WHAT TO EXPECT
High Availability (HA) is an implementation of data path redundancy and stream duplication. It protects users from data loss by replicating and sending packets through multiple network paths. xNIC compares packets received from those multiple paths and automatically reconstructs the original stream.
In this section, users will learn more about the benefits of implementing the High Availability feature in their cloudSwXtch and understand how to leverage it for their future needs.
Creating A More Resilient Network
With High Availability, critical workloads can be configured to be more resilient, stretching across regions or availability zones in a single cloud. In addition, it can be used across multiple cloud providers. Although there can only be up to eight redundant paths, there are no limits to the number of consumers that can receive the HA stream, other than bandwidth constraints.
In addition, there is no limit to the number of multicast groups per data path. If one cloud, availability zone or region should go down, then the data is still sent in the other 2-8 paths, ensuring that the consumer gets the necessary data. Consumers can also be put into different clouds, availability zones or regions so that if a consumer becomes unavailable, users can still sign into a different cloud, availability zone or region and get the data desired.
The HA feature forwards packets to the receiving application from any of the configured paths as soon as the "next" expected packet is received. Redundant packets from other paths are discarded. There is no additional latency imposed by the HA feature.
IMPORTANT
A cloudSwXtch configured in a HA path cannot be used in a cloudSwXtch mesh. They are mutually exclusive.
High Availability Example
The simple diagram above shows high availability with one multicast group 239.1.1.1:8804 originating from an on prem source. From the bridge, two paths are created with redundant packets being sent to alternate cloudSwXtches in different regions. The number of regions and cloud providers needed for High Availability will vary depending upon the customer's environment.
Independent path redundancy ensures no packet loss if every packet arrives at the consumer from at least one path. For example:
In the event that cloudSwXtch101 goes offline, the consumer will still get the multicast traffic via cloudSwXtch201 (or vice versa).
In the event that there are network issues in Region 1 where some of the packets are lost in path one, the consumer can still get the multicast traffic with High Availability pulling data from Region 2 in path two.
In the event that there are network issues in Region 1 and 2 where some of the packets are lost in both paths, both consumers can still get the multicast since the high availability function will take the valid packets and reconstruct the multicast stream from Region 1 and 2.
In each example, despite losing paths, multicast data was still able to get to the end point using high availability with no packet loss. Configuring more paths will ensure higher availability of the multicast group.
HA can be monitored via swxtch-top, see swxtch-top section 4-6.
To configure the system for high availability, refer to: High Availability Configuration.
Installing cloudSwXtch - Firewall Exceptions
When installing the cloudSwXtch, high availability requires special firewall exceptions. To learn more, see cloudSwXtch System Requirements.