The WAN link that connects to the Standby Edge can be used by the Active Edge to transmit and receive traffic. The traffic is routed over the GE1 interface to the Standby Edge, which subsequently transmits it via the WAN link.
- Mixed-mode HA: This model combines on the same site the deployments of both Standard and Enhanced HA. The Active and Standby Edges in this variant are configured identically. Both symmetric and asymmetric connections are necessary.
The SD-WAN Edge platforms 510, 510N, 520, 520v, 540, 610, 610N, 620, 620N, 640, 640N, 680, 680N, 840, 2000, 3400, 3800, 3810, and any Virtual Edge are compatible with the HA options.
HA only works between SD-WAN Edge platform types that are the same.
In High Availability deployments, you cannot mix Edges that are Wi-Fi capable and those that are not. Even though the Wi-Fi capable and non-Wi-Fi capable versions of the same Edge model (e.g., 640 and 640N) seem identical, it is not supported to deploy them as a High-Availability pair. This applies to Edge models 510N, 610N, 620N, 640N, and 680N. In order to guarantee high availability, customers should link Edges that are either both Wi-Fi capable or both non-Wi-Fi capable.
Standard HA
Standard HA is described in this section.
A Look at the Standard HA Topology
The next picture gives you a general idea of what Standard HA is all about.
In order to create a failover link, the Edges, one of which is Active and the other of which is Standby, are connected by L1 ports. All ports, with the exception of the L1 port given for the failover link, are blocked by the Standby SD-WAN Edge.
Things Needed to Complete Standard HA
- All of the following configuration explanations call for LAN side switches that are either STP capable or already set up with STP.
- The LAN and WAN ports on the SD-WAN Edge must also be linked to different L2 switches. If you need to connect both ports to the same switch, you need to separate the LAN and WAN ports.
- The two SD-WAN Edges must have WAN and LAN links that are mirror images of each other.
Types of Deployments for Standard HA
Standard HA can be set up in two different ways:
- Type 1 deployment: High availability (HA) with L2 switches
- Type 2 deployment: High Availability (HA) with L2 and L3 switches
These two types of deployment are talked about below.
First Type of Deployment: HA with L2 switches
The network connections made simply with L2 switches are depicted in the following figure.
W1 and W2 are WAN links that both ISPs use to connect to the L2 switch and get through the WAN. Two SD-WAN Edges are linked by an L1 link, which is used for “keep-alive” and contact between the Edges for HA support. The LAN connectors on the SD-WAN Edge are used to link to the L2 switches at the access layer.
A Brief Investigation into HA Deployment with L2 Switches
- On both Edges, the same port needs to be linked to the same ISP link.
- Make the same ISP link accessible to both Edges by using the L2 switch.
- The Standby SD-WAN Edge blocks all of its ports except the failover link (L1 port) so it doesn’t get in the way of any traffic.
- By utilizing the failover link, the Active and Standby SD-WAN Edges are able to acquire and maintain synchronization of session information.
- However, if the Active Edge is equipped with an Active LAN link, it will also failover to the Standby in the event that it detects the loss of a LAN link.
Type 2 deployment: HA with L2 and L3 switches
The L2 and L3 switches are used to connect the networks shown in the figure that follows.
In order to establish a WAN connection with ISP1 and ISP2, respectively, the SD-WAN Edge WAN connections (W1 and W2) are utilized to link to L2 switches. A failover link for HA support is provided by connecting the L1 connections on the SD-WAN Edge. Multiple end-user devices are linked to L2 Switches through the VMware Edge LAN connections.
Using L2 and L3 switches for HA deployment considerations
- The L3 switch pair must provide HSRP/VRRP.
- The SD-WAN Edge’s fixed route goes to the HSRP VIP on the L3 switches as the next step to get to the end stations that are behind the L2 switches.
- Both SD-WAN Edges need to have the same Internet service provider (ISP) connectivity linked to the same port. It is necessary for the L2 switch to ensure that both Edges have access to the same Internet service provider link.
Because it blocks all of its ports with the exception of the failover link (L1 port), the Standby SD-WAN Edge does not disrupt any of the traffic that is being transmitted.
- Through the failover link, the session information is synchronized between the Active SD-WAN Edge and the Standby SD-WAN Edge.
- In addition, the HA pair performs a failover from Active to Standby when it detects the loss of the L1 link between the LAN and the WAN.
- A switchover to standby will happen if Active and standby have an equal number of up LAN links but standby has more up WAN connectivity.
- There will be a shift to the Standby Edge if it has more up LAN links and at least one up WAN link. It is assumed that there are more users on the LAN side of the Standby Edge than on the Active Edge. The Standby will also let more users on the LAN side connect to the WAN if there is any WAN connection available.
Enhanced HA
Enhanced HA is described in this section. With Enhanced HA, L2 Switches on the WAN side of the edges are no longer necessary. Please see the Standard HA literature if you are seeking information about LAN side settings. If the Active Edge notices that the WAN link(s) linked to the Standby Edge are different from the link(s) connected to itself, it will select this option.
A general representation of Enhanced HA is shown in the figure below.
An HA connection is used to establish a failover link through which the Edges, one of which is Active and the other of which is Standby, are connected. In order to construct overlay tunnels on both WAN lines (connected to itself and the Standby Edge), the Active Edge uses the HA link to connect to both of these links.
There shouldn’t be mirrored physical WAN connections between the two SD-WAN Edges. For example, the Standby Edge cannot use GE2 as its WAN link if GE2 is the WAN link on the Active Edge.
The Active Edge creates the overlay tunnel via the HA connection in order to utilize the WAN link that is linked to the Standby Edge. The HA link is used to forward LAN-side traffic to the Internet. The branch’s business policy specifies how traffic will be distributed across the overlay tunnels.
Enhanced HA support for the LTE interface
Based on the GSM/EDGE and UMTS/HSPA technologies, Long-Term Evolution (LTE) is a standard for wireless internet communication for cell phones and other data terminals. It boosts speed and capacity by using a different radio interface and making changes to the core network. The 510 and 610 Edge devices from VMware SD-WAN that have two SIM slots can use LTE.
The LTE link/CELL interface is now included in the HA election as of the 4.2 version. When compared to wired links, CELL links are internally given less weight. Therefore, the Edge that has the LTE connection can be either the Active or the Standby Edge, depending on the number of wired links connected to each Edge in the eHA pair. Potential applications of eHA with an LTE interface are as follows.
Case 1: One wired link on the active edge and one LTE link on the standby edge
An illustration of the topology of enhanced high availability support for LTE interface on a standby edge is provided in the figure. An example of a failover link is provided by the presence of two Edges, one of which is Active (SD-WAN Edge 1) and the other of which is Standby (SD-WAN Edge 2), which are connected to one another through the utilization of a HA cable. It is recommended that the wired WAN link Edge be used as the Active Edge. For the purpose of tunnel establishment, the Standby Edge makes use of an LTE link. It is possible to use the LTE link on the Standby Edge as either an active link, a backup link, or a hot-standby link, depending on the configuration of the Edge. Through the use of the HA link, the Active Edge creates overlay tunnels on the WAN connection that is connected to itself as well as the LTE link that is located on the Standby Edge. Should an Active Edge fail to function properly, the Standby Edge will continue to route the traffic from the local area network (LAN) across the LTE link.
Case 2: Active Edge with a single wired and one LTE link, and Standby Edge with a single wired link
The picture shows how Enhanced HA support for LTE Interface on an Active Edge is set up. The Active Edge in this case is SD-WAN Edge 1, which has one wired link and one LTE link. The Standby Edge is SD-WAN Edge 2, which only has one wired link. The Standby Edge would become Active if the Active Edge’s cable WAN link went down. The LTE link would then be used in eHA mode.
List of Topologies Supported
For HA to work, it’s necessary for the two models to be identical. The following topologies are compatible with the upgraded LTE HA support:
- 510 – 510 LTE HA pair
- 610 – 610 LTE HA pair
- 510 LTE – 510 LTE HA pair
- 610 LTE – 610 LTE HA pair
For 510-LTE pairs and 610-LTE pairs topology, you can’t put an LTE SIM card into the Active Edge while the Standby Edge already has an LTE SIM card on the CELL interface.
